Scoring methods for anti-FRalpha antibody-drug conjugate therapy

Abstract

A method for predicting the efficacy of antibody drug conjugate (ADC) therapy includes determining an average optical density (OD) of cell membrane and optional cytoplasm staining of each cancer cell in a full slice image. The diagnostic antibody linked to the dye targets the same folate receptor alpha protein as the ADC antibody. A statistical operation is performed on the stained OD, such as calculating a median, an absolute deviation, a difference in cell membrane and cytoplasm stained OD, a percentage of cells of minimum OD, or a sum of the lowest stained cell plus inadequately stained cells near the stained cell. For example, zero, high, or low ADC doses are recommended based on whether the median OD falls below, between, or above two OD thresholds, respectively. The upper and lower OD thresholds are associated with a response of a training patient treated with the ADC and having analyzed its stained cancer tissue.

Description

Technical Field

[0001] The present invention relates to a method for determining a score that indicates how a cancer patient will respond to therapy using an anti-FRalpha antibody-drug conjugate. Background Art

[0002] Assessing the probability of a cancer patient's response to a given treatment is an important step in determining a treatment regimen for a cancer patient. This assessment is typically based on histological analysis of tissue samples from cancer patients, and involves using standard grading schemes to identify and classify cancer. Immunohistochemistry (IHC) staining can be used to distinguish marker-positive cells expressing a specific protein from marker-negative cells that do not express the protein. IHC staining typically involves a variety of dyes, including one or more dyes connected to a protein-specific antibody and another dye as a counterstain. A common counterstain is hematoxylin, which marks DNA and therefore stains the nucleus.

[0003] Protein-specific stains or biomarkers can be used to identify areas in the tissues (e.g., tumor tissues) of cancer patients that may show responses to a predetermined therapy. For example, biomarkers of stained epithelial cells can help identify suspected tumor areas. Other protein-specific biomarkers are then used to characterize cells in cancerous tissues. Cells stained with specific biomarkers can be identified and quantified, and the scores indicating the number of positively stained cells and negatively stained cells can be visually estimated by a pathologist subsequently. This score can then be compared with the scores of other cancer patients calculated in the same manner. If the response of these other patients to a given cancer treatment is known, the pathologist can predict the possibility that the cancer patient has a response to a specific treatment based on the comparison of the scores calculated for the cancer patient with the scores of other patients. However, the visual assessment of the pathologist is prone to variability and subjectivity.

[0004] A promising cancer treatment involves an antibody-drug conjugate (ADC) having a cytotoxic drug conjugated to an antibody whose antigen is expressed on the surface of cancer cells. The ADC binds to the antigen and undergoes cellular internalization so that the drug is selectively delivered to cancer cells and accumulates within those cancer cells and kills the cancer cells.

[0005] Folate receptor (FR) is a membrane-bound protein present on the cell surface and can therefore be used to develop new ADCs. The FR family includes FRα, FRβ, FRγ and FRδ. FR binds to folate molecules and transports them into cells, so that folate molecules are delivered to the folate cycle to support the metabolism of nucleotides. In particular, folate is important for DNA synthesis, methylation and repair (Cheung et al., Oncotarget. 2016; 7(32): 52553-52574).

[0006] FRα is a glycosylphosphatidylinositol (GPI) anchored membrane protein with high affinity for 5-methyltetrahydrofolate (5-MTF), the active form of folate. Previous studies have shown that FRα plays a vital role in embryogenesis (Kelemen, Int J Cancer. 2006; 119 (2): 243-250). However, folate transport in adults is mainly driven by the ubiquitous expression of reduced folate carriers and proton-coupled folate transporters (Zhao et al., Annu Rev Nutr. 2011; 31: 177-201). The distribution of FRα expression in adults is usually limited to the apical surface of polarized epithelial cells, such as the choroid plexus, kidney, lung and placenta.

[0007] Overexpression of FRα (which is also known as folate binding protein (FBP) and is related to the gene FOLR1) is common in tumor cells such as ovarian cancer, lung cancer (e.g., non-small cell lung cancer (NSCLC)), and breast cancer (Shi et al., Drug Des Devel Ther. 2015; 9: 4989-4996). In particular, previous studies have found that the level of soluble FRα is elevated in the blood of ovarian cancer patients, supporting the potential application of FRα as a biomarker for early ovarian cancer (Basal et al., PLoS One. 2009; 4 (7): e6292). Preclinical ovarian models also revealed that overexpression of FRα is associated with tumor progression, and that the binding of folate to FRα can mediate the activation of the proto-oncogene STAT3 (Hansen et al., Cell Signal. 2015; 27 (7): 1356-1368). Summary of the invention

[0008] A computer-based method was sought for generating a reproducible and objective score that predicts a cancer patient's response to treatment involving a therapeutic FRα antibody-drug conjugate.

[0009] A method for predicting how a cancer patient will respond to a therapy involving an antibody drug conjugate (ADC) comprises determining a predictive efficacy score based on the optical density of the cell membrane and optionally cytoplasmic staining by a dye attached to a diagnostic antibody. The ADC comprises an ADC payload and an ADC antibody that targets a protein on each cancer cell. Both the diagnostic antibody and the ADC antibody target the folate receptor alpha (FRα) protein on the cancer cell.

[0010] The tissue sample is immunohistochemically stained using a dye attached to a diagnostic antibody that binds to the FRα protein on the cancer cells in the tissue sample. A digital image of the tissue sample is acquired. The digital image is image analyzed using a convolutional neural network to detect cancer cells. For each cancer cell, the optical density of the staining is determined based on the staining intensity of the dye in the cell membrane and / or cytoplasm of the cancer cell and / or the cell membrane and cytoplasm of other cancer cells that are less than a predefined distance from the cancer cell. Based on various statistical operations performed on the optical density of the staining, a predictive efficacy score is generated to predict the response of a cancer patient to an ADC therapy. Patients with scores above a predetermined threshold are recommended for therapies involving ADCs.

[0011] In one embodiment, a method for predicting the response of a cancer patient to an anti-FRα ADC involves detecting cancer cells and determining the average optical density (optionally the median of the average optical density) of cell membrane staining of all cancer cells in a digital image of a tissue sample from the cancer patient. The ADC includes an ADC payload and an ADC antibody targeting the FRα protein on the cancer cell. The tissue sample is immunohistochemically stained using a dye attached to a diagnostic antibody. The diagnostic antibody binds to the FRα protein on the cancer cell in the tissue sample. A digital image of the tissue sample is acquired, and the cancer cells in the digital image are detected. For each cancer cell, the average staining intensity of the dye in the cell membrane is measured. In one embodiment, the percentage of cancer cells with a staining intensity stronger than a predetermined threshold is measured. In another embodiment, the staining intensity of the dye in the cell membrane of other cancer cells that are less than a predefined distance from the cancer cell is also measured. The staining intensity of each cell membrane is calculated based on the average optical density of the brown diaminobenzidine (DAB) signal in the pixel of the cell membrane. The response of the cancer patient to the ADC is predicted based on whether the predicted efficacy score exceeds a predetermined threshold. When the predicted efficacy score exceeds a predetermined threshold, a therapy involving ADC can be administered to the cancer patient.

[0012] In embodiments involving median optical density falling above a threshold, a predictive efficacy score is generated to predict the response of a cancer patient to an ADC, the ADC comprising an ADC payload and an ADC antibody targeting a folate receptor alpha (FRα) protein on a cancer cell. Immunohistochemical staining is performed on a tissue sample of a cancer patient using a dye attached to a diagnostic antibody, the diagnostic antibody being bound to the FRα protein on the cancer cell in the tissue sample. A digital image of the tissue sample is acquired. For each cancer cell, the average optical density of the cell membrane of the cancer cell stained by the dye is determined. The median optical density of the staining of all cancer cells in the digital image is then determined. A predictive efficacy score for the tissue sample is generated based on the median optical density. If the median optical density is equal to or greater than the optical density threshold, the predictive efficacy score is positive, and if the median optical density is less than the optical density threshold, the predictive efficacy score is negative. The optical density threshold is associated with the response of a training patient group treated with ADC. If the predictive efficacy score is positive, a therapy involving ADC is recommended to the cancer patient. When the predictive efficacy score is positive, a therapy involving ADC may be administered to the cancer patient.

[0013] In embodiments involving the percentage of cancer cells whose average optical density of staining falls above a minimum optical density threshold, a predictive efficacy score is generated to predict the response of a cancer patient to an ADC, the ADC comprising an ADC payload and an ADC antibody targeting a FRα protein on a cancer cell. A tissue sample is immunohistochemically stained using a dye attached to a diagnostic antibody that binds to a protein on a cancer cell in the tissue sample. A digital image of the tissue sample is acquired. Cancer cells in the digital image are detected using image analysis. For each cancer cell, the optical density of the staining of the dye in the cell membrane is determined. (i) If the optical density of the cancer cell is equal to or greater than the optical density threshold, each cancer cell is identified as optical density positive, or (ii) If the average optical density of the cancer cell is less than the optical density threshold, each cancer cell is identified as optical density negative. Based on the percentage of cancer cells that are optical density positive in the digital image, a predictive efficacy score for the tissue sample is generated. If the percentage of cancer cells that are optical density positive is equal to or greater than the percentage threshold, the predictive efficacy score is positive, and if the percentage of cancer cells that are optical density positive is less than the percentage threshold, the predictive efficacy score is negative. The optical density threshold and the percentage threshold are associated with the response of a training patient group treated with the ADC. If the predicted efficacy score is positive, a therapy involving ADC is recommended to the cancer patient. When the predicted efficacy score is positive, a therapy involving ADC can be administered to the cancer patient.

[0014] In an embodiment involving the spatial proximity of various stained cells, a cancer patient treated with an ADC is identified, the ADC comprising an ADC payload and an ADC antibody targeting a FRα protein on a cancer cell. A tissue sample from a cancer patient is immunohistochemically stained using a dye attached to a diagnostic antibody, the diagnostic antibody being bound to the FRα protein on the cancer cell in the tissue sample. A digital image of the tissue sample is acquired. Cancer cells in the digital image are detected. For each cancer cell, the average optical density stained by the dye in the cell membrane is determined. (i) If the average optical density of the cancer cell is equal to or greater than an optical density threshold, each cancer cell is identified as optical density positive, or (ii) If the average optical density of the cancer cell is less than the optical density threshold, each cancer cell is identified as optical density negative. A proximity score for the tissue sample is generated, the proximity score being equal to the percentage of cancer cells in the digital image that are optical density positive or optical density negative, but within a predefined distance of an optical density positive cancer cell. If the proximity score exceeds a predetermined percentage threshold, the cancer patient is identified as a patient who will likely benefit from the administration of the ADC. When the proximity score exceeds a predetermined percentage threshold, a therapy involving the ADC may be administered to the cancer patient.

[0015] In an embodiment involving the median absolute deviation of the optical density of the staining of cancer cells in a full-slice image, a predictive efficacy score is generated to predict the response of a cancer patient to an ADC, the ADC comprising an ADC payload and an ADC antibody targeting a FRα protein on a cancer cell. The tissue sample is immunohistochemically stained using a dye attached to a diagnostic antibody, the diagnostic antibody being bound to the FRα protein on the cancer cell in the tissue sample. A digital image of the tissue sample is acquired. Cancer cells in the digital image are detected. For each cancer cell, the average optical density stained by the dye in the cell membrane of the cancer cell is determined. The median optical density of all cancer cells in the digital image is then determined. The median absolute deviation of the optical density of the cancer cell in the digital image from the median optical density of all cancer cells is determined. A predictive efficacy score for the tissue sample is generated based on the median absolute deviation. If the median absolute deviation is equal to or greater than the deviation threshold, the predictive efficacy score is positive, and if the median absolute deviation is less than the deviation threshold, the predictive efficacy score is negative. The deviation threshold is associated with the response of a training patient group treated with ADC. If the predictive efficacy score is positive, a therapy involving ADC is recommended to the cancer patient. When the predictive efficacy score is positive, a therapy involving ADC may be administered to the cancer patient.

[0016] In an embodiment involving recommending an effective dose of an ADC based on cell membrane staining, a recommended dose for treating a cancer patient with an ADC is determined, the ADC comprising an ADC payload and an ADC antibody targeting a FRα protein on a cancer cell. A tissue sample is immunohistochemically stained with a dye attached to a diagnostic antibody that binds to a FRα protein on a cancer cell in the tissue sample. A digital image of the tissue sample is acquired. Cancer cells in the digital image are detected. For each cancer cell, the average optical density stained by the dye in the cell membrane of the cancer cell is determined. The median optical density of all cancer cells in the digital image is then determined. The recommended dose is determined based on whether the median optical density falls below a lower optical density threshold, falls between a lower optical density threshold and an upper optical density threshold, or falls above an upper optical density threshold. If the median optical density falls below the lower optical density threshold, the recommended dose is zero; if the median optical density falls between the lower optical density threshold and the upper optical density threshold, the recommended dose is a higher dose; if the median optical density falls above the upper optical density threshold, the recommended dose is a lower dose. The lower optical density threshold and the upper optical density threshold are associated with the response of a training patient group treated with ADC. Using this method, a treatment recommendation involving ADC and a recommended dose is provided to a cancer patient. A recommended dose of a therapy involving ADC can be administered to a cancer patient based on the median optical density of all cancer cells in the digital image.

[0017] In an embodiment involving the difference between the optical density of cell membrane staining and cytoplasm staining, a predictive efficacy score is generated to predict the response of a cancer patient to an ADC, the ADC comprising an ADC payload and an ADC antibody targeting a FRα protein on a cancer cell. The tissue sample is immunohistochemically stained using a dye attached to a diagnostic antibody that binds to proteins on the cell membrane and in the cytoplasm of cancer cells in the tissue sample. A digital image of the tissue sample, such as a full-slice image, is acquired. Cancer cells in the digital image are detected using image analysis, optionally involving a convolutional neural network. For each cancer cell, the average optical density of the cell membrane stained by the dye is determined. For each cancer cell, the average optical density of the cytoplasm stained by the dye is determined. For each cancer cell in the digital image, the difference between the average optical density of the staining of the cell membrane and the average optical density of the staining of the cytoplasm is determined. From all cancer cells in the digital image, the 85% quantile of the difference between the average optical density of the staining of the cell membrane and the average optical density of the staining of the cytoplasm is identified. If the 85% quantile of the difference exceeds a predetermined difference threshold, a therapy involving ADC is recommended to the cancer patient. When the 85% quantile of the difference exceeds a predetermined difference threshold, a therapy involving the ADC can be administered to the cancer patient.

[0018] Other embodiments and advantages will be described in the detailed description below. This summary is not intended to limit the present invention. The present invention is defined by the claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings illustrate embodiments of the present invention, wherein like numerals refer to like parts.

[0020] Figure 1 The amino acid sequences of various CDRs of various constructs of ADC anti-FRα antibodies are shown (SEQ ID No.: 1 to SEQ ID No.: 36).

[0021] Figure 2 The amino acid sequences of the heavy chain and light chain variable regions of various constructs of the anti-FRα antibody are shown (SEQ ID No.: 37 to SEQ ID No.: 48).

[0022] Figure 3 The amino acid sequences of the corresponding heavy and light chains of the three constructs of the anti-FRα antibody are shown (SEQ ID No.: 49 to SEQ ID No.: 54).

[0023] Figure 4 The amino acid sequences of the corresponding heavy and light chains of three additional constructs of anti-FRα antibodies are shown (SEQ ID No: 55 to SEQ ID No.: 60).

[0024] Figure 5 The amino acid sequences of various light chain and heavy chain variable framework regions of anti-FRα antibodies are shown (SEQ ID No.: 61 to SEQ ID No.: 108).

[0025] Figure 6 The amino acid sequences of the three constant domain regions of the anti-FRα antibody are shown (SEQ ID No.: 109 to SEQ ID No.: 111).

[0026] Figure 7 is a table showing three exemplary numbering systems for identifying sequences to define complementarity determining regions (CDRs).

[0027] Figure 8 The amino acid sequences of the FRα proteins are shown (SEQ ID No.: 112 to SEQ ID No.: 113).

[0028] Fig. 9 is a flow chart of the steps by which an analysis system analyzes a digital image of a stained tissue to predict how a cancer patient will likely respond to an anti-FRα ADC based on the percentage of cancer cells that are positively stained or non-positively stained, but in the vicinity of positively stained cells.

[0029] Fig.10 Show instructions Fig. 9 Step 12 of the image analysis process is to process the digital image.

[0030] Fig.11 The image analysis steps for detecting the cell nucleus objects of cancer cells are shown.

[0031] Fig.12 Shows the image analysis steps for detecting the cell membrane using the cell nucleus object.

[0032] Fig.13 is a screenshot of the results of the image analysis step in the image analysis software environment.

[0033] Fig.14 Shown are sample calculations of binary spatial proximity scores for ten exemplary cells based on cell separation to reflect uptake of ADC payload into neighboring cells.

[0034] Fig.15 The mechanism by which anti-FRα ADC therapy kills cancer cells is shown.

[0035] Fig.16 Shown are sample quantification results of staining intensity from image analysis using gray values ​​of cell membrane and cytoplasm pixels.

[0036] Fig.17 List Fig.16 Images are given for exemplary quantification of the staining on the cell membrane and in the cytoplasm.

[0037] Fig.18 The cell separation Fig.17 Calculation of a continuous spatial proximity score for each of the three cells shown in to reflect the uptake of ADC payload into neighboring cells.

[0038] Fig.19 Shown are graphs and bar charts correlating binary spatial proximity scores with actual results from patient-derived xenograft (PDX) studies of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0039] Fig. 20 is a graph showing the binary spatial proximity scores in increasing order of size for 28 PDX models of ovarian cancer, NSCLC, and endometrial cancer.

[0040] Fig.21 Receiver operating characteristic (ROC) curve used to determine the optimal cut-off point to stratify responding and non-responding PDX models.

[0041] Fig. 22is a bar graph of binary space proximity scores for 197 tissue samples from a database of ovarian cancer patients sorted by decreasing size.

[0042] Fig.23 is a flow chart of the steps of a method for predicting the efficacy of an anti-FRα ADC based on the percentage of cancer cells that exhibit at least minimal staining in whole-slide images.

[0043] Fig.24 Shown are graphs and bar charts correlating predictive efficacy scores based on the percentage of minimally stained cells with actual results for PDX studies of patients with ovarian, NSCLC, colon, and endometrial cancers.

[0044] Fig.25 is a graph showing the predictive efficacy scores (based on % of cells least stained) in ascending order of size for 28 PDX models from patients with ovarian cancer, NSCLC, and endometrial cancer.

[0045] Fig.26 is a ROC curve for determining the optimal cutoff point for stratifying responding and non-responding PDX models based on the percentage of minimally stained cells.

[0046] Fig. 27 is a bar graph of the predictive efficacy scores (based on % of cells least stained) sorted by size in descending order for a database of 197 tissue samples from ovarian cancer patients.

[0047] Fig.28 is a flow chart of the steps of a method for predicting the efficacy of an anti-FRα ADC based on the median of the absolute deviation of the staining of each cell in a digital image from the median cell membrane staining of all cancer cells.

[0048] Fig.29 Shown are graphs and bar charts correlating predictive efficacy scores based on the median absolute deviation of staining intensity with actual results for PDX studies of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0049] Fig.30 is a graph showing the predictive efficacy scores (median absolute deviation of staining) in ascending order of size for 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients.

[0050] Fig.31 is a ROC curve for determining the optimal cutoff point for stratifying responding and non-responding PDX models based on the median absolute deviation of staining.

[0051] Fig.32is a bar graph of the predictive efficacy scores (median absolute deviation of staining) sorted by decreasing size for a database of 197 tissue samples from ovarian cancer patients.

[0052] Fig.33 is a flow chart of the steps of a method for determining a recommended dose of an anti-FRα ADC based on the median optical density of cell membrane staining in a digital image.

[0053] Fig.34 Shown is a graph showing the correlation between the median optical density of cell membrane staining and tumor response to 5 mg / kg anti-FRα ADC for PDX models from ovarian cancer, NSCLC, and endometrial cancer patients.

[0054] Fig.35 is a graph showing the predicted efficacy scores (based on median optical density of cell membrane staining) in ascending order of size for 28 PDX models from patients with ovarian cancer, NSCLC, and endometrial cancer for 5 mg / kg of anti-FRα ADC.

[0055] Fig.36 Figure 2 is a ROC curve for determining the optimal cutoff point for stratifying responding and non-responding PDX models using 5 mg / kg anti-FRα ADC based on the median optical density of cell membrane staining.

[0056] Fig.37 Shown is a graph showing the correlation between the median optical density of cell membrane staining from PDX models of ovarian cancer, NSCLC, and endometrial cancer patients and tumor response to 2.5 mg / kg anti-FRα ADC.

[0057] Fig.38 is a graph showing the predicted efficacy scores (based on median optical density of cell membrane staining) in ascending order of size for 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients for 2.5 mg / kg of anti-FRα ADC.

[0058] Fig.39 Figure 2 is a ROC curve for determining the optimal cutoff point for stratifying responding and non-responding PDX models using 2.5 mg / kg anti-FRα ADC based on the median optical density of cell membrane staining.

[0059] Fig.40 is a graph showing similar distribution of FRα protein in ovarian cancer PDX models and human ovarian tumors.

[0060] Fig.41is a bar graph of predicted efficacy scores (based on median optical density of cell membrane staining) for 197 ovarian cancer patients sorted by size in descending order, showing upper and lower optical density thresholds for stratifying patients who will receive higher, lower, and zero doses of anti-FRα ADC.

[0061] Fig.42 is a flow chart of the steps of a method for predicting the efficacy of an anti-FRα ADC based on the difference between the optical density of cell membrane staining and cytoplasmic staining.

[0062] Fig.43 Shown is a graph correlating the predictive efficacy score based on the difference between cell membrane and cytoplasmic staining intensity with actual results of PDX studies of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0063] Fig.44 is a graph showing the predictive efficacy score (difference between cell membrane and cytoplasm OD) in ascending order of size for 28 PDX models from ovarian cancer, NSCLC, and endometrial cancer patients.

[0064] Fig.45 Figure 2 is a ROC curve for determining the optimal cutoff point for stratifying responding and non-responding PDX models based on the difference between cell membrane and cytoplasmic staining intensity.

[0065] Fig.46 is a bar graph of the predictive efficacy scores (difference between cell membrane and cytoplasm OD) for 197 tissue samples from a database of ovarian cancer patients sorted by decreasing size. DETAILED DESCRIPTION

[0066] The present invention relates to a novel method for predicting a cancer patient's response to an antibody drug conjugate (ADC), the ADC comprising an ADC antibody targeting a folate receptor alpha (FRα) protein on a cancer cell, wherein the response is predicted based on a statistical operation using measured optical density of dye staining linked to a diagnostic antibody that also targets the FRα protein. Other aspects of the invention relate to methods for treating cancer patients by administering a therapy involving an ADC based on a quantitative continuous score (QCS) calculated using one of the statistical operations. Other aspects of the invention relate to methods for identifying cancer patients for treatment with an ADC based on a QCS score.

[0067] The first embodiment relates to a novel method for determining a recommended dose of a FRα antibody-drug conjugate (ADC) for cancer patients based on whether the median optical density of all cancer cells in a tissue sample stained with a diagnostic antibody that binds to the same protein as the ADC antibody falls below, between, or above two optical density thresholds determined based on the responses of a training cohort of patients treated with the ADC.

[0068] A second embodiment is directed to a method for generating a predictive efficacy score indicating how the cancer patient will respond to a treatment involving an ADC based on the median optical density of all cancer cells in a tissue sample of a cancer patient stained with a diagnostic antibody that binds to the same protein as the ADC antibody.

[0069] A third embodiment is directed to a method for generating a predictive efficacy score indicative of how a cancer patient will respond to an ADC based on the percentage of cancer cells in a tissue sample of the patient that exhibit a staining optical density equal to or greater than an optical density threshold.

[0070] A fourth embodiment is directed to identifying cancer patients who would likely benefit from ADC administration based on a proximity score equal to the percentage of cancer cells in a tissue sample of the patient that have an average optical density of staining above an optical density threshold or that have an average optical density of staining below an optical density threshold but are located within a predefined distance of cells having an average optical density of staining above the optical density threshold.

[0071] A fifth embodiment relates to a method for generating a predictive efficacy score indicating how a cancer patient will respond to an ADC based on a median absolute deviation (MAD) of the optical density of stained cancer cells in a digital image of a tissue sample of the patient and the median optical density of all cancer cells in the digital image, wherein if the predictive efficacy score exceeds a predetermined deviation threshold, then the predictive efficacy score predicts a positive response of the patient to the ADC.

[0072] A sixth embodiment relates to a method for generating a predictive efficacy score indicating how a cancer patient will respond to an ADC based on identifying the 85% quantile of the difference between the optical density of cell membrane staining and the optical density of cytoplasmic staining. If the predictive efficacy score exceeds a difference threshold, the predictive efficacy score predicts a positive response of the patient to the ADC.

[0073] I. General Definitions

[0074] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present disclosure belongs. Singleton et al., DICTIONARY OF MICROBIOLOGY AND MOLECULAR BIOLOGY, 20th Edition, John Wiley and Sons, New York (1994); and Hale and Marham, THE HARPER COLLINSDICTIONARY OF BIOLOGY, Harper Perennial, NY (1991) provide the skilled artisan with a general dictionary of many of the terms used in the present disclosure.

[0075] Unless otherwise indicated, any nucleic acid sequence is written from left to right in a 5' to 3' orientation. Amino acid sequences are written from left to right in an amino to carboxyl orientation, respectively. Amino acids are referred to herein by amino acid names, three-letter abbreviations or single-letter abbreviations. The term "protein" as used herein includes proteins, polypeptides and peptides. As used herein, the term "amino acid sequence" is synonymous with the term "polypeptide" and / or the term "protein". In some cases, the term "amino acid sequence" is synonymous with the term "peptide". The terms "protein" and "polypeptide" are used interchangeably herein. In the present disclosure and claims, conventional single-letter and three-letter codes for amino acid residues may be used. The 3-letter code for amino acids is defined according to the Joint Commission on Biochemical Nomenclature (JCBN) of IUPAC IUB. Due to the degeneracy of the genetic code, a polypeptide may be encoded by more than one nucleotide sequence.

[0076] As used herein, the term "about" encompasses an acceptable degree of error for a quantity measured given the nature or accuracy of the measurement. An exemplary degree of error is within 20% (%) of a given value or range of values, typically within 10%, and more typically within 5%. Preferably, the term "about" as used herein includes plus or minus (±) 5%, preferably ±4%, ±3%, ±2%, ±1%, ±0.5%, ±0.1% of the numerical value of the number to which it is used.

[0077] Some embodiments of the novel prediction methods disclosed herein relate to a quantitative continuous score (QCS). The QCS may indicate an indication of a predicted efficacy score, a recommended dose, or a predicted survival time. As used herein, the term "QCS positive" refers to a cancer that is likely to show a response to an anti-FRα ADC therapy. The term "QCS negative" refers to a cancer that is unlikely to show a response to an anti-FRα ADC therapy.

[0078] II. Anti-FRα Antibodies for ADC

[0079] A. Antibody Sequences

[0080] The antibody or antigen-binding fragment of an ADC as defined herein includes the listed complementary determining region (CDR) sequences or variable heavy chain and variable light chain sequences (reference antibody), as well as functional variants thereof. The functional variant binds to the same target antigen as the reference antibody, and preferably exhibits the same antigen cross-reactivity as the reference antibody. Compared to the reference antibody, the functional variant may have a different affinity for the target antigen, but preferably substantially the same affinity.

[0081] In some embodiments, the functional variant of the reference antibody exhibits sequence variation at one or more CDRs compared to the corresponding reference CDR sequence. Therefore, functional antibody variants may include functional variants of CDRs. When the term "functional variant" is used in the context of a CDR sequence, this means that the CDR has at most 2, preferably at most 1 amino acid difference compared to the corresponding reference CDR sequence, and when combined with the remaining 5 CDRs (or variants thereof), the variant antibody is able to bind to the same target antigen as the reference antibody, and preferably exhibits the same antigen cross-reactivity as the reference antibody. Functional variants may be referred to as "variant antibodies".

[0082] Figure 1-6 The CDR sequences, VH and VL sequences, heavy and light chain sequences, FR sequences and constant domain sequences of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095 and AB1370117 are shown. Figure 1-6 If there is any difference in the sequence Figure 1-6 The sequence in shall prevail.

[0083] Figure 1 Various CDRs associated with ADC antibodies are shown. In some embodiments, the ADC antibody or antigen-binding fragment thereof includes CDRs associated with any of the six constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095, or AB1370117, wherein the CDRs are determined by Kabat, Chothia, or IMGT. Figure 1 The sequences of the CDRs as defined by Kabat are listed as SEQ ID NO: 1 to SEQ ID NO: 36.

[0084] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0085] The heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN);

[0086] Heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS);

[0087] The heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY);

[0088] Light chain CDR1 of SEQ ID NO:4 (RASQSISSWLA);

[0089] The light chain CDR2 of SEQ ID NO:5 (KASGLES); and

[0090] Light chain CDR3 of SEQ ID NO:6 (QQYNSYSQLT).

[0091] Figure 2 The sequences of the CDRs are listed as SEQ ID NO:37 to SEQ ID NO:48.

[0092] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a heavy chain variable region (VH) comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 37; and a light chain variable region (VL) comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, or at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 38. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the VH of SEQ ID NO: 37 and the VL of SEQ ID NO: 38.

[0093] Figure 3 The amino acid sequences of the heavy chains (VH-CH) and light chains (VL-CL) of constructs AB1370049, AB1370026 and AB1370035 are shown as SEQ ID NO: 49 to SEQ ID NO: 54. Figure 4 The amino acid sequences of VH-CH and VL-CL of constructs AB1370083, AB1370095 and AB1370117 are shown as SEQ ID NO: 55 to SEQ ID NO: 60.

[0094] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 49; and a light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 50. In some embodiments, the ADC anti-FRα antibody comprises a heavy chain amino acid sequence of SEQ ID NO: 49 and a light chain amino acid sequence of SEQ ID NO: 50.

[0095] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0096] Heavy chain CDR1 of SEQ ID NO:7 (SYAMS);

[0097] Heavy chain CDR2 of SEQ ID NO: 8 (SISSGRSYIYYADSVKG);

[0098] The heavy chain CDR3 of SEQ ID NO:9 (EMQQLALDY);

[0099] Light chain CDR1 of SEQ ID NO: 10 (RASQGISNFLA);

[0100] The light chain CDR2 of SEQ ID NO: 11 (AASSLQS); and

[0101] Light chain CDR3 of SEQ ID NO: 12 (QQYNSYPFT).

[0102] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 39; and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 40. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 39 and a VL of SEQ ID NO: 40.

[0103] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 51; and a light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 52. In some embodiments, the ADC anti-FRα antibody comprises a heavy chain amino acid sequence of SEQ ID NO: 51 and a light chain amino acid sequence of SEQ ID NO: 52.

[0104] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0105] Heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN);

[0106] Heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS);

[0107] The heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS);

[0108] Light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA);

[0109] The light chain CDR2 of SEQ ID NO: 17 (KASSLES); and

[0110] Light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT).

[0111] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 41; and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 42. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 41 and a VL of SEQ ID NO: 42.

[0112] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 53; and a light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 54. In some embodiments, the ADC anti-FRα antibody comprises a heavy chain amino acid sequence of SEQ ID NO: 53 and a light chain amino acid sequence of SEQ ID NO: 54.

[0113] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0114] The heavy chain CDR1 of SEQ ID NO:19 (SYNMN);

[0115] Heavy chain CDR2 of SEQ ID NO:20 (SISSGSSYIYYADSMKG);

[0116] The heavy chain CDR3 of SEQ ID NO:21 (GMTTLTFDY);

[0117] Light chain CDR1 of SEQ ID NO:22 (RASQGISTFLA);

[0118] The light chain CDR2 of SEQ ID NO:23 (AASSLQS); and

[0119] Light chain CDR3 of SEQ ID NO:24 (QQYISYPLT).

[0120] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 43; and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 44. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 43 and a VL of SEQ ID NO: 44.

[0121] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 55; and a light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 56. In some embodiments, the ADC anti-FRα antibody comprises a heavy chain amino acid sequence of SEQ ID NO: 55 and a light chain amino acid sequence of SEQ ID NO: 56.

[0122] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0123] The heavy chain CDR1 of SEQ ID NO:25 (SYSMN);

[0124] Heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG);

[0125] The heavy chain CDR3 of SEQ ID NO:27 (GMTTLTFDY);

[0126] Light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA);

[0127] The light chain CDR2 of SEQ ID NO:29 (AASSLQS); and

[0128] Light chain CDR3 of SEQ ID NO: 30 (QQYNSYPLT).

[0129] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 45; and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 46. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 45 and a VL of SEQ ID NO: 46.

[0130] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 57; and a light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 58. In some embodiments, the ADC anti-FRα antibody comprises a heavy chain amino acid sequence of SEQ ID NO: 57 and a light chain amino acid sequence of SEQ ID NO: 58.

[0131] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises the following CDRs:

[0132] The heavy chain CDR1 of SEQ ID NO:31 (SDSATWN);

[0133] Heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS);

[0134] The heavy chain CDR3 of SEQ ID NO:33 (GGAPFDY);

[0135] Light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA);

[0136] The light chain CDR2 of SEQ ID NO:35 (KASSLES); and

[0137] Light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

[0138] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof has a VH comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 47; and a VL comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 48. In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises a VH of SEQ ID NO: 47 and a VL of SEQ ID NO: 48.

[0139] In some embodiments, the ADC anti-FRα antibody comprises a heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 59; and a light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 60. In some embodiments, the ADC anti-FRα antibody comprises a heavy chain amino acid sequence of SEQ ID NO: 59 and a light chain amino acid sequence of SEQ ID NO: 60.

[0140] Figure 5 The amino acid sequences of the heavy and light chain framework regions of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095 and AB1370117 are shown as SEQ ID NO: 61 to SEQ ID NO: 108.

[0141] In some embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof comprises: (a) a light chain VL-FR1, VL-FR2, VL-FR3, and VL-FR4 that is at least 80%, 85%, 90%, or 95% identical or identical to a reference light chain VL-FR1, VL-FR2, VL-FR3, and VL-FR4 of any of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095, or AB1370117, respectively. and (b) heavy chain VH-FR1, VH-FR2, VH-FR3 and VH-FR4 that are at least 80%, 85%, 90% or 95% identical or identical to a reference heavy chain VH-FR1, VH-FR2, VH-FR3 and VH-FR4 of any of constructs AB1370049, AB1370026, AB1370035, AB1370083, AB1370095 or AB1370117, respectively. Figure 5 shown.

[0142] Figure 6 The amino acid sequences of the heavy chain constant domain, the light chain constant domain and the heavy chain constant domain CH1 are listed as SEQ ID NO: 109 to SEQ ID NO: 111.

[0143] In some embodiments, the ADC anti-FRα antibody comprises a constant heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 109; and a constant light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 110. In some embodiments, the ADC anti-FRα antibody comprises a constant heavy chain amino acid sequence of SEQ ID NO: 109 and a constant light chain amino acid sequence of SEQ ID NO: 110.

[0144] In some embodiments, the ADC anti-FRα antigen-binding fragment comprises a constant heavy chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 111; and a constant light chain comprising an amino acid sequence that is at least 85% identical, at least 90% identical, at least 95% identical, at least 97% identical, at least 99% identical, or identical to the amino acid sequence of SEQ ID NO: 110. In some embodiments, the ADC anti-FRα antigen-binding fragment comprises a constant heavy chain amino acid sequence of SEQ ID NO: 111 and a constant light chain amino acid sequence of SEQ ID NO: 110.

[0145] The ADC antibody may include minor variations in amino acid sequence, provided that the variation in amino acid sequence retains at least 75%, more preferably at least 80%, at least 90%, at least 95%, and most preferably at least 99% sequence identity to the ADC antibody or antigen-binding fragment thereof as defined anywhere herein.

[0146] ADC antibodies may include variants in which amino acid residues from one class are substituted with corresponding residues in another class at conservative or non-conservative positions. In some embodiments, amino acid residues at non-conservative positions are substituted with conservative or non-conservative residues. In particular, conservative amino acid substitutions are contemplated.

[0147] "Conservative amino acid substitution" is an amino acid substitution in which an amino acid residue is replaced by an amino acid residue with a similar side chain. Families of amino acid residues with similar side chains have been defined in the art, including basic side chains (e.g., lysine, arginine, or histidine), acidic side chains (e.g., aspartic acid or glutamic acid), uncharged polar side chains (e.g., glycine, asparagine, glutamine, serine, threonine, tyrosine, or cysteine), non-polar side chains (e.g., alanine, valine, leucine, isoleucine, proline, phenylalanine, methionine, or tryptophan), β-branched side chains (e.g., threonine, valine, isoleucine), and aromatic side chains (e.g., tyrosine, phenylalanine, tryptophan, or histidine). Therefore, if an amino acid in a polypeptide is replaced by another amino acid from the same side chain family, the amino acid substitution is considered conservative. Conservatively modified variants in ADC antibodies do not exclude other forms of variants, such as polymorphic variants, cross-species homologs, and alleles.

[0148] "Non-conservative amino acid substitutions" include those in which (i) a residue having a positively charged side chain (e.g., Arg, His or Lys) is substituted for or is replaced by a negatively charged residue (e.g., Glu or Asp), (ii) a hydrophilic residue (e.g., Ser or Thr) is substituted for or is replaced by a hydrophobic residue (e.g., Ala, Leu, Ile, Phe or Val), (iii) cysteine ​​or proline is substituted for or is replaced by any other residue, or (iv) a residue having a bulky hydrophobic or aromatic side chain (e.g., Val, His, Ile or Trp) is substituted for or is replaced by a residue having a smaller side chain (e.g., Ala or Ser) or a residue without a side chain (e.g., Gly).

[0149] In addition to the twenty standard amino acids, non-standard amino acids (such as 4-hydroxyproline, 6-N-methyllysine, 2-aminoisobutyric acid, isovaline and α-methylserine) can also replace the amino acid residues of the ADC antibody. A limited number of non-conservative amino acids, amino acids not encoded by the genetic code and non-natural amino acids can replace amino acid residues. The ADC antibody may also include non-naturally occurring amino acid residues.

[0150] Non-naturally occurring amino acids include, but are not limited to, trans-3-methylproline, 2,4-methylene-proline, cis-4-hydroxyproline, trans-4-hydroxy-proline, N-methylglycine, allo-threonine, methyl-threonine, hydroxy-ethylcysteine, hydroxyethylhomo-cysteine, nitro-glutamine, homoglutamine, pipecolic acid, tert-leucine, norvaline, 2-azaphenylalanine, 3-azaphenylalanine, 4-azaphenyl-alanine and 4-fluorophenylalanine. Many methods are known in the art for incorporating non-naturally occurring amino acid residues into proteins. For example, an in vitro system can be used in which nonsense mutations are suppressed using chemical aminoacylation inhibitor tRNAs. Methods for synthesizing amino acids and aminoacylation of tRNAs are known in the art. Transcription and translation of plasmids containing nonsense mutations are performed in a cell-free system comprising an E. coli S30 extract and commercially available enzymes and other reagents. The protein is purified by chromatography. See, e.g., Robertson et al., J. Am. Chem. Soc. 113:2722, 1991; Ellman et al., Methods Enzymol. 202:301, 1991; Chung et al., Science 259:806-9, 1993; and Chung et al., Proc. Natl. Acad. Sci. USA 90:10145-9, 1993). In a second approach, translation is performed in Xenopus oocytes by microinjection of mutant mRNA and chemically aminoacylated inhibitor tRNA (Turcatti et al., J. Biol. Chem. 271:19991-8, 1996). In the third method, E. coli cells are cultured in the absence of the natural amino acid to be replaced (e.g., phenylalanine) and in the presence of the desired non-natural amino acid (e.g., 2-piperazinephenylalanine, 3-piperazinephenylalanine, 4-piperazinephenylalanine, or 4-fluorophenylalanine). Non-natural amino acids are incorporated into polypeptides in place of their natural counterparts. See Koide et al., Biochem. 33: 7470-6, 1994. Naturally occurring amino acid residues can be converted to non-naturally occurring species by in vitro chemical modification. Chemical modification can be combined with site-directed mutagenesis to further expand the range of substitutions (Wynn and Richards, Protein Sci. 2: 395-403, 1993).

[0151] A limited number of non-conservative amino acids, amino acids not encoded by the genetic code, non-naturally occurring amino acids, and unnatural amino acids can be substituted for amino acid residues of the ADC antibody.

[0152] Essential amino acids in ADC antibodies can be identified according to procedures known in the art, such as site-directed mutagenesis or alanine scanning mutagenesis (Cunningham and Wells, Science 244: 1081-5, 1989). Mutations of putative contact site amino acids can also be combined with physical analysis of the structure to determine the site of biological interaction by techniques such as nuclear magnetic resonance, crystallography, electron diffraction or photoaffinity labeling. See, for example, deVos et al., Science 255: 306-12, 1992; Smith et al., J. Mol. Biol. 224: 899-904, 1992; Wlodaver et al., FEBS Lett. 309: 59-64, 1992. The identity of essential amino acids can also be inferred from analysis of homology with related components of the ADC antibody (e.g., translocation or protease components).

[0153] Multiple amino acid substitutions can be prepared and tested using known mutagenesis and screening methods, such as those disclosed in Reidhaar-Olson and Sauer (Science 241:53-7, 1988) or Bowie and Sauer (Proc. Natl. Acad. Sci. USA 86:2152-6, 1989). These authors disclose methods for simultaneously randomizing two or more positions in a polypeptide, selecting functional polypeptides, and then sequencing the mutagenized polypeptides to determine the range of permissible substitutions at each position. Other methods that can be used include phage display (e.g., Lowman et al., Biochem. 30:10832-7, 1991; Ladner et al., U.S. Patent No. 5,223,409; Huse, WIPO Publication WO 92 / 06204) and region-directed mutagenesis (Derbyshire et al., Gene 46:145, 1986; Ner et al., DNA 7:127, 1988).

[0154] The "percentage of sequence identity" between two or more nucleic acid or amino acid sequences varies with the number of identical positions shared by the sequences. Thus, the percentage of sequence identity can be calculated as the number of identical nucleotides / amino acids divided by the total number of nucleotides / amino acids, multiplied by 100. The calculation of the percentage of sequence identity can also take into account the number of spaces and the length of each space that needs to be introduced to optimize the comparison of two or more sequences. Sequence comparisons and determination of percentage identity between two or more sequences can be performed using specific mathematical algorithms familiar to those skilled in the art, such as BLAST.

[0155] Any of a variety of sequence alignment methods can be used to determine the percentage of sequence identity, including but not limited to global methods, local methods, and hybrid methods, such as segment approximation methods. The scheme for determining the percentage of sequence identity is a routine procedure within the skill of those skilled in the art. The global method aligns the sequence from the beginning to the end of the molecule and determines the optimal alignment by adding the scores of individual residue pairs and applying gap penalties. Non-limiting methods include, for example, CLUSTAL W. See, for example, Julie D. Thompson et al., CLUSTAL W: Improving the Sensitivity of Progressive Multiple Sequence Alignment Through Sequence Weighting, Position-Specific Gap Penalties and Weight Matrix Choice, 22 (22) Nucleic Acids Research 4673-4680 (1994); and iterative refinement. See, for example, Osamu Gotoh, Significant Improvement in Accuracy of Multiple Protein Sequence Alignments by Iterative Refinement as Assessed by Reference to Structural Alignments, 264(4) J. MoI. Biol. 823-838 (1996). Local methods align sequences by identifying one or more conserved motifs shared by all input sequences.Non-limiting methods include, for example, the Match-box method, see, for example, Eric Depiereux and Ernest Feytmans, Match-Box: A Fundamentally New Algorithm for the Simultaneous Alignment of Several Protein Sequences, 8(5) CABIOS 501-509 (1992); Gibbs sampling, see, for example, CE Lawrence et al., Detecting Subtle Sequence Signals: A Gibbs Sampling Strategy for Multiple Alignment, 262(5131) Science 208-214 (1993); Align-M, see, for example, Ivo Van WaIIe et al., Align-M-A New Algorithm for Multiple Alignment of Highly Divergent Sequences, 20(9) Bioinformatics: 1428-1435 (2004).

[0156] The percent sequence identity can be determined by conventional methods. See, for example, Altschul et al., Bull. Math. Bio. 48:603-16, 1986; and Henikoff and Henikoff, Proc. Natl. Acad. Sci. USA 89:10915-19, 1992. Two amino acid sequences are aligned using a gap opening penalty of 10, a gap extension penalty of 1, and the "blosum62" scoring matrix of Henikoff and Henikoff (supra) to optimize the alignment score.

[0157] In some embodiments, the variable domains in the heavy and light chains of the ADC antibody or its antigen-binding fragment are changed by at least partially replacing one or more CDRs and / or by replacing and changing the sequence of a partial framework region. Although CDRs may be derived from antibodies of the same class or even subclass as the antibody from which the framework region is derived, it is envisioned that CDRs will be derived from antibodies of different classes and preferably from antibodies from different species in certain embodiments. It is not necessary to replace all CDRs with complete CDRs from the donor variable region to transfer the antigen binding ability of one variable domain to the antigen binding ability of another variable domain. On the contrary, only those residues necessary for maintaining the activity of the antigen binding site need to be transferred. In view of the explanations set forth in U.S. Patent Nos. 5,585,089, 5,693,761 and 5,693,762 (each of which is incorporated herein by reference), those skilled in the art are fully capable of performing routine experiments to obtain functional antibodies with reduced immunogenicity.

[0158] In some embodiments, the ADC antibody or antigen-binding fragment thereof may further comprise a heavy chain constant region or a fragment thereof in addition to VH and VL. In some embodiments, the heavy chain constant region is a human heavy chain constant region, such as a human IgG constant region or a human IgG1 constant region.

[0159] In some embodiments, residues are inserted into the heavy chain constant region for site-specific conjugation. For example, a cysteine ​​residue can be inserted between amino acids S239 and V240 in the CH2 region of IgG1, which can be referred to as a "239 insertion" or "239i".

[0160] In some embodiments, the ADC antibodies disclosed herein may be modified to include changes or modifications to one or more of the three heavy chain constant domains (CH1, CH2 or CH3) and / or to the light chain constant domain (CL). In some embodiments, a modified constant region in which one or more domains are partially or completely deleted is considered. In some embodiments, the modified ADC antibody will include a domain-deleted construct or variant in which the entire CH2 domain has been removed (ΔCH2 construct). In some embodiments, the omitted constant region domain may be replaced by a short amino acid spacer (e.g., 10 residues) that provides some molecular flexibility typically conferred by the lacking constant region. The deletion or inactivation of the constant region domain (by point mutation or other methods) can reduce the Fc receptor binding of the circulating modified antibody. In other cases, it may be that the constant region modification weakens complement binding and thus reduces the serum half-life and non-specific association of the conjugated ADC payload (e.g., cytotoxin). Other modifications of the constant region can be used to eliminate disulfide linkages or oligosaccharide moieties that allow enhanced localization due to increased antigen specificity or antibody flexibility. In some embodiments, the ADC antibody or antigen-binding fragment thereof does not have antibody-dependent cellular cytotoxicity (ADCC) activity and / or complement-dependent cytotoxicity (CDC) activity.

[0161] In some embodiments, ADC antibodies or antigen-binding fragments thereof may be engineered to fuse the CH3 domain directly to the hinge region of the corresponding modified antibody or fragment thereof. In other constructs, a peptide spacer may be inserted between the hinge region and the modified CH2 and / or CH3 domains. For example, a compatible construct may be expressed in which the CH2 domain is missing and the remaining CH3 domain (modified or unmodified) is connected to the hinge region using a spacer of 5-20 amino acids. Such a spacer may be added, for example, to ensure that the regulatory elements of the constant domain remain free and accessible or to ensure that the hinge region remains flexible. The amino acid spacer may be shown to be immunogenic in some cases and to induce an undesirable immune response against the construct. In some embodiments, any spacer added to the construct may be relatively non-immunogenic, or even completely omitted, in order to maintain the desired biochemical properties of the modified antibody.

[0162] In addition to deleting the entire constant region domain, the ADC antibody or its antigen-binding fragment provided herein can also be modified by partially deleting or replacing a few or even single amino acids in the constant region. For example, mutation of a single amino acid in a selected region of the CH2 domain can be sufficient to substantially reduce Fc binding and thereby increase tumor localization. Similarly, one or more constant region domains that control effector functions (e.g., complement C1Q binding) can be completely or partially deleted. Such partial deletions of the constant region can improve the selected features (e.g., serum half-life) of the ADC antibody or its antigen-binding fragment, but keep other desired functions related to the subject's constant region domain intact. In addition, the constant region of the ADC antibody and its antigen-binding fragment can be modified by mutation or substitution of one or more amino acids that enhance the properties of the resulting construct. In this regard, it is possible to destroy the activity provided by the conserved binding site (e.g., Fc binding) while substantially maintaining the configuration and immunogenicity characteristics of the modified antibody or its antigen-binding fragment. In some embodiments, one or more amino acids may be added to the constant region to enhance the desired features, such as reducing or increasing effector functions or providing more ADC payloads (e.g., cytotoxins) or carbohydrate attachment. In some embodiments, it may be desirable to insert or replicate specific sequences derived from selected constant region domains.

[0163] In some embodiments, the heavy chain constant region or fragment thereof (e.g., human IgG constant region or fragment thereof) may comprise one or more amino acid substitutions relative to a wild-type IgG constant domain, wherein the modified IgG has an increased half-life compared to the half-life of an IgG with a wild-type IgG constant domain. For example, the IgG constant domain may contain one or more amino acid substitutions of amino acid residues at positions 251-257, 285-290, 308-314, 385-389, and 428-436, wherein the amino acid position numbering is according to the EU index as shown in Kabat. In some embodiments, the IgG constant domain may contain one or more of the following: substitution of the amino acid at Kabat position 252 with tyrosine (Y), phenylalanine (F), tryptophan (W), or threonine (T), substitution of the amino acid at Kabat position 254 with threonine (T), substitution of the amino acid at Kabat position 256 with serine (S), arginine (R), glutamine (Q), glutamic acid (E), aspartic acid (D), or threonine (T), substitution of the amino acid at Kabat position 257 with leucine (L), substitution of the amino acid at Kabat position 258 with proline (P), substitution of the amino acid at Kabat position 259 with leucine (L), substitution of the amino acid at Kabat position 260 with leucine (L), substitution of the amino acid at Kabat position 261 with leucine (L), substitution of the amino acid at Kabat position 262 with leucine (L), substitution of the amino acid at Kabat position 263 with leucine (L), substitution of the amino acid at Kabat position 264 with leucine (L), substitution of the amino acid at Kabat position 265 with leucine (L), substitution of the amino acid at Kabat position 266 with leucine (L), substitution of the amino acid at Kabat position 267 with leucine (L), substitution of the amino acid at Kabat position 268 with leucine (L), substitution of the amino acid at Kabat position 269 with leucine (L), substitution of the amino acid at Kabat position 270 with leucine (L), substitution of the amino acid at Kabat position 271 with leucine (L), substitution of the amino acid at Kabat position 272 with leucine (L), substitution of the amino acid at Kabat position 273 with leucine (L), substitution of the amino acid at Kabat position 274 with The amino acid at Kabat position 309 is substituted with serine (S), the amino acid at Kabat position 311 is substituted with threonine (T), leucine (L), phenylalanine (F) or serine (S), the amino acid at Kabat position 428 is substituted with arginine (R), serine (S), isoleucine (I), proline (P) or glutamine (Q), or the amino acid at Kabat position 434 is substituted with tryptophan (W), methionine (M), serine (S), histidine (H), phenylalanine (F) or tyrosine.

[0164] In some embodiments, the ADC antibody or its antigen-binding fragment comprises a YTE mutant. The term "YTE" or "YTE mutant" refers to a mutation in IgG1 Fc that results in increased binding to human FcRn and improves the serum half-life of antibodies with the mutation. The YTE mutant comprises a combination of three mutations introduced into the heavy chain of IgG1, M252Y / S254T / T256E (EU numbering Kabat et al. (1991) Sequences of Proteins of Immunological Interest, US Public Health Service, National Institutes of Health, Washington, DC). See U.S. Patent No. 7,658,921, which is incorporated herein by reference. The YTE mutant has been shown to increase the serum half-life of the antibody approximately four-fold compared to the wild-type version of the same antibody (Dall'Acqua et al., J. Biol. Chem. 281: 23514-24 (2006); Robbie et al., (2013) Antimicrob. Agents Chemother. 57, 6147-6153). See also U.S. Pat. No. 7,083,784, which is incorporated herein by reference in its entirety.

[0165] In some embodiments, the ADC antibody or antigen-binding fragment thereof comprises:

[0166] L at the N-terminus (e.g., position 1) of VH;

[0167] E at the N-terminus (e.g., position 1) of VH;

[0168] Q at the N-terminus (eg, position 1) of VH.

[0169] B. Definition and Antibody Formats

[0170] As used herein, the term "antibody" refers to an immunoglobulin molecule that specifically binds to or immunoreacts with a particular antigen.

[0171] In general, antibodies comprise at least two "light chains" (LC) and two "heavy chains" (HC). The light and heavy chains of such antibodies are polypeptides comprising several domains. Each heavy chain comprises a heavy chain variable region (abbreviated herein as "VH") and a heavy chain constant region (abbreviated herein as "CH"). The heavy chain constant region comprises heavy chain constant domains CH1, CH2 and CH3 (antibody classes IgA, IgD and IgG) and optional heavy chain constant domains CH4 (antibody classes IgE and IgM). Each light chain comprises a light chain variable domain (abbreviated herein as "VL") and a light chain constant domain (abbreviated herein as "CL").

[0172] In some embodiments, the ADC antibody is a full-length antibody.As used herein, an "intact" or "full-length" antibody refers to an antibody having two heavy (H) chain polypeptides and two light (L) chain polypeptides interconnected by disulfide bonds.

[0173] The "variable region" of an antibody refers to the variable region of the antibody light chain or the variable region of the antibody heavy chain, alone or in combination. The variable regions VH and VL can be further subdivided into regions with high variability, called complementarity determining regions (CDRs) (also called hypervariable regions), interspersed with more conserved regions called framework regions (FRs). Preferably, each VH and VL is composed of three CDRs and four FRs, arranged in the following order from amino terminus to carboxyl terminus: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The VH or VL chain of an antibody may also contain all or part of the constant region of the heavy chain or light chain.

[0174] The binding between an antibody and its target antigen or epitope is mediated by CDR. The term "epitope" refers to a target protein region (e.g., a polypeptide) that can bind to (e.g., be bound by) an antibody or antigen-binding fragment. CDR is the main determinant of antigen specificity. There are at least two techniques for determining CDR: (1) a method based on cross-species sequence variability (i.e., Kabat et al. Sequences of Proteins of Immunological Interest, (5th edition, 1991, National Institutes of Health, Bethesda Md.)); and (2) a method based on crystallographic studies of antigen-antibody complexes (Al-lazikani et al. (1997) J. Molec. Biol. 273: 927-948)). In addition, a combination of these two methods is sometimes used in the art to determine CDR.

[0175] The sequence of a CDR may be identified by reference to any numbering system known in the art. Figure 7Three exemplary numbering systems are shown: the Kabat system (Kabat, EA et al., Sequences of Proteins of Immunological Interest, 5th ed., Public Health Service, National Institutes of Health, Bethesda, MD (1991); the Chothia system (Chothia and Lesk, “Canonical Structures for the Hypervariable Regions of Immunoglobulins,” J. Mol. Biol. 196, 901–917 (1987)); and the IMGT system (Lefranc et al., “IMGT Unique Numbering for Immunoglobulin and Cell Receptor Variable Domains and Ig superfamily V-like domains,” Dev. Comp. Immunol. 27, 55–77 (2003)).

[0176] The "constant domains" (or "constant regions") of the heavy and light chains are not directly involved in the binding of antibodies to targets, but rather exhibit various effector functions. The constant regions of antibodies mediate the binding of immunoglobulins to host tissues or factors, including different cells of the immune system (e.g., effector cells) and the first component (Clq) of the classical complement system.

[0177] There are five major classes of heavy chain constant regions, classified as IgA, IgG, IgD, IgE and IgM, each with a characteristic effector function specified by isotype. Ig molecules interact with various classes of cell receptors. For example, IgG molecules interact with three classes of Fcγ receptors (FcγR) specific for antibodies of IgG classes, i.e., FcγRI, FcγRII and FcγRIII. The binding of antibodies to Fc receptors on cell surfaces triggers many important and multiple biological responses, including the engulfment and destruction of antibody-coated particles, the removal of immune complexes, the dissolution of target cells coated with antibodies by killer cells (referred to as antibody-dependent cell-mediated cytotoxicity, or ADCC), the release of inflammatory mediators, placental transfer and the control of immunoglobulin production. It is reported that the important sequences for the binding of IgG to FcγR receptors are located in the CH2 and CH3 domains.

[0178] In a preferred embodiment, the ADC anti-FRα antibody or antigen-binding fragment thereof is an IgG isotype. The ADC anti-FRα antibody or antigen-binding fragment may be any IgG subclass, such as an IgG1, IgG2, IgG3, or IgG4 isotype. In a preferred embodiment, the ADC anti-FRα antibody or antigen-binding fragment thereof is based on an IgG1 isotype. Using wild-type human IgG1 molecules that are close to natural IgG can reduce developability and other risks.

[0179] For the heavy chain constant region amino acid positions of the ADC antibodies described herein, the numbering is according to the EU index first described in Edelman, GM et al., Proc. Natl. Acad. Sci. USA 63 (1969) 78-85). Edelman's EU numbering is also listed in Kabat et al. (1991) (above). Therefore, the terms "EU index as shown in Kabat", "EU index", "EU index of Kabat" or "EU numbering" in the context of the heavy chain refer to the residue numbering system of the human IgG1EU antibody based on Edelman et al. as described in Kabat et al. (1991). The numbering system for the amino acid sequence of the light chain constant region is similar to that listed in Kabat et al. (above). Therefore, as used herein, "numbering according to Kabat" refers to the Kabat numbering system listed in Kabat et al. (above).

[0180] The terms "Fc region", "Fc portion" and "Fc" are used interchangeably herein and refer to the portion of a natural immunoglobulin formed by two Fc chains. Each "Fc chain" comprises a constant domain CH2 and a constant domain CH3. Each Fc chain may also comprise a hinge region. The native Fc region is a homodimer. In some embodiments, the Fc region may be a heterodimer because it may contain modifications that force Fc heterodimerization. The Fc region contains a carbohydrate moiety and binding sites for complement and Fc receptors (including FcRn receptors) and does not have antigen binding activity. Fc may refer to this region in isolation or in the context of an antibody, antibody fragment or Fc fusion protein. Polymorphisms have been found in many Fc domain sites, including but not limited to EU positions 270, 272, 312, 315, 356 and 358, resulting in minor variations between the sequences described in this application and sequences known in the art. Therefore, each naturally occurring IgG Fc region is referred to as a "wild-type IgG Fc domain" or "WT IgG Fc domain" (i.e., any allele). Human IgG1, IgG2, IgG3, and IgG4 heavy chain sequences are available in a variety of sequence databases, including the UniProt database (www.uniprot.org), under accession numbers P01857 (IGHG1_human), P01859 (IGHG2_human), P01860 (IGHG3_human), and P01861 (IGHG4_human), respectively.

[0181] In some embodiments, the ADC anti-FRα antibody is a monoclonal antibody. "Monoclonal antibody" (mAb) refers to a homogeneous antibody population that participates in highly specific recognition and binding to a single antigenic determinant or epitope. This is in contrast to polyclonal antibodies that typically include different antibodies against different antigenic determinants. The term "monoclonal antibody" can encompass full-length monoclonal antibodies as well as antibody fragments (such as Fab, Fab', F(ab')2, Fv), single-chain (scFv) mutants, fusion proteins comprising antibody portions, and any other modified immunoglobulin molecules comprising antigen recognition sites. In addition, "monoclonal antibody" refers to such antibodies made in any number of ways, including, but not limited to, hybridomas, phage selection, recombinant expression, and transgenic animals. More preferably, the ADC anti-FRα antibody is an isolated monoclonal antibody. In a more preferred embodiment, the ADC antibody is a fully human monoclonal antibody.

[0182] ADC anti-FRα antibodies and antigen-binding fragments thereof can be obtained from any species by recombinant means. For example, the ADC antibody or antigen-binding fragment can be a mouse, rat, goat, horse, pig, cow, chicken, rabbit, camel, donkey, human or a chimeric version thereof. For administration to humans, ADC antibodies or antigen-binding fragments of non-human origin can be genetically or structurally altered to reduce immunogenicity when administered to human patients. Particularly preferred are human or humanized antibodies, especially as recombinant human or humanized antibodies.

[0183] The term "human antibody" refers to an antibody produced in a human or an antibody having an amino acid sequence corresponding to an antibody produced in a human, made using any technology known in the art. Human antibodies may include complete or full-length antibodies, fragments thereof and / or antibodies comprising at least one human heavy chain and / or light chain polypeptide, such as antibodies comprising mouse light chain and human heavy chain polypeptide. Human antibodies may include amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or during gene rearrangement or by somatic mutations in vivo). Human antibodies may be made in human cells (by recombinant expression), non-human animals or prokaryotic or eukaryotic cells capable of expressing functionally rearranged human immunoglobulin (such as heavy and light chains) genes. Single-chain human antibodies may include a linker peptide not found in natural human antibodies. For example, Fv may have a linker peptide connecting the variable region of the heavy chain and the variable region of the light chain, such as 2 to about 8 glycine or other amino acid residues. These linker peptides are considered to have human origin. Human antibodies can be produced using a variety of techniques, including phage display technology using an antibody library derived from human immunoglobulin sequences. Transgenic mice that cannot express functional inherent immunoglobulins but can express human immunoglobulin genes can also be used to prepare human antibodies (see, e.g., PCT Publication Nos. WO 1998 / 24893; WO 1992 / 01047; WO 1996 / 34096; WO 1996 / 33735; U.S. Pat. Nos. 5,413,923; 5,625,126; 5,633,425; 5,569,825; 5,661,016; 5,545,806; 5,814,318; 5,885,793; 5,916,771; and 5,939,598, each of which is incorporated herein by reference). Human antibodies can also be prepared directly using various techniques known in the art. Immortalized human B lymphocytes that produce antibodies to a target antigen that are immunized in vitro or isolated from an immunized individual can be produced. See, e.g., Cole et al., Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, p. 77 (1985); Boemer et al., J. Immunol. 147(1):86-95 (1991); U.S. Pat. No. 5,750,373.

[0184] The term "humanized antibody" refers to an antibody in which the framework or CDR has been modified to contain a CDR of an immunoglobulin with a different specificity than that of the parent immunoglobulin. For example, a mouse CDR can be transplanted into the framework region of a human antibody to prepare a "humanized antibody". See, for example, Riechmann, L. et al., Nature 332 (1988) 323-327; and Neuberger, MS et al., Nature 314 (1985) 268-270. In some embodiments, a "humanized antibody" is an antibody in which the constant region has been additionally modified or altered from that of the original antibody to produce desired properties.

[0185] Humanized antibodies can be prepared optionally by analyzing parental sequences and various conceptual humanization and engineered products using three-dimensional models of parental sequences, engineered sequences and humanized sequences. Three-dimensional immunoglobulin models are generally available and are familiar to those skilled in the art. Computer programs that illustrate and display the possible three-dimensional conformational structures of selected candidate immunoglobulin sequences are available. Inspection of these displays allows analysis of the possible effects of residues in the function of candidate immunoglobulin sequences, i.e., analysis of residues that affect the ability of candidate immunoglobulins to bind their antigens (such as FRα). In this way, folate receptor (FR) residues can be selected and combined with consensus sequences and input sequences to achieve desired antibody characteristics, such as increased affinity to one or more target antigens.

[0186] Humanized antibodies can be further modified to improve and optimize antibody specificity, affinity and / or ability by replacing the extra residues in the Fv framework region and / or the non-human residues replaced.In general, humanized antibodies will substantially comprise at least one and usually two or three variable domains, the variable domains containing all or substantially all CDR regions corresponding to non-human immunoglobulins, and all or substantially all FR regions are those of human immunoglobulin consensus sequences.Humanized antibodies may also comprise at least a portion (Fc) of an immunoglobulin constant region or domain, typically a portion of a human immunoglobulin.Examples of methods for producing humanized antibodies are described in U.S. Patent numbers 5,225,539 and 5,639,641, each of which is incorporated herein by reference.

[0187] The term "chimeric antibody" refers to a monoclonal antibody comprising a variable region (i.e., a binding region) from one source or species and at least a portion of a constant region derived from a different source or species, usually prepared by recombinant DNA technology. Chimeric antibodies comprising murine variable regions and human constant regions are preferred. Other preferred forms of "chimeric antibodies" are antibodies in which the constant region has been modified or changed from the constant region of the original antibody to produce the desired properties. Such chimeric antibodies are also referred to as "class-shifted antibodies." Chimeric antibodies are the product of expressed immunoglobulin genes, which contain DNA segments encoding immunoglobulin variable regions and DNA segments encoding immunoglobulin constant regions. Methods for producing chimeric antibodies involving conventional recombinant DNA and gene transfection techniques are well known in the art. See, for example, Morrison, SL et al., Proc. Natl. Acad. Sci. USA 81 (1984) 6851-6855; U.S. Pat. Nos. 5,202,238 and 5,204,244, each of which is incorporated herein by reference.

[0188] In some embodiments, the ADC antibody is a full-length antibody as described above. Alternatively, the ADC antibody may be an antigen-binding fragment. As used herein, the term "antigen-binding fragment" includes any naturally occurring or artificially constructed antigen-binding polypeptide configuration comprising one, two or three light chain CDRs, and / or one, two or three heavy chain CDRs, wherein the polypeptide is capable of binding to an antigen.

[0189] In some embodiments, the ADC antigen-binding fragment is a Fab fragment. The ADC antibody can also be Fab', Fv, scFv, Fd, VNAR domain, IgNAR, intracellular antibody, IgGCH2, mini antibody, single domain antibody, Fcab, scFv-Fc, F(ab')2, di-scFv, bispecific T cell engager F(ab')3, tetrabody, triabody, diabody, DVD-Ig, (scFv)2, mAb2 or DARPin.

[0190] The terms "Fab fragment" and "Fab" are used interchangeably herein and contain a single light chain (e.g., constant domains CL and VL) and a single heavy chain (e.g., constant domains CH1 and VH). The heavy chain of the Fab fragment cannot form a disulfide bond with another heavy chain.

[0191] "Fab' fragments" contain a single light chain and a single heavy chain, but in addition to CH1 and VH, "Fab' fragments" also contain the heavy chain region between the CH1 and CH2 domains, which is necessary for forming interchain disulfide bonds. Therefore, two "Fab' fragments" can associate by forming a disulfide bond to form a F(ab')2 molecule.

[0192] "F(ab')2 fragment" contains two light chains and two heavy chains. Each chain contains a portion of the constant region necessary for forming interchain disulfide bonds between the two heavy chains.

[0193] An "Fv fragment" contains only the variable regions of the heavy and light chains. It does not contain the constant regions.

[0194] "Single-domain antibodies" are antibody fragments that contain a single antibody domain unit (eg, VH or VL).

[0195] "Single-chain Fv" ("scFv") is an antibody fragment containing the VH and VL domains of an antibody linked together to form a single chain. A polypeptide linker is often used to link the VH and VL domains of a scFv.

[0196] "Tandem scFv" (also known as ) is a single-chain Fv molecule formed by two scFvs covalently bonded in a tandem orientation via a flexible peptide linker.

[0197] “Bispecific T cell engager” It is a fusion protein containing two single-chain variable fragments (scFv) on a single peptide chain. One of the scFv binds to T cells via the CD3 receptor and the other binds to tumor cell antigens.

[0198] "Diabodies" are small bivalent and bispecific antibody fragments comprising a heavy chain variable domain (VH) connected to a light chain variable domain (VL) on the same polypeptide chain (VH-VL) connected by a peptide linker that is too short to allow pairing between the two domains on the same chain (Kipriyanov, Int. J. Cancer 77 (1998), 763-772). This forces pairing with the complementary domain of the other chain and promotes the assembly of a dimeric molecule with two functional antigen binding sites.

[0199] "DARPin" is a bispecific ankyrin repeat molecule. DARPin is derived from natural ankyrin, which can be found in the human genome and is one of the most abundant types of binding proteins. The DARPin library module is defined by the natural ankyrin repeat protein sequence, using 229 ankyrin repeat sequences for initial design and another 2200 for subsequent refinement. The module serves as the building block of the DARPin library. The library module is similar to the human genome sequence. DARPin consists of 4 to 6 modules. Because each module is about 3.5 kDa, the size of the average DARPin is 16-21 kDa. The selection of binders is completed by ribosome display, which is completely cell-free and described in He M and Taussig MJ., Biochem Soc Trans. 2007, November; 35 (Pt 5): 962-5.

[0200] In some embodiments, the ADC antibody or its antigen-binding fragment can be further modified to include additional chemical parts that are not normally part of the protein. Those derivative parts can improve the solubility, biological half-life or absorption of the protein. The parts can also reduce or eliminate any desired side effects of the protein, etc. An overview of those parts can be found in Remington's Pharmaceutical Sciences, 22nd edition, Ed. Lloyd V. Allen, Jr. (2012).

[0201] C. FRα binding

[0202] In preferred embodiments, the ADC anti-FRα antibody or its antigen-binding fragment specifically binds to FRα. The term "specifically binds to FRα" refers to an antibody that is capable of binding to a defined target with sufficient affinity so that the antibody can be used as a therapeutic agent targeting FRα. In some embodiments, the ADC antibody that specifically binds to FRα does not bind to other antigens, or does not bind to other antigens with sufficient affinity to produce a physiological effect. In some embodiments, the ADC anti-FRα antibody or its antigen-binding fragment specifically binds to human FRα (UniProt ID: P15328) and / or cynomolgus monkey FRα (UniProt ID: A0A2K5U044). In particularly preferred embodiments, the ADC anti-FRα antibody or its antigen-binding fragment specifically binds to human FRα. In preferred embodiments, the ADC anti-FRα antibody or its antigen-binding fragment specifically binds to human FRα and cynomolgus monkey FRα.

[0203] In some embodiments, the FRα protein has a sequence of SEQ ID NO: 112 or SEQ ID NO: 113, such as Figure 8In one embodiment, FRα has the sequence of SEQ ID NO: 112 and is a human FRα protein (predicted mature secreted polypeptide). In another embodiment, FRα has the sequence of SEQ ID NO: 113 and is a cynomolgus monkey FRα protein (predicted mature secreted polypeptide).

[0204] In some embodiments, the ADC antibody or antigen-binding fragment thereof does not bind to one or more selected from mouse FRα (UniProt ID: P35846), rat FRα (UniProt ID: G3V8M6), human FRβ (UniProt ID: P14207), human FRγ (UniProt ID: P41439), or a combination thereof.

[0205] The term "does not bind" means that the ADC antibody or its antigen-binding fragment does not substantially bind to one or more of the molecules (e.g., mouse FRα, rat FRα, human FRβ, human FRγ, or a combination thereof). The term "substantially not" when used herein in the context of binding may mean that less than 5%, 2%, 1%, 0.5% or 0.1% of cells expressing one or more of the molecules in cell culture become bound by the ADC antibody or its antigen-binding fragment (when in contact with it). Suitably, the term "substantially not" when used herein in the context of binding may mean that no such cells become bound.

[0206] D. Antibody Preparation

[0207] ADC antibodies or antigen-binding fragments thereof (e.g., as monoclonal antibodies) can be prepared using recombinant DNA methods, as described in U.S. Pat. No. 4,816,567, which is incorporated herein by reference. Such as by RT-PCR using oligonucleotide primers that specifically amplify the genes encoding the antibody heavy and light chains, the polynucleotides encoding the monoclonal antibodies are isolated from mature B cells or hybridoma cells, and their sequences are determined using conventional procedures. The isolated polynucleotides encoding the heavy and light chains are then cloned into suitable expression vectors, which enter host cells such as E. coli cells, simian COS cells, Chinese hamster ovary (CHO) cells or myeloma cells that do not otherwise produce immunoglobulin proteins when transfected, and monoclonal antibodies are produced by these host cells. In addition, recombinant monoclonal antibodies or antigen-binding fragments thereof of the desired species can be isolated from phage display libraries expressing CDRs of the desired species as described in McCafferty et al., Nature, 348:552-554 (1990); Clackson et al., Nature, 352:624-628 (1991); and Marks et al., J. Mol. Biol. 222:581-597 (1991).

[0208] Affinity maturation strategies and chain shuffling strategies are known in the art and can be used to generate high affinity human antibodies or antigen-binding fragments thereof. See Marks et al., BioTechnology 10:779-783 (1992), which is incorporated by reference in its entirety.

[0209] Different techniques are known for producing antibody fragments. Traditionally, these fragments are obtained by proteolytic digestion of intact antibodies, for example, as described by Morimoto et al., J. Biochem. Biophys. Meth. 24: 107-117 (1993) and Brennan et al., Science 229: 81 (1985). In some embodiments, ADC antibody fragments are recombinantly produced. Fab, Fv and scFv antibody fragments can all be expressed in Escherichia coli or other host cells and secreted therefrom, thereby allowing the production of large amounts of these fragments. Such ADC antibody fragments can also be separated from the antibody phage library discussed above. ADC antibody fragments can also be linear antibodies, as described in U.S. Patent No. 5,641,870, which is incorporated herein by reference. Other techniques for producing antibody fragments will be apparent to skilled practitioners.

[0210] The technology can also be adapted to produce single-chain antibodies specific for FRα. See, for example, U.S. Patent No. 4,946,778. In addition, the method can be adapted to construct a Fab expression library to allow rapid and efficient identification of monoclonal Fab fragments or derivatives, fragments, analogs or homologs thereof with the desired specificity for FRα. See, for example, Huse et al., Science 246: 1275-1281 (1989). Antibody fragments can be produced by techniques known in the art, including but not limited to: F(ab')2 fragments produced by pepsin digestion of antibody molecules; Fab fragments produced by reducing the disulfide bridge bonds of F(ab')2 fragments; Fab fragments produced by treating antibody molecules with papain and a reducing agent; or Fv fragments.

[0211] III. ADC connector

[0212] The ADC antibody or antigen-binding fragment thereof can be conjugated to an ADC payload (eg, a cytotoxic agent or cytotoxin) via a linker.

[0213] As used herein, the term "linker" or "spacer" means a divalent chemical part comprising a covalent bond or a chain of atoms, which covalently connects an ADC antibody or its antigen-binding fragment to an ADC payload (e.g., a cytotoxin) to form an ADC. In some embodiments, the linker or spacer is a peptide spacer. In some embodiments, the linker or spacer is a non-peptide (e.g., chemical) spacer. A suitable linker has two reactive ends, one for antibody conjugation and the other for ADC payload (e.g., cytotoxin) conjugation. Due to the formation of a bond between a linker and / or ADC payload (e.g., cytotoxin) and between a linker and / or ADC antibody or its antigen-binding fragment, one or both of the reactive ends will be missing or incomplete (e.g., only the carbonyl of a carboxylic acid). These conjugation reactions will be discussed in detail below.

[0214] In preferred embodiments, the linker is cleavably attached (eg, conjugated) to an amino residue, such as an amino acid, of an ADC antibody or antigen-binding fragment described herein.

[0215] In some embodiments, the linker is cleavable under intracellular conditions, thereby allowing the drug unit to be released from the ADC antibody in the intracellular environment.

[0216] Alternatively, the linker unit may not be cleavable. In such embodiments, the drug is released, for example, by antibody degradation. However, a non-cleavable payload requires complete mAb digestion in the lysosome, and the resulting drug-containing product may be too polar, for example, to achieve a bystander effect.

[0217] ADC is preferably stable and intact before transport or delivery into the cell, i.e. the antibody should be linked to the drug moiety. Outside the target cell, the linker is stable, but inside the cell, they can be cleaved at a higher rate. Effective linkers will: (i) maintain the specific binding properties of the antibody; (ii) allow the conjugate or drug moiety to be delivered inside the cell; (iii) remain stable and intact, i.e. not cleaved, until the conjugate is delivered or transported to its target site; and (iv) maintain the cell killing or cell inhibition of the cytotoxic moiety. Standard analytical techniques such as mass spectrometry, HPLC, and separation / analysis techniques LC / MS can be used to assess the stability of the ADC.

[0218] The linker can be cleaved, for example, by enzymatic hydrolysis, photolysis, hydrolysis under acidic conditions, hydrolysis under basic conditions, oxidation, disulfide reduction, nucleophilic cleavage, or organometallic cleavage (see, e.g., Leriche et al., Bioorg. Med. Chem., 20:571-582, 2012).

[0219] Linkers that can be hydrolyzed under acidic conditions include, for example, hydrazones, semicarbazones, thiosemicarbazones, cis-aconitamides, orthoesters, acetals, ketals, etc. (See, for example, U.S. Pat. Nos. 5,122,368, 5,824,805, 5,622,929; Dubowchik and Walker, 1999, Pharm. Therapeutics 83:67-123; Neville et al., 1989, Biol. Chem. 264:14653-14661). Linkers that can be cleaved under reducing conditions include, for example, disulfides. A variety of disulfide linkers are known in the art, including, for example, those that can be formed using SATA (N-succinimidyl-S-acetylthioacetate), SPDP (N-succinimidyl-3-(2-pyridyldithio) propionate), SPDB (N-succinimidyl-3-(2-pyridyldithio) butyrate), and SMPT (N-succinimidyl-oxycarbonyl-α-methyl-α-(2-pyridyl-dithio) toluene) (see, e.g., Thorpe et al., 1987, Cancer Res. 47:5924-5931; Wawrzynczak et al., In Immunoconjugates: Antibody Conjugates in Radioimaging and Therapy of Cancer (CW Vogel, ed., Oxford U. Press, 1987)).

[0220] In a preferred embodiment, the joint is susceptible to enzymatic hydrolysis. Relative to pH-sensitive cleavable joints, such joints are preferred because pH-sensitive cleavable joints may not be stable enough and are cleaved prematurely before reaching the target cell, so potential off-target toxicity may be observed. Enzymatic cleavable joints can be, for example, peptide-containing joints cleaved by intracellular peptidases or proteases (including but not limited to lysosomal or endosomal proteases). One benefit of using intracellular proteolytic release of therapeutic drugs is that the agent is usually weakened when conjugated, and the serum stability of the conjugate is usually high. In some embodiments, the peptidyl joint is at least two amino acids long or at least three amino acids long. Exemplary amino acid joints include dipeptides, tripeptides, tetrapeptides or pentapeptides. Peptides comprising amino acids valine, alanine, citrulline (Cit), phenylalanine, lysine, leucine and glycine are examples of suitable peptides. Natural amino acids, minor amino acids and non-naturally occurring amino acid analogs such as citrulline are all examples of amino acid residues constituting amino acid joint components. Exemplary dipeptides include valine-citrulline (VC or Val-Cit), alanine-phenylalanine (AF or Ala-Phe). Exemplary tripeptides include glycine-valine-citrulline (Gly-Val-Cit) and glycine-glycine-glycine (Gly-Gly-Gly). In some embodiments, the linker includes a dipeptide such as Val-Cit, Ala-Val or Phe-Lys, Val-Lys, Ala-Lys, Phe-Cit, Leu-Cit, Ile-Cit, Phe-Arg or Trp-Cit.

[0221] In some embodiments, the linker comprises PEG. A stable protease-cleavable linker containing PEG can limit the hydrophobicity of the payload and can selectively cleave and release free drugs in target cancer cells. The lower hydrophobicity of the linker as described herein can allow the drug to be highly loaded on the antibody or antigen-binding fragment (e.g., DAR8) without aggregation, which will be significantly higher than Mituximab Soraftansine (DAR3-4) or its derivatives, such as IMGN151 (DAR3.5). This can enable ADC to deliver significantly higher concentrations of ADC payload (e.g., cytotoxin) to target cancer cells by binding to FRα on cancer cells.

[0222] In some embodiments, the linker comprises maleimide. The use of maleimide in the linker can generate DAR8 and DAR4 ADCs by utilizing the natural interchain disulfides in antibodies. This is superior to the conjugation of surface amines from lysine residues, which can lead to mixtures of DAR substances and batch-to-batch variability. If the conjugation site interferes with antigen binding, there may also be reproducibility issues that affect the efficacy of the ADC. In addition, other conjugation methods, such as azide-alkyne click chemistry involving engineered antibodies, may not easily achieve DARs exceeding 4.

[0223] In certain embodiments, the ADC anti-FRα antibody or antigen-binding fragment thereof is connected via a linker R L is linked to an ADC payload (e.g., a cytotoxin). In some embodiments, R L Selected from:

[0224] [Formula 1]

[0225]

[0226] Where Q is:

[0227] [Formula 2]

[0228]

[0229] Where Q X is such that Q is an amino acid residue, a dipeptide residue, a tripeptide residue or a tetrapeptide residue, and X in Formula 1 is:

[0230] [Formula 3]

[0231]

[0232] Where a=0 to 5, b1=0 to 16, b2=0 to 16, c1=0 or 1, c2=0 or 1, d=0 to 5, wherein at least b1 or b2=0 (i.e., only one of b1 and b2 can be non-0) and at least c1 or c2=0 (i.e., only one of c1 and c2 can be non-0). G in Formula 1 L is a linker for connecting to an ADC antibody or an antigen-binding fragment thereof. Alternatively, G in Formula 1 L is a compound denoted as "Ib":

[0233] [Formula 4]

[0234]

[0235] Where R L1 and R L2 are independently selected from H and methyl, or together with the carbon atoms to which they are attached form cyclopropylene or cyclobutylene; and e is 0 or 1; or a compound represented by "Ib'":

[0236] [Formula 5]

[0237]

[0238] Where R L1 and R L2 As defined above.

[0239] For example, we will outline G L , X, QX (e.g., within the linker of Ia of Formula 1) and preferred embodiments of the linker of Ib of Formula 4.

[0240] The following preferences may apply to all embodiments, or may relate to a single embodiment. The preferences may be combined together in any combination.

[0241] Various definitions related to certain terms in this section are provided under the heading "Chemical Definitions" provided below.

[0242] G of the linker of formula 1 L You can choose from:

[0243]

[0244]

[0245] Where Ar represents C 5-6 Arylene, such as phenylene, and X represents C 1-4 alkyl.

[0246] In some embodiments, G L Selected from G L1-1 and G L1-2 In some of these embodiments, G L It's G L1-1 .

[0247] X in Formula 1 is preferably:

[0248] [Formula 6]

[0249]

[0250] Wherein a=0 to 5, b1=0 to 16, b2=0 to 16, c=0 or 1, d=0 to 5, wherein at least b1 or b2=0 and at least c1 or c2=0.

[0251] "a" can be 0, 1, 2, 3, 4, or 5. In some embodiments, a is 0 to 3. In some of these embodiments, a is 0 or 1. In other embodiments, a is 0.

[0252] "b1" can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b1 is 0 to 12. In some of these embodiments, b1 is 0 to 8, and can be 0, 2, 3, 4, 5, or 8.

[0253] "b2" can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 16. In some embodiments, b2 is 0 to 12. In some of these embodiments, b2 is 0 to 8, and can be 0, 2, 3, 4, 5 or 8. Preferably, only one of b1 and b2 can be non-zero.

[0254] "c1" may be 0 or 1. c2 may be 0 or 1. Preferably, only one of c1 and c2 may not be 0.

[0255] "d" can be 0, 1, 2, 3, 4, or 5. In some embodiments, d is 0 to 3. In some of these embodiments, d is 1 or 2. In other embodiments, d is 2. In other embodiments, d is 5.

[0256] In some embodiments of X, a is 0, b1 is 0, c1 is 1, c2 is 0 and d is 2, and b2 can be 0 to 8. In some of these embodiments, b2 is 0, 2, 3, 4, 5, or 8. In some embodiments of X, a is 1, b2 is 0, c1 is 0, c2 is 0 and d is 0, and b1 can be 0 to 8. In some of these embodiments, b1 is 0, 2, 3, 4, 5, or 8. In some embodiments of X, a is 0, b1 is 0, c1 is 0, c2 is 0 and d is 1, and b2 can be 0 to 8. In some of these embodiments, b2 is 0, 2, 3, 4, 5, or 8. In some embodiments of X, b1 is 0, b2 is 0, c1 is 0, c2 is 0, and one of a and d is 0. The other of a and d is 1 to 5. In some of these embodiments, the other of a and d is 1. In other of these embodiments, the other of a and d is 5. In some embodiments of X, a is 1, b2 is 0, c1 is 0, c2 is 1, d is 2, and b1 can be 0 to 8. In some of these embodiments, b2 is 0, 2, 3, 4, 5, or 8.

[0257] In some embodiments, Q of Formula 1 is an amino acid residue. The amino acid can be a natural amino acid or a non-natural amino acid. For example, Q can be selected from: Phe, Lys, Val, Ala, Cit, Leu, Ile, Arg and Trp, wherein Cit is citrulline.

[0258] In some embodiments, Q comprises a dipeptide residue. The amino acids in the dipeptide can be any combination of natural amino acids and non-natural amino acids. In some embodiments, the dipeptide comprises natural amino acids. When the linker is a cathepsin-labile linker, the dipeptide is the action site of cathepsin-mediated cleavage. The dipeptide is then a recognition site for cathepsin.

[0259] In some embodiments, Q is selected from:

[0260] NH -Phe-Lys- C＝O ,

[0261] NH -Val-Ala- C＝O ,

[0262] NH -Val-Lys- C＝O ,

[0263] NH -Ala-Lys- C＝O ,

[0264] NH -Val-Cit- C＝O ,

[0265] NH -Phe-Cit- C＝O ,

[0266] NH -Leu-Cit- C＝O ,

[0267] NH -Ile-Cit- C＝O ,

[0268] NH -Phe-Arg- C＝O ,

[0269] NH -Trp-Cit- C＝O and

[0270] NH -Gly-Val- C＝O ;

[0271] Wherein Cit is citrulline. Preferably, Q is selected from:

[0272] NH -Phe-Lys- C＝O ,

[0273] NH -Val-Ala- C＝O,

[0274] NH -Val-Lys- C＝O ,

[0275] NH -Ala-Lys- C＝O and

[0276] NH -Val-Cit- C＝O .

[0277] More preferably, Q is selected from NH -Phe-Lys- C＝O , NH -Val-Cit- C＝O or NH -Val-Ala- C＝O Other suitable dipeptide combinations include:

[0278] NH -Gly-Gly- C＝O ,

[0279] NH -Gly-Val- C＝O ,

[0280] NH -Pro-Pro- C＝O and

[0281] NH -Val-Glu- C＝O .

[0282] Other dipeptide combinations may be used, including those described by Dubowchik et al., Bioconjugate Chemistry, 2002, 13, 855-869, which is incorporated herein by reference.

[0283] In some embodiments, Q is a tripeptide residue. The amino acids in the tripeptide can be any combination of natural amino acids and non-natural amino acids. In some embodiments, the tripeptide comprises natural amino acids. When the linker is a cathepsin-labile linker, the tripeptide is the site of action for cathepsin-mediated cleavage. The tripeptide is then a recognition site for cathepsin. Particularly interesting tripeptide linkers are:

[0284] NH -Glu-Val-Ala- C＝O ,

[0285] NH -Glu-Val-Cit- C＝O ,

[0286] NH -αGlu-Val-Ala- C＝O and

[0287] NH -αGlu-Val-Cit- C＝O .

[0288] In some embodiments, Q is a tetrapeptide residue. The amino acids in the tetrapeptide can be any combination of natural amino acids and non-natural amino acids. In some embodiments, the tetrapeptide comprises natural amino acids. When the linker is a cathepsin labile linker, the tetrapeptide is the site of action for cathepsin-mediated cleavage. The tetrapeptide is then a recognition site for cathepsin. Particularly interesting tetrapeptide linkers are:

[0289] NH -Gly-Gly-Phe-Gly C＝O ;and

[0290] NH -Gly-Phe-Gly-Gly C＝O .

[0291] In some embodiments, the tetrapeptide is:

[0292] NH -Gly-Gly-Phe-Gly C＝O .

[0293] In the above representation of peptide residues, NH - represents the N-terminus of the residue, and - C＝O Represents the C-terminus of the residue. The C-terminus is bound to the NH of the "drug unit" (e.g., A* in Formula 14 below). Glu represents the residue of glutamic acid, that is:

[0294] [Formula 7]

[0295]

[0296] αGlu represents the residue when glutamate is bound via the α chain, that is:

[0297] [Formula 8]

[0298]

[0299] In some embodiments, the amino acid side chains are chemically protected when appropriate. The side chain protecting group can be the group discussed above. The protected amino acid sequence can be cleaved by enzymes. For example, a dipeptide sequence containing a Lys residue protected by a Boc side chain can be cleaved by histones.

[0300] Protecting groups for amino acid side chains are well known in the art and are described in the Novabiochem catalogue and as described above.

[0301] R of the linker Ib shown in Formula 4 L1 and R L2 They may be independently selected from H and methyl, or together with the carbon atom to which they are attached form cyclopropylene or cyclobutylene.

[0302] In some embodiments, R L1 and R L2 Both are H. In some embodiments, R L1 is H and R L2 In some embodiments, R L1 and R L2 Both are methyl groups.

[0303] In some embodiments, R L1 and R L2 Together with the carbon atom to which they are attached, they form a cyclopropylene group. L1 and R L2 Together with the carbon atom to which they are attached they form a cyclobutylene group.

[0304] In the group Ib of formula 4, in some embodiments, e is 0. In other embodiments, e is 1, and the nitro group can be located in any available position of the ring. In some embodiments of these embodiments, it is in the ortho position. In other embodiments of these embodiments, it is in the para position.

[0305] In some embodiments, the linker R shown in Formula 1 L Selected from:

[0306]

[0307]

[0308] Preferably, R L is: [Formula 9]

[0309]

[0310] For example, an ADC may have the following general formula:

[0311] [Formula 10]

[0312] L–(D L ) p

[0313] or a pharmaceutically acceptable salt or solvate thereof, wherein L is an ADC antibody or an antigen-binding fragment thereof, and D Lis a linker R connected to an ADC antibody or an antigen-binding fragment thereof LL In some embodiments, the linker R LL Preferably selected from "Ia'":

[0314] [Formula 11]

[0315]

[0316] Where Q and X are as defined above, and G LL It is a linker connected to the ADC antibody or its antigen-binding fragment. Ib' is represented by the following formula:

[0317] [Formula 12]

[0318]

[0319] Where R L1 and R L2 As defined above.

[0320] Antibody-drug conjugate of formula 10: L-(D L ) p The "p" in represents drug loading, and is the number of "drug units" (e.g., cytotoxins such as TOPOi) per antibody or antigen-binding fragment thereof. The drug loading per antibody or antigen-binding fragment thereof is in the range of 1 to 20 drug units (D). For a composition, "p" represents the average drug loading of the conjugate in the composition, and p is in the range of 1 to 20. In some embodiments, the range of "p" is selected from 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10; preferably p is 8.

[0321] G of formula 11 LL You can choose from:

[0322]

[0323]

[0324] Where Ar represents C 5-6 Arylene, such as phenylene, and X represents C 1-4 alkyl.

[0325] In some embodiments, G LL Selected from G LL1-1 and G LL1-2 In some of these embodiments, G LL It's G LL1-1 .

[0326] In some embodiments, R LL It is derived from the above R L Group of groups.

[0327] One skilled in the art will recognize that any one or more of the chemical groups, moieties, and features disclosed herein can be combined in a variety of ways to form linkers useful for conjugating an ADC antibody and an ADC payload (e.g., a cytotoxin) as disclosed herein.

[0328] In some embodiments, when the compounds described herein are provided as single enantiomers or in enantiomerically enriched form, the enantiomerically enriched form has an enantiomeric ratio greater than 60:40, 70:30, 80:20, or 90: 10. In other embodiments, the enantiomeric ratio is greater than 95:5, 97:3, or 99:1.

[0329] IV. Other Chemical Definitions

[0330] C 5-6 Arylene: As used herein, the term "C 5-6 "Arylene" refers to a divalent moiety obtained by removing two hydrogen atoms from an aromatic ring atom of an aromatic compound.

[0331] In this context, a prefix (such as C 5-6 ) represents the number of ring atoms or a range of the number of ring atoms, whether carbon atoms or heteroatoms. The ring atoms may all be carbon atoms, as in "carboarylene", in which case the group is phenylene (C 6 ). Alternatively, the ring atoms may include one or more heteroatoms, as in "heteroarylene". Examples of heteroarylene groups include, but are not limited to, those derived from:

[0332] N 1 :Pyrrole(azole)(C 5 ), pyridine (azine) (C 6 );

[0333] O 1 :Furan (oxol) (C 5 );

[0334] S 1 :Thiophene (thiolene) (C 5 );

[0335] N 1 O 1 :Oxazole (C 5 ), isoxazole (C 5 ), isoxazine (C 6 );

[0336] N 2 O1 :Oxadiazole (furazan) (C 5 );

[0337] N 3 O 1 :Oxatriazole (C 5 );

[0338] N 1 S 1 :Thiazole (C 5 ), isothiazole (C 5 );

[0339] N 2 :Imidazole (1,3-oxadiazole) (C 5 ), pyrazole (1,2-oxadiazole) (C 5 ), pyridazine (1,2-diazine) (C 6 ), pyrimidine (1,3-diazine) (C 6 ) (e.g., cytosine, thymine, uracil), pyrazine (1,4-diazine) (C 6 );and

[0340] N 3 :Triazole (C 5 ), triazine (C 6 ).

[0341] C 1-4 Alkyl: As used herein, the term "C 1-4 "Alkyl" refers to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of a hydrocarbon compound having 1 to 4 carbon atoms, which may be aliphatic or alicyclic, and may be saturated or unsaturated (e.g., partially unsaturated, fully unsaturated). As used herein, the term "C 1-n "Alkyl" refers to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of a hydrocarbon compound having from 1 to n carbon atoms, which may be aliphatic or alicyclic, and may be saturated or unsaturated (e.g., partially unsaturated, fully unsaturated). Thus, the term "alkyl" includes the subclasses alkenyl, alkynyl, cycloalkyl, etc. discussed below.

[0342] Examples of saturated alkyl groups include, but are not limited to, methyl (C 1 ), ethyl (C 2 ), propyl (C 3 ) and butyl (C 4 Examples of saturated straight chain alkyl groups include, but are not limited to, methyl (C 1 ), ethyl (C 2 ), n-propyl (C 3 ) and n-butyl (C 4 Examples of saturated branched alkyl groups include isopropyl (C 3), isobutyl (C 4 ), sec-butyl (C 4 ) and tert-butyl (C 4 ).

[0343] C 2-4 Alkenyl: As used herein, the term "C 2-4 "Alkenyl" refers to an alkyl group having one or more carbon-carbon double bonds. Examples of unsaturated alkenyl groups include, but are not limited to, vinyl (vinyl, -CH=CH 2 ), 1-propenyl (-CH=CH-CH 3 ), 2-propenyl (allyl, -CH-CH=CH 2 ), isopropenyl (1-methylvinyl, -C(CH 3 )=CH 2 ) and butenyl (C 4 ).

[0344] C 2-4 Alkynyl: As used herein, the term "C 2-4 "Alkynyl" refers to an alkyl group having one or more carbon-carbon triple bonds. Examples of unsaturated alkynyl groups include, but are not limited to, ethynyl (-C≡CH) and 2-propynyl (propargyl, -CH 2 -C≡CH).

[0345] C 3-4 Cycloalkyl: As used herein, the term "C 3-4 "Cycloalkyl" refers to an alkyl group that is also a cyclic group; that is, a monovalent moiety obtained by removing a hydrogen atom from an alicyclic ring atom of a cyclic hydrocarbon (carbocyclic) compound, the moiety having 3 to 7 carbon atoms, including 3 to 7 ring atoms. Examples of cycloalkyl groups include, but are not limited to, those derived from saturated monocyclic hydrocarbon compounds such as cyclopropane (C 3 ) and cyclobutane (C 4 ); and unsaturated monocyclic hydrocarbon compounds, such as cyclopropene (C 3 ) and cyclobutene (C 4 ).

[0346] Connection mark: In the formula:

[0347] [Formula 13]

[0348]

[0349] Superscript mark C(＝O) and NH Indicates the group to which the atom is bonded. For example, an NH group is shown bonded to a carbonyl group (not part of the moiety shown), and a carbonyl group is shown bonded to an NH group (not part of the moiety shown).

[0350] V. Cytotoxicity of ADC

[0351] The anti-FRα ADC used in the novel methods disclosed herein comprises an ADC payload and an ADC anti-FRα antibody or antigen-binding fragment thereof. In a preferred embodiment, the anti-FRα ADC comprises an ADC anti-FRα antibody or antigen-binding fragment thereof conjugated to one or more cytotoxins.

[0352] In some embodiments, the number of ADC payloads (e.g., cytotoxins) per antibody (or its antigen-binding fragment) can be expressed as a ratio of ADC payload (i.e., drug) to antibody. This ratio is referred to as drug-to-antibody ratio (DAR). DAR is the average number of drugs (i.e., ADC payloads) attached to each antibody. In some embodiments, DAR is in the range of about 1 to 20. In some embodiments, the range of DAR is selected from about 1 to 10, about 2 to 10, about 2 to 8, about 2 to 6, and about 4 to 10. In preferred embodiments, DAR is about 4 (e.g., 3.8-4.2) or about 8 (e.g., 7.6-8.4), more preferably about 8 (e.g., 7.6-8.4).

[0353] Cytotoxins can be any molecules known in the art that inhibit or prevent cell function and / or cause cell destruction (cell death) and / or exert anti-tumor / anti-proliferative effects. It is known that various types of cytotoxic agents have potential utility in ADC molecules. Suitable cytotoxic agents for ADC include, but are not limited to, topoisomerase I inhibitors (TOPOi), amanitin, auristatin, daunomycin, doxorubicin, duocarmycin, dolestadin, enediynes, lexitropsin, taxanes, puromycin, maytansinoids, vinca alkaloids, tubulysin and pyrrolobenzodiazepines (PBD). Examples of such cytotoxic agents are AFP, MMAF, MMAE, AEB, AEVB, auristatin E, paclitaxel, docetaxel, CC-1065, SN-38, topotecan, morpholino-adriamycin, root mycin, cyanomorpholino-adriamycin, tail sea suline-10, echinomycin, combretatstatin, calicheamicin, maytansine, DM-1, vinblastine, methotrexate and fusobacterium, and their derivatives and analogs. Additional disclosures about cytotoxins suitable for ADC can be found, for example, in International Patent Application Publication Nos. WO 2015 / 155345 and WO 2015 / 157592, which are incorporated herein by reference in their entirety.

[0354] In some embodiments, the ADC comprises one or more selected from topoisomerase I inhibitors, tubulysin derivatives, pyrrolobenzodiazepines, An anti-FRα antibody or antigen-binding fragment thereof conjugated with a cytotoxin or a combination thereof. For example, the ADC antibody or antigen-binding fragment thereof is conjugated with one or more cytotoxins selected from the group consisting of topoisomerase I inhibitor SG3932 (also known as AZ14170133), SG4010, SG4057 or SG4052 (whose structures are provided below), or a combination thereof.

[0355] Preferably, the ADC antibody or antigen-binding fragment thereof is conjugated to a topoisomerase I inhibitor, more preferably to the topoisomerase I inhibitor SG3932.

[0356] Topoisomerase inhibitors are chemical compounds that block the action of topoisomerases (topoisomerases I and II), a type of enzyme that controls changes in DNA structure during the normal cell cycle by catalyzing the breakage and reconnection of the phosphodiester backbone of the DNA chain. Topoisomerase I inhibitors are advantageous because they mediate highly effective tumor cell killing with less toxicity to patients. In particular, alternative payloads (such as microtubule inhibitors) that have been commonly used to develop anti-FRα ADCs to date are known to have toxicity issues (Hinrichs, et al. AAPS J. 2015 September; 17(5): 1055–1064). In addition, the use of less hydrophobic linkers and less potent warheads (e.g., TOPOi) will contribute to bystander killing in heterogeneous tumors. Although bystander activity can be achieved by increasing potency and / or improving warhead permeability by increased hydrophobicity, this may lead to increased toxicity due to nonspecific uptake.

[0357] General examples of suitable topoisomerase I inhibitors are represented by the following compounds:

[0358] [Formula 14]

[0359]

[0360] The compound shown above is represented as A* and may be referred to herein as a "drug unit". Compound A* is preferably provided with a linker for connection (preferably conjugation) to an ADC antibody or antigen-binding fragment. In a preferred embodiment, the linker is cleavably connected (e.g., conjugated) to an amino residue, such as an amino acid of an ADC antibody or antigen-binding fragment.

[0361] More particularly, examples of suitable topoisomerase I inhibitors are represented by the following compounds "I" having the formula:

[0362] [Formula 15]

[0363]

[0364] Compound I includes salts and solvates thereof, wherein R L is a linker for attachment to an ADC antibody or antigen-binding fragment thereof described herein. L As mentioned above.

[0365] Therefore, the ADC used in the new method can have the following general formula:

[0366] [Formula 10]

[0367] L–(D L ) p

[0368] or a pharmaceutically acceptable salt or solvate thereof, wherein L and p are as defined above, D L is a topoisomerase I inhibitor having a linker (eg, a drug linker unit) of a compound III having the formula:

[0369] [Formula 16]

[0370]

[0371] Compound III includes salts and solvates thereof, wherein R LL is a linker for attachment to an ADC antibody or antigen-binding fragment thereof described herein. LL As defined above.

[0372] In some embodiments, the compound I of formula 15 belongs to I having the following formula P type:

[0373] [Formula 17]

[0374]

[0375] Compound I P Including salts and solvates thereof, wherein R LP is a linker for attachment to an ADC antibody or an antigen-binding fragment thereof, wherein the linker is selected from a compound Ia having the following formula P :

[0376] [Formula 18]

[0377]

[0378] Where Q P yes:

[0379] [Formula 19]

[0380]

[0381] Where Q XP is such that Q P is an amino acid residue, a dipeptide residue or a tripeptide residue. P yes:

[0382] [Formula 20]

[0383]

[0384] Where aP = 0 to 5, bP = 0 to 16, cP = 0 or 1, dP = 0 to 5. Alternatively, G in Formula 18 L is a compound Ib having the formula:

[0385] [Formula 21]

[0386]

[0387] Where R L1 and R L2 are independently selected from H and methyl, or together with the carbon atom to which they are attached form a cyclopropylene or cyclobutylene group; and e is 0 or 1. aP in Formula 20 may be 0, 1, 2, 3, 4, or 5. In some embodiments, aP is 0 to 3. In some of these embodiments, aP is 0 or 1. In other embodiments, aP is 0. bP in Formula 20 may be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b is 0 to 12. In some of these embodiments, bP is 0 to 8, and may be 0, 2, 4, or 8. cP in Formula 20 may be 0 or 1. dP in Formula 20 may be 0, 1, 2, 3, 4, or 5. In some of these embodiments, dP is 0 to 3. In some of these embodiments, dP is 1 or 2. In other embodiments, dP is 2.

[0388] In formula 18, X P In some embodiments of the present invention, aP is 0, cP is 1 and dP is 2, and bP can be from 0 to 8. In some of these embodiments, bP is 0, 4 or 8.

[0389] For compound I of formula 15, Q of formula 13 X The preference can be applied to Q in Equation 19 XP (e.g. where appropriate).

[0390] The above G of compound I of formula 15 L , R L1 , R L2 The preference for and e can be applied to compound I of formula 17 P .

[0391] In some embodiments, the conjugate of the ADC of Formula 10 is an ADC having the formula P :

[0392] [Formula 22]

[0393] L–(D LP ) p

[0394] or a pharmaceutically acceptable salt or solvate thereof, wherein L is an ADC antibody or an antigen-binding fragment thereof, and D LP is a topoisomerase I inhibitor (eg, a drug linker unit), designated as compound III having the formula P :

[0395] [Formula 23]

[0396]

[0397] Where R LLP is a linker connected to an ADC antibody or an antigen-binding fragment thereof, wherein the linker is selected from group Ia P ' and Ib'. Group Ia P ' is represented by the following formula:

[0398] [Equation 24]

[0399]

[0400] Where Q P , X P and G LL As defined above. Group Ib' is represented by the formula:

[0401] [Formula 25]

[0402]

[0403] Where R L1 and R L2 As defined above; and p is an integer from 1 to 20.

[0404] In some embodiments, the compound I of formula 15 belongs to I having the following formula P2 type:

[0405] [Equation 26]

[0406]

[0407] And its salts and solvates. LP2 is a linker for connecting to an ADC antibody or an antigen-binding fragment thereof, wherein the linker is selected from group IaP2 and Ib. Group Ia P2 It is expressed by the following formula:

[0408] [Formula 27]

[0409]

[0410] Where Q is:

[0411] [Equation 28]

[0412]

[0413] Where Q X is such that Q is an amino acid residue, a dipeptide residue, a tripeptide residue or a tetrapeptide residue. P2 yes:

[0414] [Equation 29]

[0415]

[0416] Wherein aP2=0 to 5, b1P2=0 to 16, b2P2=0 to 16, cP2=0 or 1, dP2=0 to 5, wherein at least b1P2 or b2P2=0 (i.e., only one of b1 and b2 may not be 0). GL in Formula 27 is a linker for connecting to an ADC antibody or an antigen-binding fragment thereof.

[0417] The group Ib is represented by the following formula:

[0418] [Formula 30]

[0419]

[0420] Where R L1 and R L2 are independently selected from H and methyl, or together with the carbon atom to which they are attached form cyclopropylene or cyclobutylene; and e is 0 or 1.

[0421] In formula 27, X P2In some embodiments of, aP2 can be 0, 1, 2, 3, 4, or 5. In some embodiments, aP2 is 0 to 3. In some of these embodiments, aP2 is 0 or 1. In other embodiments, aP2 is 0. In some embodiments, b1P2 can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b1P2 is 0 to 12. In some of these embodiments, b1P2 is 0 to 8, and can be 0, 2, 3, 4, 5, or 8. In some embodiments, b2P2 can be 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16. In some embodiments, b2P2 is 0 to 12. In some of these embodiments, b2P2 is 0 to 8, and can be 0, 2, 3, 4, 5, or 8. Preferably, only one of b1P2 and b2P2 may not be 0. In some embodiments, cP2 may be 0 or 1. In some embodiments, dP2 may be 0, 1, 2, 3, 4, or 5. In some embodiments, dP2 is 0 to 3. In some of these embodiments, dP2 is 1 or 2. In other embodiments, dP2 is 2. In other embodiments, dP2 is 5.

[0422] In X P2 In some embodiments of X, aP2 is 0, b1P2 is 0, cP2 is 1 and dP2 is 2, and b2P2 can be 0 to 8. In some of these embodiments, b2P2 is 0, 2, 3, 4, 5 or 8. P2 In some embodiments of X, aP2 is 1, b2P2 is 0, cP2 is 0 and dP2 is 0, and b1P2 can be 0 to 8. In some of these embodiments, b1P2 is 0, 2, 3, 4, 5 or 8. P2 In some embodiments of X, aP2 is 0, b1P2 is 0, cP2 is 0 and dP2 is 1, and b2P2 can be 0 to 8. In some of these embodiments, b2P2 is 0, 2, 3, 4, 5 or 8. P2 In some embodiments of the present invention, b1P2 is 0, b2P2 is 0, cP2 is 0, and one of aP2 and dP2 is 0. The other of aP2 and dP2 is 1 to 5. In some of these embodiments, the other of aP2 and dP2 is 1. In other of these embodiments, the other of aP2 and dP2 is 5.

[0423] For the compound I of formula 15, Q of formula 2 X The preference can be applied to group Ia of formula 27 P2 QX (e.g. where appropriate). L , R L1 , R L2 The preference of and e can be applied to compound type I of formula 26 P2 .

[0424] In some embodiments, the conjugate of ADC of formula 10 is a compound ADC having the formula P2 :

[0425] [Equation 31]

[0426] L–(D LP2 ) p

[0427] or a pharmaceutically acceptable salt or solvate thereof, wherein L is an ADC antibody or an antigen-binding fragment thereof, and D LP2 is a topoisomerase I inhibitor (eg, a drug linker unit) which is a compound III having the formula P2 :

[0428] [Formula 32]

[0429]

[0430] R LLP2 is a linker connected to an ADC antibody or an antigen-binding fragment thereof, wherein the linker is selected from group Ia P2 ' and Ib'. Group Ia P2 ' is represented by the following formula:

[0431] [Formula 33]

[0432]

[0433] Where Q and X P2 As defined above, and G LL It is a linker connected to the ADC antibody or its antigen-binding fragment. Group Ib' is represented by the following formula:

[0434] [Equation 34]

[0435]

[0436] Where R L1 and R L2 As defined above; and p is an integer from 1 to 20.

[0437] Particularly suitable topoisomerase I inhibitors include those having the formula:

[0438] [Formula 35]

[0439]

[0440] [Equation 36]

[0441]

[0442] [Equation 37]

[0443]

[0444] [Equation 38]

[0445]

[0446] and / or

[0447] [Equation 39]

[0448]

[0449] The topoisomerase I inhibitor SG3932 is particularly preferred. Therefore, in a preferred embodiment, the ADC used in the new method comprises an antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor having the formula of SG3932):

[0450] [Formula 35]

[0451]

[0452] Synthetic methods for preparing topoisomerase I inhibitors are described, for example, in WO 2020 / 200880, which is incorporated herein by reference.

[0453] While topoisomerase I inhibitors are preferred as outlined above, it should be noted that any suitable ADC payload (eg, drug / cytotoxin) may be attached to the ADC antibody or antigen-binding fragment thereof.

[0454] In certain embodiments, the anti-FRα ADC used in the present invention does not contain a microtubule inhibitor, such as a tubulin inhibitor (eg, maytansinoid, auristatin). Microtubule inhibitor molecules have potential difficult-to-treat toxicity, thereby limiting dosing.

[0455] VI. Preparation of ADC

[0456] The ADCs of the present disclosure can be made in a variety of ways using known organic chemistry reactions, conditions, and reagents, such as: (1) reacting reactive substituents of an ADC antibody or antigen-binding fragment with a divalent linker reagent, followed by reaction with an ADC payload (e.g., a cytotoxin), preferably a topoisomerase I inhibitor; or (2) reacting reactive substituents of an ADC payload (e.g., a cytotoxin), preferably a topoisomerase I inhibitor, with a divalent linker reagent, followed by reaction with reactive substituents of an ADC antibody or antigen-binding fragment thereof.

[0457] As disclosed herein, reactive substituents that may be present in an ADC antibody or antigen-binding fragment thereof include, but are not limited to, nucleophilic groups such as (i) an N-terminal amine group, (ii) a side chain amine group, such as lysine, (iii) a side chain thiol group, such as cysteine, and (iv) a sugar hydroxyl or amino group, wherein the antibody is glycosylated. As disclosed herein, reactive substituents that may be present in an ADC antibody or antigen-binding fragment thereof include, but are not limited to, hydroxyl moieties of serine, threonine, and tyrosine residues; amino moieties of lysine residues; carboxyl moieties of aspartic acid and glutamic acid residues; and thiol moieties of cysteine ​​residues, as well as propargyl, azido, halogenated aryl (e.g., fluoroaryl), halogenated heteroaryl (e.g., fluoroheteroaryl), halogenated alkyl, and halogenated heteroalkyl moieties of non-naturally occurring amino acids. In some embodiments, reactive substituents present in an ADC antibody or antigen-binding fragment thereof as disclosed herein include amine or thiol moieties. Certain antibodies have cysteine ​​bridges that are reducible interchain disulfides. By treating antibodies with reducing agents such as DL-dithiothreitol (DTT) and tris(2-carboxyethyl)phosphine (TCEP), they can be made reactive for conjugation with linker reagents. Thus, each cysteine ​​bridge will theoretically result in the formation of two reactive thiol nucleophiles. The reaction of lysine with 2-iminothiolane (Trouble's reagent) results in the conversion of amines into thiols, which can be used to introduce additional nucleophilic groups into antibodies. One, two, three, four or more cysteine ​​residues can be used to insert reactive thiol groups into ADC antibodies (or fragments thereof) (e.g., to prepare mutant antibodies comprising one or more non-natural cysteine ​​amino acid residues). Engineering antibodies with reactive cysteine ​​amino acids is described in U.S. Pat. No. 7,521,541, which is incorporated herein by reference.

[0458] In some embodiments, the ADC antibody or antigen-binding fragment thereof may have one or more carbohydrate groups that can be chemically altered to contain one or more sulfhydryl groups. The ADC is then formed by conjugation via the sulfur atom of the sulfhydryl group.

[0459] In some embodiments, the ADC antibody may contain one or more carbohydrate groups that can be oxidized to produce an aldehyde (-CHO) group (see, e.g., Laguzza et al., J. Med. Chem. 1989, 32(3), 548-55). Conjugation via the corresponding aldehyde results in the formation of the ADC. Other protocols for modifying proteins to attach or associate ADC payloads (e.g., cytotoxins) are described in Coligan et al., Current Protocols in Protein Science, Vol. 2, John Wiley & Sons (2002). Methods for conjugating linker-drug moieties to cell targeting proteins (e.g., antibodies, immunoglobulins, or fragments thereof) can be found in, e.g., U.S. Pat. No. 5,208,020; U.S. Pat. No. 6,441,163; WO 2005 / 037992; WO 2005 / 081711; and WO 2006 / 034488, each of which is incorporated herein by reference.

[0460] Conventional conjugation strategies for antibodies or their antigen-binding fragments rely on the payload being conjugated to the antibody or fragment arbitrarily or randomly by lysine or cysteine. In some embodiments, the ADC antibody or its antigen-binding fragment is randomly conjugated to an ADC payload (e.g., a cytotoxin), preferably a topoisomerase I inhibitor, for example, by partially reducing the ADC antibody or fragment, and then reacting with the desired ADC payload, with or without a connected linker portion. DTT or other reducing agents (e.g., TCEP) can be used to reduce the ADC antibody or fragment for similar reduction. The ADC payload with or without a connected linker portion can then be added to the reduced antibody or fragment in a molar excess in the presence of DMSO. After conjugation, a quencher such as N-acetyl-L-cysteine ​​can be added to quench unreacted agents. The reaction mixture can then be purified (e.g., by TFF, SEC-FPLC, CHT, spin filtration centrifugation) and the buffer exchanged to PBS or other related formulation buffers.

[0461] In some embodiments, the ADC payload (e.g., cytotoxin) is conjugated to the ADC antibody or antigen-binding fragment thereof by site-specific conjugation. In some embodiments, site-specific conjugation of the therapeutic moiety to the ADC antibody using reactive amino acid residues at specific positions can produce a homogeneous ADC preparation with uniform stoichiometry.

[0462] Site-specific conjugation can be carried out by cysteine, residues or non-natural amino acids. In a preferred embodiment, the ADC payload (e.g., cytotoxin) is conjugated to the ADC antibody or its antigen-binding fragment through at least one cysteine ​​residue. Cysteine ​​amino acids can be engineered at reactive sites in the ADC antibody (or its antigen-binding fragment), and preferably do not form intrachain or intermolecular disulfide bonds (Junutula, et al. 2008b Nature Biotech., 26 (8): 925-932; Dornan et al. (2009) Blood 114 (13): 2721-2729; US 7521541; US ​​7723485; WO2009 / 052249). In some embodiments, the ADC payload (e.g., cytotoxin) is conjugated to the antibody or antigen-binding fragment thereof by cysteine ​​substitution at at least one of positions 239, 248, 254, 273, 279, 282, 284, 286, 287, 289, 297, 298, 312, 324, 326, 330, 335, 337, 339, 350, 355, 356, 359, 360, 361, 375, 383, 384, 389, 398, 400, 413, 415, 418, 422, 440, 441, 442, 443, and 446, wherein the numbering corresponds to the EU index of Kabat. In some embodiments, the specific Kabat positions are 239, 442, or both. In some embodiments, the specific position is Kabat position 442, an amino acid insertion between Kabat positions 239 and 240, or both. In some embodiments, the ADC payload (e.g., cytotoxin) is conjugated to the ADC antibody or antigen-binding fragment thereof via a thiol-maleimide linkage. In some aspects, the amino acid side chain is a sulfhydryl side chain.

[0463] In the case where more than one nucleophilic or electrophilic group of the ADC antibody or antigen-binding fragment thereof reacts with an ADC payload (e.g., a cytotoxin), then the resulting product may be a mixture of ADC and a distribution (e.g., 1, 2, 3, etc.) of ADC payload units attached to the antibody. Liquid chromatography methods such as hydrophobic interaction chromatography (HIC) can separate compounds in the mixture according to ADC payload loading values. Preparations of ADC with a single ADC payload loading value (p) can be separated.

[0464] The average ADC payload (e.g., cytotoxin) number of each ADC antibody (or antigen-binding fragment) in the ADC preparation from the conjugation reaction can be characterized by conventional means (such as UV, reversed-phase HPLC, HIC, mass spectrometry, ELISA assays, and electrophoresis). The number distribution of ADCs in terms of p can also be determined. By ELISA, the average value of p in a particular preparation of ADC can be determined (Hamblett et al. (2004) Clin. Cancer Res. 10: 7063-7070; Sanderson et al. (2005) Clin. Cancer Res. 11: 843-852). In some cases, the separation, purification, and characterization of homogeneous ADCs with a certain value of p and antibodies with other ADC payloads (e.g., cytotoxins) can be achieved by means such as reversed-phase HPLC, electrophoresis, TFF, SEC-FPLC, CHT, and spin filter centrifugation. Such techniques are also applicable to other types of conjugates.

[0465] VII. Pharmaceutical Compositions

[0466] The anti-FRα ADCs may be administered as pharmaceutical compositions containing one or more pharmaceutically compatible components.

[0467] The term "pharmaceutical composition" refers to a preparation that is in a form that permits the biological activity of the active ingredients to be effective and does not contain additional components that are unacceptably toxic to the subject to which the combination is to be administered. Such compositions may be sterile and may include a pharmaceutically acceptable carrier, such as normal saline. Suitable pharmaceutical compositions may include one or more of the following: a buffer (e.g., acetate, phosphate or citrate buffer), a surfactant (e.g., polysorbate), a stabilizer (e.g., human albumin), a preservative (e.g., benzyl alcohol), and an absorption promoter and / or other conventional solubilizing or dispersing agents for enhancing bioavailability.

[0468] As used herein, the term "pharmaceutically acceptable" means approved by a regulatory agency of the Federal or a state government or listed in the US Pharmacopeia, European Pharmacopeia or other generally recognized pharmacopeia for use in animals, and more particularly in humans.

[0469] In some embodiments, the pharmaceutical composition may be contained in one or more formulations selected from capsules, tablets, aqueous suspensions, solutions, nasal aerosols, lyophilized powders that can be reconstituted prior to use to make suspensions or solutions, or combinations thereof.

[0470] For example, the anti-FRα ADC AB1370049-SG3932-DAR8 used in the new method can be administered as a pharmaceutical composition containing a buffer (such as a histidine buffer), an excipient (such as sucrose) and a surfactant (such as polysorbate 80). The pharmaceutical composition containing the anti-FRα ADC used in the new method can be used as an injection, as an aqueous injection or a lyophilized injection, or even as a lyophilized injection.

[0471] The pharmaceutical compositions disclosed herein can be administered to a patient by any suitable systemic or local route of administration. For example, administration can be oral, buccal, sublingual, ocular, intranasal, intratracheal, pulmonary, topical, transdermal, urogenital, rectal, subcutaneous, intravenous, intraarterial, intraperitoneal, intramuscular, intracranial, intrathecal, epidural, intraventricular, or intratumoral.

[0472] The pharmaceutical compositions disclosed herein can be formulated for administration by any appropriate means, such as by epidermal or transdermal patches, ointments, lotions, creams or gels; by nebulizers, vaporizers or inhalers; by injection or infusion; or in the form of capsules, tablets, liquid solutions or suspensions in water or non-aqueous media, drops, suppositories, enemas, sprays or powders. The most appropriate route of administration in any given case will depend on the patient's physical and mental condition, the nature and severity of the disease, and the desired characteristics of the formulation.

[0473] Pharmaceutical compositions for oral administration can be in tablet, capsule, powder or liquid form. Tablets can contain solid carriers or adjuvants. Liquid pharmaceutical compositions usually contain liquid carriers, such as water, petroleum, animal or vegetable oils, mineral oils or synthetic oils. Physiological saline solutions, dextrose or other sugar solutions or glycols such as ethylene glycol, propylene glycol or polyethylene glycol can be contained. Capsules can contain solid carriers such as gelatin.

[0474] For intravenous, cutaneous or subcutaneous injection, or injection at the site of affliction, the active ingredient will be in the form of a parenterally acceptable aqueous solution that is pyrogen-free and has a suitable pH, isotonicity and stability. Those skilled in the art are well able to prepare suitable solutions using, for example, isotonic vehicles such as sodium chloride injection, Ringer's injection, and lactated Ringer's injection. Preservatives, stabilizers, buffers, antioxidants and / or other additives may be included as needed.

[0475] If the pharmaceutical composition containing the anti-FRα ADC used in the new method is an aqueous injection, it can be diluted with a suitable diluent and then administered as an intravenous infusion. Examples of diluents may include dextrose solution and physiological saline.

[0476] If the pharmaceutical composition containing the anti-FRα ADC used in the new method is a lyophilized injection, it can be dissolved with injection-grade water, then diluted with a suitable diluent to the required amount, and then administered as an intravenous infusion. Examples of diluents include dextrose solution and physiological saline.

[0477] VIII. Therapeutic Uses of Anti-FRα ADCs

[0478] Preferably, the ADC disclosed herein can be used to prevent, treat or ameliorate symptoms associated with a disease, disorder or infection, preferably cancer. The term "cancer" is used to have the same meaning as the term "tumor".

[0479] "Treatment" refers to therapeutic measures that cure, slow down, alleviate the symptoms of the diagnosed pathological disorder or condition and / or prevent its progression. Therefore, those in need of treatment include those who already have the condition. In some embodiments, if the patient shows, for example, total, partial or temporary relief or elimination of symptoms associated with the disease or condition (preferably cancer), then the subject's disease or condition (preferably cancer) is successfully "treated" according to the new methods disclosed herein.

[0480] "Prevention" refers to prophylactic or preventative measures to prevent and / or slow the development of a targeted pathological condition or disorder, preferably cancer. Thus, those in need of prevention include those predisposed to or suspected of having the disorder.

[0481] The terms "patient", "individual" and "subject" are used interchangeably herein to refer to mammalian subjects. In some embodiments, "subject" is a human, domestic animal, farm animal, sports animal and zoo animal, such as a human, non-human primate, dog, cat, guinea pig, rabbit, rat, mouse, horse, cattle, etc. In some embodiments, the subject is a cynomolgus monkey (acacafascicularis). In a preferred embodiment, the subject is a human. In some embodiments of the new method, the patient may not have been diagnosed with cancer before. Alternatively, the patient may have been diagnosed with cancer before. The patient may also be a patient who exhibits disease risk factors, or a patient with no cancer symptoms. The patient may also be a patient with cancer or a patient at risk of developing cancer. Therefore, in one embodiment, the method of the present invention can be used to confirm the presence of cancer in a patient. For example, the patient may have been previously diagnosed with cancer by alternative means. In one embodiment, the patient has previously been administered cancer therapy.

[0482] In some embodiments, the anti-FRα ADC disclosed herein can be used to treat cancers associated with FRα expression. In other words, the cancers mentioned herein may include cancerous cells expressing FRα. The cancerous cells may be contained within a tumor. In some embodiments, the cancer includes cancer cells with heterogeneous expression of FRα and / or low expression of FRα.

[0483] After undergoing cellular internalization into cancer cells, the anti-FRα ADC used in the present invention is cleaved at the linker unit and releases the drug unit into the cancer cells. The anti-FRα antibody-drug conjugate used in the new method also has a bystander effect, in which the anti-FRα antibody-drug conjugate is internalized into cancer cells expressing the target protein FRα, and the drug unit then also exerts an anti-tumor effect on neighboring cancer cells that do not express the target protein FRα.

[0484] Preferably, the cancer is selected from ovarian cancer, lung cancer, endometrial cancer, breast cancer (e.g., TNBC), cervical cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC) and malignant pleural mesothelioma. More preferably, the cancer is ovarian cancer or lung cancer. In some embodiments, the cancer is non-small cell lung cancer (NSCLC), preferably selected from one or more of squamous NSCLC, adenocarcinoma NSCLC or a combination thereof.

[0485] Further examples of cancer include, but are not limited to, benign, pre-malignant and malignant cell proliferations, including, but not limited to, neoplasms and tumors (e.g., histiocytomas, gliomas, astrocytomas, osteomas), cancers (e.g., ovarian cancer, lung cancer, non-small cell lung cancer (squamous cell carcinoma or adenocarcinoma), endometrial cancer, pancreatic cancer, gastric cancer, colorectal cancer, head and neck squamous cell carcinoma, malignant pleural mesothelioma, breast cancer (e.g., TNBC), and renal cancer. Any type of cell can be treated, including, but not limited to, lung, gastrointestinal, breast (breast), ovarian, kidney (kidney), and pancreatic.

[0486] The anti-FRα ADCs disclosed herein can be expected to exert therapeutic effects by being administered to patients as systemic therapy, and additionally by being applied locally to cancer tissue.

[0487] IX. Diagnostic Antibodies

[0488] The diagnostic antibody used in the new method binds to FRα on cells (e.g., cancer cells) in a tissue sample of a cancer patient. In a preferred embodiment, the diagnostic antibody specifically binds to human FRα. The diagnostic antibody is labeled to help detect cell binding. In some embodiments, the label can be a fluorophore or a dye. In some embodiments, the label can be used for immunohistochemistry. In a preferred embodiment, the diagnostic antibody is linked to 3,3'-diaminobenzidine (DAB). In some embodiments, the diagnostic antibody is an anti-FRα antibody obtained from Ventana (VMSI, #742-5065). In some embodiments, the diagnostic antibody is an anti-rabbit FRα antibody. In a preferred embodiment, the diagnostic antibody is an anti-rabbit FRα antibody obtained from Abcam (#ab221543). In some embodiments, the diagnostic antibody is tightly associated with an ADC antibody or antigen-binding fragment as described herein. In some embodiments, the diagnostic antibody is tightly associated with the ADC antibody of ADC AB1370049-SG3932.

[0489] In some embodiments, the diagnostic antibody binds to cancer cells from ovarian cancer, lung cancer, endometrial cancer, colon cancer, breast cancer (e.g., TNBC), cervical cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), and malignant pleural mesothelioma. More preferably, the diagnostic antibody binds to cancer cells from ovarian cancer, lung cancer, endometrial cancer, or breast cancer. In some embodiments, the diagnostic antibody binds to cancer cells from non-small cell lung cancer (NSCLC), preferably selected from squamous NSCLC, adenocarcinoma NSCLC, or a combination thereof.

[0490] X. Methods of Predicting Patient Response to Administration of Anti-FRα ADCs

[0491] Fig. 9is a flow chart of steps 11-17 of method 10, by which the analysis system analyzes digital images of tissues from cancer patients and predicts how the cancer patient will likely respond to a therapy involving an anti-FRα antibody-drug conjugate (ADC). In one embodiment, the ADC is an anti-FRα ADC AB1370049-SG3932-DAR8. In one embodiment, the method predicts the response of a patient with a cancer selected from ovarian cancer, lung cancer, endometrial cancer, breast cancer (e.g., TNBC), cervical cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), and malignant pleural mesothelioma to the ADC. More preferably, the method predicts the response of a patient with a cancer selected from ovarian cancer, non-small cell lung cancer, endometrial cancer, and colon cancer to the ADC. In some embodiments, the method predicts the response of a patient with non-small cell lung cancer (NSCLC) to the ADC, and the non-small cell lung cancer is preferably selected from squamous NSCLC, adenocarcinoma NSCLC, or a combination thereof.

[0492] In a first step 11, a high-resolution digital image of a tissue section from a cancer patient is acquired, the tissue section having been stained with one or more biomarkers or stains.

[0493] To predict the efficacy of an ADC therapy, a diagnostic antibody (e.g., a diagnostic biomarker) with an attached dye is used that targets the same protein as the protein targeted by the ADC therapy. In one embodiment, the anti-FRα ADC therapy targeted by the score is an anti-FRα antibody conjugated to a topoisomerase I inhibitor.

[0494] Thus, in embodiments, the diagnostic antibody also targets the FRα protein.

[0495] In step 12, a pre-trained convolutional neural network processes a digital image of a cancer patient's tissue that has been stained with a diagnostic antibody attached to a dye such as 3,3'-diaminobenzidine (DAB). The staining intensity of the dye in the cell membrane of the cancer cell is determined based on the average staining intensity of the dye for all pixels associated with the corresponding segmented cell membrane object. In addition, the staining intensity of the dye in a single pixel is calculated based on the red, green, and blue components of the pixel. The result of the image analysis process is two posterior image layers that represent, for each pixel in the digital image, the probability that the pixel belongs to the cell nucleus and the probability that the pixel belongs to the cell membrane.

[0496] In step 13, a single cancer cell is detected based on a heuristic image analysis of the posterior layer of the cell nucleus and the cell membrane, and a cancer cell object including a cell membrane object is generated.

[0497] In step 14, each cancer cell is identified as optical density positive or optical density negative based on the amount of DAB in the cell membrane. Based on the average (average) optical density of the brown diaminobenzidine (DAB) signal in all pixels of the cell membrane, the amount of DAB is determined by the staining intensity of each cell membrane. (i) If the average optical density of the cancer cell is equal to or greater than the optical density threshold, each cancer cell is identified as optical density positive, or (ii) If the average optical density of the cancer cell is less than the optical density threshold, each cancer cell is identified as optical density negative. In one embodiment, the optical density threshold is in the range of 10 to 15 on a scale where the maximum optical density is 220. For example, the optical density threshold is 12.

[0498] In step 15, a binary proximity score is generated for the digital image of the tissue sample, the proximity score being equal to the percentage of cancer cells in the digital image that are either optical density positive or optical density negative but within a predefined distance of an optical density positive cancer cell. In one embodiment, the predefined distance is twenty-five microns.

[0499] In step 16, if the proximity score exceeds a predetermined percentage threshold, the cancer patient is identified as a patient who will likely benefit from administration of an anti-FRα ADC. In one embodiment, the predetermined percentage threshold is 20.

[0500] The threshold optical density used in step 14, the predefined distance used in step 15, and the predetermined percentage threshold used in step 16 are optimized using a training cohort of patients with known responses to ADC therapy. Optimization is performed using Spearman correlation analysis of proximity score versus patient response to obtain the lowest p-value and rank R value. Alternatively, a receiver operating characteristic (ROC) analysis is performed to optimize the cutoff point to distinguish binary clinical outcomes (e.g., responsiveness R / non-responsiveness NR). The binary proximity score indicates how a cancer patient will respond to anti-FRα ADC therapy.

[0501] In step 17, if the score exceeds a predetermined percentage threshold, a therapy involving an anti-FRα ADC is recommended to the patient with a positive score.

[0502] XI. Examples of Prediction and Scoring Methods

[0503] A. Image Analysis of Stained Tissues

[0504] Now let's describe the specific images of stained cancer tissue. Fig. 9 method.

[0505] In step 11, the tissue sample is immunohistochemically stained using a dye linked to a diagnostic antibody that binds to a relevant protein on cancer cells in the tissue sample. Fig.10(Upper left image) is a digital image 18 of a portion of the stained tissue acquired in step 11. Image 18 shows tissue from a cancer patient that has been immunohistochemically stained with an anti-FRα diagnostic antibody linked to a dye. In this example, the diagnostic antibody is an anti-FRα generated clone (denoted as FRαIHC Tool Ab), reformatted as a mouse anti-human FRαIgG1 clone. IgG1 represents the isotype of the anti-FRα antibody. The staining protocol was developed using the Dako Autostainer platform. The anti-FRα antibody binds to the membrane protein FRα, such that 3,3'-diaminobenzidine (DAB) staining indicates the location of the protein FRα in the tissue sample.

[0506] In step 12, image analysis is performed on the digital image 18 to generate a posterior image layer of the cancer cell nucleus and cell membrane using a convolutional neural network. Image analysis is used to detect cancer cells and their components, such as the cell nucleus, cell membrane, and cytoplasm. Fig.10 The image analysis process of step 12 is shown. The convolutional neural network generates a posterior layer (gray value image) that indicates, for each pixel of the digital image 18, that each pixel belongs to the nucleus of a cell ( Fig.10 upper right image) or the cell membrane ( Fig.10 The lower left image shows the probability of a cell membrane object being generated by growing outward from the cell nuclear center until encountering growth from another cell nuclear center.

[0507] In one embodiment, the convolutional neural network includes a series of convolutional layers from the input image 18 toward a bottleneck layer with very low spatial size (1 to 16 pixels), and a series of deconvolutional layers toward a posterior layer with the same size as the input image 18. This network architecture is called U-Net. The training of the weights of the convolutional neural network is performed by generating manually annotated layers for cell nuclei and cell membranes in multiple training images, and then adjusting the network weights through an optimization algorithm so that the generated posterior layer is most similar to the manually generated annotation layer.

[0508] In another embodiment, an annotation layer of the nucleus and cell membrane is automatically generated and manually corrected in multiple training images. The epithelial region and the center of the cell nucleus are manually annotated as regions and points, respectively. For each training image, a cell membrane segmentation is automatically generated by applying a regional growth algorithm (e.g., watershed segmentation) seeded by the center of the annotated cell nucleus and constrained by the range of the annotated epithelial region. Given a training image, a cell nucleus segmentation is automatically generated by applying a spot detection algorithm (e.g., by the maximum stable extreme region MSER algorithm) and by selecting only the detected spots containing the annotated cell nucleus center as the cell nucleus. The automatically generated cell membrane and cell nucleus segmentation are visually inspected and manually corrected if necessary. The correction step involves one of the following methods: rejecting incorrectly segmented cell membranes or cell nuclei, explicitly accepting correctly annotated cell membranes or cell nuclei, or refining the shape of the membrane or cell nucleus. For each image with annotated cell membranes or cell nuclei, an annotation layer is created. In one embodiment, if each pixel of the annotation layer belongs to an annotation object (cell membrane or cell nucleus), it is assigned to "1"; otherwise it is assigned to "0". In another embodiment, the pixels of the annotation layer represent the distance to the nearest annotation object. The network weights are adjusted by an optimization algorithm so that the generated posterior layer is most similar to the automatically generated cell membrane and cell nucleus annotation layers.

[0509] Fig.11 Step 13 is shown, where individual cancer cell objects, each comprising a cell membrane and cytoplasm, are detected. The heuristic image analysis process uses watershed segmentation to segment the cell nuclei using a cell nucleus posterior layer generated by a convolutional neural network. Segmentation generates cell nucleus objects. Each cell nucleus object is assigned a unique identifier (UID). The individually identified cell nuclei are Fig.11 (lower right image) as a dark object. The detected cell nuclei are also shown as an overlay in the input image 18 (upper left image) and in the posterior layer of the cell nucleus (upper right image) and cell membrane (lower left image).

[0510] In one embodiment, watershed segmentation involves thresholding the nucleus posterior layer using a predefined first size threshold. All single connected pixels above the first size threshold are considered to belong to the nucleus object. Nucleus objects with an area less than 16um^2 will be discarded. Each nucleus object is assigned a UID. In a subsequent step, the nucleus object grows toward the smaller nucleus posterior, where the added nucleus posterior pixels must be greater than a second predefined threshold.

[0511] Fig.12Detection and segmentation of cell membrane objects are shown by growing the region of the detected cell's border pixels outward to the cell membrane probability layer and a predefined cell membrane layer posterior threshold. The thicker border region becomes the cell membrane object. Each cell membrane object is assigned the same UID as the associated cell nucleus object.

[0512] Use the UID of the nucleus to assign the space between the cell membrane and the nucleus to the cytoplasm. For each cell membrane ( Fig.12 see upper left image) and cytoplasm ( Fig.12 See upper left image), the average optical density of the DAB staining is exported along with the UID to a file on the hard drive. For each cell (defined to include the nucleus, cytoplasm and cell membrane), the center of gravity position (x, y) of the cell within the slice is also exported. This file can reside on a hard drive, solid state drive, or a portion of dedicated RAM in a computer system.

[0513] Fig.13 Shows the results of image analysis in an image analysis software environment. Fig.13 (Upper left image) Segmentation of cell nucleus objects and cell membrane objects is shown as an overlay on a digital image of stained tissue. Fig.13 (Lower left image) Segmentation of a cell nucleus object is shown as an overlay of the optical density representation of the digital image. Dark optical density pixels are associated with high DAB amounts, and bright optical density pixels are associated with low DAB amounts. The DAB optical density of each image pixel is calculated from the red-green-blue representation of the image pixel by transforming the red-green-blue space to make the brown DAB component a separate color and by taking the logarithm of the brown component. Fig.13 (Upper right image) Shown is the image analysis script used to generate the segmented image. Fig.13 (Lower right image) shows the derived measurements for all cell membrane objects and cytoplasm objects in image 18.

[0514] In another embodiment, the cell membrane of each tumor cell is detected by the single instance segmentation method "StarDist" as described in [https: / / squidpy.readthedocs.io / en / stable / external_tutorials / tutorial_stard ist.html]. This method approximates the cell membrane as a closed polygon with N star-shaped support vectors originating from the cell center. The network predicts N+1 posterior layers, and layer N+1 associates each pixel value with the probability that this pixel belongs to the cell center. In each of the N posterior layers, the distance of the cell membrane to the cell center pixel is predicted. The average optical density of each cell membrane is calculated by iterating the predicted closed polygon and measuring the optical density of the associated pixels that intersect the polygon.

[0515] B. Calculation of Predictive Efficacy Score

[0516] Based on the optical density of the DAB staining within the cell membrane object and optionally the cytoplasm object, a spatial proximity score is determined for the tissue sample shown in the digital image 18. The spatial proximity score also takes into account the staining intensity of the DAB dye in the cell membrane object and optionally the cytoplasm object of the neighboring cancer cells whose distance from the cancer cell for which the single cell ADC score is being calculated is less than a predefined distance. There are two types of spatial proximity scores: binary spatial proximity scores (bSPS) and continuous spatial proximity scores (cSPS).

[0517] Fig.14 An embodiment showing how to determine a binary spatial proximity score (bSPS) is a percentage score that is equal to the sum of (i) the number of cancer cells in the digital image 18 whose staining intensity equals or exceeds the optical density threshold plus (ii) the number of cancer cells in the digital image whose staining intensity is less than the optical density threshold but within a predefined distance of at least one cancer cell whose staining intensity equals or exceeds the optical density threshold, divided by the total number of cancer cells in the digital image 18.

[0518] Fig.14 Ten exemplary cancer cells (also called tumor cells) with different staining optical densities from the image analysis of steps 12-13 are shown. Figure 10-12 The heuristic image analysis steps shown in are used to obtain segmentation of image objects including cell membrane objects and cancer cell objects. Circles with solid lines on the circumference represent tumor cells whose optical density of brown DAB staining (corresponding to the amount of target protein FRα) is greater than or equal to a predetermined optical density threshold. In this embodiment, the optical density threshold is 12 from the maximum scale of the optical density of 220. Fig.14 Three cancer cells 19-21 in the figure are classified as optical density positive. The circle with a dotted circumference represents a tumor cell whose optical density of brown DAB staining is less than a predetermined optical density threshold. Seven cancer cells 22-28 are classified as optical density negative. The dark gray circle with a dotted circumference represents an optical density negative cancer cell located within a predefined distance of at least one optical density positive cancer cell. Both cancer cells 22-23 are optical density negative and are located within a predefined distance from an optical density positive cancer cell. In this embodiment, the predefined distance is 25 microns.

[0519] In this embodiment, the binary spatial proximity score is equal to a percentile score of 50%, which is Fig.14 The calculation shown in is the sum of three optical density positive cells and two optical density negative cells within a predefined distance of the optical density positive cells, divided by the total number of ten cancer cells ([3+2] / 10=0.5).

[0520] exist Fig.14 In the schematic image of FIG. 1 , three cancer cells 19-21 express a large amount of the target protein FRα and are likely to be killed by the ADC payload (e.g., cytotoxin) that enters the cell linked to the ADC antibody ( Fig.15 Effect 1). The two cancer cells 22-23 do not express sufficient amounts of the target protein FRα and therefore cannot be killed directly by the anti-FRα ADC. However, due to the proximity of at least one of the optical density-positive cells 19-21, the toxic payload released from cells 19-21 will also kill cancer cells 22-23 ( Fig.15 The remaining optical density-negative cancer cells 24-28 will remain active and may become the origin of resistance mechanisms, which may eventually lead to the death of the patient.

[0521] Fig.15 The mechanism of anti-FRαADC therapy killing cancer cells is shown. In the first step, the ADC antibody binds to the target protein FRα and inhibits the natural function of the target protein, which can cause cell death. In the second step, the payload (e.g., type I topoisomerase inhibitor) is internalized into the cell and kills the cell by the toxicity of the payload. This uptake of the payload depends on the amount of the target protein on the cell membrane, and also depends on the difference in the amount of the target protein on the cell membrane and in the cytoplasm. After uptake, the payload can be released from the cell into the surrounding tissue. In the third step, the payload can enter nearby cells and can also kill them. The spatial distribution of the payload in the tissue is diffused by passive diffusion.

[0522] Traditional IHC scores reflect both the inhibitory effect of ADC binding on the target protein and the effect of the cytotoxic payload entering the cancer cell along with the ADC antibody. Therefore, traditional scores for ADC therapeutics cannot reflect the importance of the presence of the target protein in the cytoplasm and the impact of the cytotoxic payload that diffuses into the tissue after being released from the first killed cancer cell. In contrast, the new predictive spatial proximity score measures the impact of the release of the cytotoxic payload on neighboring cancer cells.

[0523] Figure 16-18 An example of how to determine a continuous spatial proximity score is shown. Although the continuous spatial proximity score applies a predefined distance, optical density negative cancer cells are still given a certain weight even though their distance from optical density positive cancer cells exceeds the predefined distance. In addition, this embodiment of the continuous spatial proximity score also takes into account staining in the cytoplasm. The continuous spatial proximity score is calculated for the tissue sample shown in the digital image 18 based on the optical density of the DAB staining within the cell membrane object and the optional cytoplasm object.

[0524] Fig.16An exemplary quantification result of the optical density of the staining from the image analysis of steps 12-13 is shown in a schematic diagram using the grayscale values ​​of the cell membrane and cytoplasm pixels. Figure 10-12 The steps of heuristic image analysis shown in are used to obtain Fig.16 Example segmentation of the nucleus, cell membrane and cytoplasm in . Fig.16 The bright gray values ​​in the graph are associated with high DAB optical density, and therefore with a large amount of protein targeted by the diagnostic antibody. Dark gray values ​​are associated with low DAB optical density. Brighter pixels correspond to higher DAB optical density.

[0525] Fig.17 List Fig.16 The images are an example of quantitative staining on the cell membrane and in the cytoplasm. Fig.17 The optical density of the brown DAB signal from the cell membrane of the first, second and third cells is 0.949, 0.369 and 0.498, respectively. In this example, the optical density is expressed as a maximum staining percentage of 255. The optical density of the brown DAB signal from the cytoplasm of the first, second and third cells is 0.796, 0.533 and 0.369, respectively. Fig.17 In the schematic image of FIG. 1 , the first cancer cell 29 expresses a large amount of the target protein FRα and is likely to be killed by the ADC payload (e.g., cytotoxin) that enters the cell linked to the ADC antibody ( Fig.15 The second cancer cell 30 and the third cancer cell 31 do not express sufficient amounts of the target protein FRα and therefore cannot be directly killed by the anti-FRα ADC. However, since the second cancer cell 30 is adjacent to the first cancer cell 29, the toxic payload released from the first cancer cell will also kill the second cancer cell 30 ( Fig.15 Effect 3). The third cancer cell 31 will remain active and may become the origin of drug resistance mechanisms, which may eventually lead to the death of the patient.

[0526] exist Fig. 9 In step 14 of the method, the average optical density of each cancer cell is determined. Then in a modification of method 10, a single cell spatial proximity score is determined. Fig.18 Show Fig.17 The single cell spatial proximity score is calculated for each of the three cells shown in , and an exponential weighting factor for cell separation is incorporated to reflect the uptake of the ADC payload (e.g., cytotoxin) into neighboring cells. Single cell scores can be based on Fig.18 The predefined distance used in the formula is 50 microns. Fig.17 The optical densities of the DAB signals from the cell membrane (0.949, 0.369, and 0.498) and from the cytoplasm (0.796, 0.533, and 0.369) listed in Fig.18 In the calculation shown in , the single cell spatial proximity scores of the first, second and third cells 29, 30, 31 are 0.145, 0.012 and 0.064, respectively.

[0527] Thus, continuous spatial proximity scoring combines the amount of target protein on the cell membrane using DAB optical density measurement and optionally an estimate of the ADC payload (e.g., cytotoxin) uptake. Fig.17 As shown, the uptake of the ADC payload (e.g., cytotoxin) of a first cell depends on the amount of dye in its cell membrane and in its cytoplasm, as well as the amount of dye in the cell membrane and cytoplasm of a second cell in the vicinity of the first cell. More specifically, the vicinity can be a disk with a predefined radius surrounding the first cancer cell. In one embodiment, the continuous spatial proximity score of the first cancer cell is determined by the distance-weighted sum of several powers of the DAB optical density of the cell membrane and cytoplasm objects whose associated cancer cell centers are closer to the first cancer cell center than the predefined distance. In one embodiment, the predefined distance is 50 μm, such as Fig.18 In another embodiment, the distance is 20 μm. In another embodiment, the distance is determined by minimizing the KaplanMeier log-rank p-value of the spatial proximity score positive group compared to the spatial proximity score negative group in a predefined training group of patients treated with ADC. In another embodiment, the distance weighting involves calculating the exponent of the scaled negative Euclidean distance from the center of the first cancer cell to the center of the other cancer cells in the sum. In another embodiment, the power in the sum is limited to 0, 1, and 2 (constant, linear term, square).

[0528] C. Validation of the binary spatial proximity score

[0529] according to Fig. 9The accuracy of the predicted ADC score generated by the new method and its variants is based on the patient-derived xenograft (PDX) study of patients with different FRα expression levels, and is associated with tumor growth as an indicator of response to ADC. Tumor tissue fragments were subcutaneously implanted in female athymic nude mice aged 6 to 8 weeks. When the tumor reached the appropriate tumor volume range (usually 150-300mm), the animals were randomly divided into treatment and control groups and began to be administered with anti-FRαADC AB1370049-SG3932-DAR8. Doses of 2.5mg / kg and 5mg / kg were used. The human equivalent dose (HED) was 0.2mg / kg and 0.4mg.kg, respectively. A single dose of anti-FRαADC was administered to tumor-bearing mice via intravenous injection. Mice were observed daily, and tumor size and body weight were measured and recorded twice a week. Tumor volume was measured with a digital caliper and calculated using the following formula: Tumor volume = [length (mm) x width (mm) x 0.52, where length and width are the longest and shortest diameters of the tumor, respectively.

[0530] In step 15 of method 10, a predictive efficacy score in the form of a binary spatial proximity score (bystander_memb(meanOD)_binary_r25_cut12) for the tissue sample is generated based on the percentage of cancer cells in the digital image that are optical density (OD) positive or optical density negative, but within a predefined distance of an optical density positive cancer cell. In this embodiment, the predefined distance is 25 microns. If the mean optical density is greater than or equal to an optical density threshold, each cancer cell is identified as optical density positive, and if the mean optical density is less than the optical density threshold, each cancer cell is identified as optical density negative. In this embodiment, the optical density threshold is 12 on a scale with a maximum optical density of 220.

[0531] In step 15 of method 10, if the binary spatial proximity score exceeds a predetermined threshold, the cancer patient is identified as one who will likely benefit from administration of the ADC. In this embodiment, the predetermined threshold is ninety-eight percentile. The predefined distance, optical density threshold, and predetermined threshold are associated with the responses of training patients treated with the ADC.

[0532] Fig.19 It is illustrated how the predefined distance, optical density threshold, and predetermined threshold are correlated to the response of tissue samples from cancer patients treated with an anti-FRα ADC. Fig.19 Correlation between actual results from PDX studies of 112 patients with ovarian, NSCLC, colon, and endometrial cancers is shown. Fig.19 The tissues analyzed in the study were from mice bearing tumor cells from 112 human cancer patients. A dose of 5 mg / kg was administered in the PDX study. Fig.19The relationship between tumor growth rate (e.g., 1.5 = 150%) and the predicted efficacy score: bystander_memb(mean OD)_binary_r25_cut12 (the percentage of cancer cells that are (i) optical density (OD) positive or (ii) optical density (OD) negative and within 25 microns of an OD positive cell) is shown. Y-axis values ​​greater than 0 indicate that the tumor increased in size during the observation period, and values ​​less than 0 indicate that the tumor decreased in size during the observation period.

[0533] Fig.19 A negative correlation was shown between the predicted efficacy score bystander_memb(mean OD)_binary_r25_cut12 and the change in tumor size after treatment. Fig.19 In the lower graph, Spearman's rho is -0.57 and p equals 5.77E-8. Therefore, a greater reduction in tumor size is associated with a higher predicted efficacy score. Fig.19 The upper bar graph shows the predictive efficacy score in descending order of size for each of the 112 PDX models of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0534] Fig. 20 is a graph showing the predictive efficacy scores in ascending order of size for 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer. The predetermined threshold of 98% used in step 16 of method 10 is determined based on the 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. The circled shapes represent PDX models that are not responsive to the ADC, and the shapes without circles represent PDX models that show a positive response to the ADC.

[0535] Fig.21 It shows how to stratify patients to determine the optimal predetermined threshold between responding and non-responding patients. Fig.21It is a receiver operating characteristic (ROC) curve for evaluating the discriminative performance of a predictor (in this case, a predictive efficacy score). The "R package cutpointr" ROC curve analysis tool was used to determine the optimal predetermined threshold, as well as predictive performance parameters such as AUC (area under the curve), Youden index value, sensitivity, and specificity. AUC values ​​closer to 1 indicate better discriminative performance when stratifying patients between binary clinical outcomes such as response (R) and non-response (NR). Sensitivity is the true positive rate and is equal to the number of true positives divided by the sum of true positives and false negatives [TP / (TP+FN)]. Sensitivity indicates the probability that a positive score actually corresponds to a positive clinical outcome. Specificity is the false positive rate and is equal to the number of true negatives divided by the sum of true negatives and false positives [TN / (TN+FP)]. Specificity indicates the probability that a positive score actually corresponds to a negative clinical outcome. Youden index value is equal to sensitivity plus specificity minus 1.

[0536] For the predicted efficacy score bystander_memb(mean OD)_binary_r25_cut12, Fig.21 It was shown that the optimal predetermined threshold was 98%, and the area under the curve (AUC) was 0.8235. The sensitivity was 0.8824, the specificity was 0.8182, and the Youden index value was 0.7005.

[0537] Fig. 22 The distribution of values ​​for the predictive efficacy score bystander_memb(mean OD)_binary_r25_cut12 in 197 tissue samples from a database of ovarian cancer patients is shown. The analysis was performed on stained tissue samples from human patients, not tissues from PDX models. The bar graph with decreasing score values ​​indicates the prevalence of different ranges of predictive efficacy scores (binary spatial proximity scores) in ovarian cancer patients. One quarter of the patients had a score of 60.11% or lower. The average score was 77.69%. Fig. 22 It was shown that tissue samples from more than half (109) of the 197 ovarian cancer patients had a bystander_memb(mean OD)_binary_r25_cut12 score of 98% or higher. These ovarian cancer patients are likely to have a positive response to therapy involving anti-FRα antibody-drug conjugates.

[0538] exist Fig. 9 In step 17 of method 10 of , if the predicted efficacy score exceeds a predetermined threshold of 98%, a therapy involving an anti-FRα ADC is recommended to the cancer patient. Thus, anti-FRα ADC therapy will be recommended to 109 of the 197 ovarian cancer patients.

[0539] D. Verification of the percentage of cells with the minimum OD

[0540] The results were also validated by treating tumors in PDX models with 2.5 mg / kg and 5 mg / kg doses of the anti-FRα ADC AB1370049-SG3932-DAR8. Fig. 9 Changes to the new method. Fig.23 is a flow chart of steps 33-38 of method 32 for predicting a cancer patient's response to an anti-FRα ADC. Method 32 is Fig. 9 Another embodiment of method 10 in , and predicting the efficacy of a therapy involving an anti-FRα ADC based on the percentage of cancer cells that exhibit the least amount of staining. In step 33, a digital image of a tissue section from a cancer patient is acquired, the tissue section having been stained with a dye attached to a diagnostic antibody that targets the same protein as the protein targeted by the ADC therapy, in this case, the FRα protein. In step 34, image analysis is performed on the digital image to generate an image object of the cancer cell and the cell membrane. In step 35, the average optical density (OD) of the staining of the dye in the cell membrane of each cancer cell is determined.

[0541] In step 36, each cancer cell is identified as optical density positive or optical density negative based on the average optical density of the staining of the cell membrane. If the average optical density is greater than or equal to the optical density threshold, the cancer cell is identified as optical density positive. If the average optical density is less than the optical density threshold, the cancer cell is optical density negative. In this embodiment, the optical density threshold is 12 on a scale with a maximum optical density of 220.

[0542] In step 37, a predicted efficacy score for the tissue sample is generated based on the percentage of cancer cells that are optical density positive in the digital image. If the percentage of cancer cells that are optical density positive is greater than or equal to a percentage threshold, the predicted efficacy score is positive. If the percentage of cancer cells that are optical density positive is less than the percentage threshold, the predicted efficacy score is negative. The optical density threshold and the percentage threshold are associated with the response of training patients treated with ADC. In this embodiment, the percentage threshold is ninety percent.

[0543] Fig.24 It is illustrated how the optical density threshold and percentage threshold are correlated with the response of tissue samples from cancer patients treated with anti-FRα ADC. Fig.24 Correlation between actual results from PDX studies of 112 patients with ovarian, NSCLC, colon, and endometrial cancers is shown. Fig.24 The tissues analyzed in the study were from mice bearing tumor cells from 112 human cancer patients. A dose of 5 mg / kg was administered in the PDX study. Fig.24The relationship between percent tumor growth (1 = 100%) and the predicted efficacy score memb(mean OD)_perc_pos_12 (the percentage of cancer cells stained with a mean optical density of at least 12 on a scale of 220) is shown. Y-axis values ​​greater than 0 indicate that the tumor increased in size during the observation period, and values ​​less than 0 indicate that the tumor decreased in size during the observation period.

[0544] Fig.24 A negative correlation was shown between the predicted efficacy score memb(mean OD)_perc_pos_12 and the change in tumor size after treatment. Fig.24 In the lower panel, Spearman's rho is -0.54 and p equals 3.36E-7. Therefore, a greater reduction in tumor size is associated with a higher predicted efficacy score. Fig.24 The upper bar graph shows the predictive efficacy score for each of the 112 PDX models of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer, sorted by decreasing size.

[0545] Fig.25 is a graph showing the predictive efficacy scores for 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer, sorted in ascending order by size. The percentage threshold of 90% used in step 37 of method 32 is determined based on 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. The circled shapes represent PDX models that are not responsive to the ADC, and the shapes without circles represent PDX models that show a positive response to the ADC.

[0546] Fig.26 It shows how to stratify patients to determine the optimal percentage threshold between responding and non-responding patients. Fig.26 is a receiver operating characteristic (ROC) curve used to evaluate the discriminative performance of a predictor (in this case, the memb(mean OD)_perc_pos_12 score). Fig. 20 The same ROC curve analysis tool was used to determine Fig.25 For the predictive efficacy score memb(mean OD)_perc_pos_12, Fig.26 It was shown that the optimal percentage threshold was 90%, and the area under the curve (AUC) was 0.7968. The sensitivity was 0.8235, the specificity was 0.8182, and the Youden index value was 0.6417.

[0547] Fig. 27The distribution of values ​​for the predictive efficacy score memb(mean OD)_perc_pos_12 in 197 tissue samples from a database of ovarian cancer patients is shown. Fig. 27 The analysis was performed on stained tissue samples from human patients, not tissue from PDX models. The bar graph with decreasing score values ​​shows the prevalence of different ranges of predictive efficacy scores in ovarian cancer patients. A quarter of the patients had a score of 26.08% or lower. The average score was 66.75%. Fig. 27 It was shown that tissue samples from slightly less than half (96) of the 197 ovarian cancer patients had a memb(mean OD)_perc_pos_12 score of 90% or higher. These ovarian cancer patients are likely to have a positive response to therapy involving anti-FRα antibody-drug conjugates.

[0548] exist Fig.23 In step 38 of method 32 of , if the predicted efficacy score is positive, a therapy involving an anti-FRα ADC is recommended to the cancer patient. If the memb(mean OD)_perc_pos_12 score is 90% or higher, the score is positive. Therefore, anti-FRα ADC therapy will be recommended to slightly less than half of the 197 ovarian cancer patients.

[0549] E. Verification of the median absolute deviation of OD

[0550] Fig. 9 Another embodiment of method 10 involves determining the median of the absolute deviations of the staining of each cell in the digital image from the median cell membrane staining of all cancer cells. Fig.28 4 is a flow chart of steps 41-47 of a method 40 for predicting a cancer patient's response to an anti-FRα ADC. Method 40 predicts the efficacy of a therapy involving an anti-FRα ADC based on the absolute deviation of cell membrane staining in a digital image. In step 41, a digital image of a tissue section from a cancer patient is acquired, the tissue section having been stained with a dye attached to a diagnostic antibody that targets the FRα protein. In step 42, an image analysis is performed on the digital image to generate an image object of a cancer cell and a cell membrane. In step 43, the average optical density (OD) of the staining of the dye in the cell membrane of each cancer cell is determined. In step 44, the median optical density of all cancer cells in the digital image is determined.

[0551] In step 45, the median absolute deviation of the optical density of cancer cells in the digital image from the median optical density of all cancer cells is determined. In step 46, a predicted efficacy score for the tissue sample is generated based on the median absolute deviation. If the median absolute deviation is equal to or greater than the deviation threshold, the predicted efficacy score is positive, and if the median absolute deviation is less than the deviation threshold, the predicted efficacy score is negative. The deviation threshold is associated with the response of a training patient cohort treated with ADC. In this embodiment, the deviation threshold is twenty-four percent.

[0552] Fig.29 Illustrate how the optical density threshold and the deviation threshold are associated with the response of tissue samples from cancer patients treated with anti-FRα ADC. Fig.29 Show the correlation between the actual results of a PDX study of 112 patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer. Fig.29 The tissues analyzed were from mice carrying tumor cells from 112 human cancer patients. A dose of 5 mg / kg was administered in the PDX study. Fig.29 Show the relationship between the percentage of tumor growth (1 = 100%) and the predicted efficacy score hist_memb(mean OD)_mad (the median absolute deviation of the optical density of cancer cells in the digital image from the median optical density of all cancer cells). Y-axis values greater than 0 indicate an increase in tumor size during the observation period, and values less than 0 indicate a decrease in tumor size during the observation period.

[0553] Fig.29 Show that there is a negative correlation between the predicted efficacy score hist_memb(mean OD)_mad and the change in tumor size after treatment. In Fig.29 the lower graph below, Spearman's rho is -0.60, and p is equal to 7.03E-9. Thus, a larger decrease in tumor size is associated with a higher predicted efficacy score. Fig.29 The upper bar graph above shows the predicted efficacy scores sorted in decreasing order of size for each of the 112 PDX models of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer.

[0554] Fig.30 is a graph showing the predicted efficacy scores sorted in increasing order of size for 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer. The deviation threshold of 24% used in step 46 of method 40 was determined based on the 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. Shapes with circles represent PDX models that are non-responsive to ADC, and shapes without circles represent PDX models that show a positive response to ADC.

[0555] Fig.31 It shows how to stratify patients to determine the optimal bias threshold between responding and non-responding patients. Fig.31 is a receiver operating characteristic (ROC) curve used to evaluate the discriminative performance of a predictor (in this case, hist_memb(mean OD)_mad score). Fig. 20 The same ROC curve analysis tool was used to determine Fig.30 For the prediction efficacy score hist_memb(mean OD)_mad, Fig.31 It was shown that the optimal deviation threshold was 24%, and the area under the curve (AUC) was 0.8128. The sensitivity was 0.7059, the specificity was 0.9091, and the Youden index value was 0.6150.

[0556] Fig.32 The distribution of values ​​for the predictive efficacy score hist_memb (mean OD)_mad in 197 tissue samples from a database of ovarian cancer patients is shown. Fig.32 The analysis was performed on stained tissue samples from human patients, rather than on Fig.30 The tissues of the PDX models shown in were analyzed. The bar chart with decreasing score values ​​shows the prevalence of different ranges of predictive efficacy scores in ovarian cancer patients. A quarter of the patients had a score of 3.20% or lower. A quarter of the patients had a score of 44.59% or higher. The average score was 25.90%. Fig.32 It was shown that tissue samples from just over half (103) of the 197 ovarian cancer patients had a hist_memb (mean OD)_mad score of 24% or higher. These ovarian cancer patients are likely to have a positive response to therapy involving anti-FRα antibody-drug conjugates.

[0557] exist Fig.28 In step 47 of method 40 of , if the predicted efficacy score is positive, a therapy involving an anti-FRα ADC is recommended to the cancer patient. If the hist_memb (mean OD)_mad score is 24% or higher, the score is positive. Thus, anti-FRα ADC therapy will be recommended to just over half of the 197 ovarian cancer patients.

[0558] F. Validation of mean OD for predicting dose efficacy

[0559] Fig. 9 Another embodiment of method 10 involves determining the median optical density of all cancer cells in the digital image to determine the recommended dose of the anti-FRα ADC. Fig.3349-54 is a flow chart of steps 49-54 of method 48, which is used to predict the efficacy of various doses of anti-FRα ADC based on the median optical density of cell membrane staining in a digital image. In step 49, a digital image of a tissue section from a cancer patient is acquired, the tissue section having been stained with a dye attached to a diagnostic antibody that targets the FRα protein. In step 50, image analysis is performed on the digital image to generate an image object of cancer cells and cell membranes. In step 51, the average optical density (OD) of the staining of the dye in the cell membrane of each cancer cell is determined. In step 52, the median optical density of all cancer cells in the digital image is determined. This is equivalent to determining the optical density of the 50th percentile (q50) of cancer cells in the digital image, which is a whole slice image (WSI).

[0560] In step 53, the recommended dose of the anti-FRα ADC is determined based on whether the median optical density (i) falls below the lower optical density threshold, (ii) falls between the lower optical density threshold and the upper optical density threshold, or (iii) falls above the upper optical density threshold. If the median optical density falls below the lower optical density threshold, the recommended dose is zero. Therefore, if the median optical density is below the lower optical density threshold, the ADC is not administered to the cancer patient. If the median optical density falls between the lower optical density threshold and the upper optical density threshold, the recommended dose is a higher dose. If the median optical density falls above the upper optical density threshold, the recommended dose is a lower dose. The lower optical density threshold and the upper optical density threshold are associated with the response of a training patient group treated with the ADC. In this embodiment, the lower dose is 2.5 mg / kg and the higher dose is 5 mg / kg. In this embodiment, the lower optical density threshold is 25 and the upper optical density threshold is 39, both on a scale with a maximum optical density of 220.

[0561] Figure 34-39 It is illustrated how the lower and upper optical density thresholds correlate with the response of tissue samples from cancer patients treated with an anti-FRα ADC. Fig.34 Correlation between actual results of PDX studies of 28 patients with ovarian cancer, NSCLC, and endometrial cancer is shown. Fig.34 The tissues analyzed in the study were from mice bearing tumor cells from 28 human cancer patients. A dose of 5 mg / kg was administered in the PDX study. Fig.34 The relationship between the percentage of tumor growth and the predicted efficacy score double_dose_memb(mean OD)_q50 (mean OD of cell membrane staining) is shown. Y-axis values ​​greater than 0 indicate that the tumor increased in size during the observation period, while values ​​less than 0 indicate that the tumor size decreased during the observation period.

[0562] Fig.34A negative correlation was shown between the predicted efficacy score double_dose_memb(mean OD)_q50 and the change in tumor size after treatment. A greater reduction in tumor size was associated with a higher predicted efficacy score.

[0563] Fig.35 Is displayed for Fig.34 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer, shown in , are plotted in ascending order of predicted efficacy scores. The lower optical density threshold of 25 used in step 53 of method 48 is determined based on 28 PDX models treated with 5 mg / kg anti-FRα ADC. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. The circled shapes represent PDX models that are unresponsive to the ADC, and the shapes without circles represent PDX models that show a positive response to the ADC.

[0564] Fig.36 It shows how to stratify patients to determine the optimal lower optical density threshold between responding and non-responding patients. Fig.36 is a receiver operating characteristic (ROC) curve used to evaluate the discriminative performance of a predictor (in this case, the double_dose_memb(mean OD)_q50 score). For the predicted efficacy score of the PDX model treated with 5 mg / kg anti-FRα ADC, Fig.36 The optimal lower optical density threshold was shown to be optical density 25, and the area under the curve (AUC) was 0.7914. The sensitivity was 0.8824, the specificity was 0.7273, and the Youden index value was 0.6096.

[0565] Fig.37 Demonstrating how the upper optical density threshold correlates with the response of tissue samples from cancer patients treated with anti-FRα ADCs. Fig.37 Correlation between actual results of a PDX study of 23 patients with ovarian cancer, NSCLC, and endometrial cancer is shown. A dose of 2.5 mg / kg was administered in the PDX study. Fig.37 The tissues analyzed were from mice bearing tumor cells from 23 human cancer patients. Fig.37 Shown is the relationship between the predicted efficacy score double_dose_memb(mean OD)_q50 and the percentage of tumor growth after administration of a 2.5 mg / kg dose. Fig.37 A negative correlation was shown between the predicted efficacy score double_dose_memb(mean OD)_q50 and the change in tumor size after treatment. A greater reduction in tumor size was associated with a higher predicted efficacy score.

[0566] Fig.38 Is displayed for Fig.37 A graph of predicted efficacy scores for 23 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer, sorted by increasing size, is shown in The upper optical density threshold 39 used in step 53 of method 48 was determined based on the 23 PDX models treated with 2.5 mg / kg of anti-FRα ADC.

[0567] Fig.39 It shows how to stratify patients to determine the optimal upper optical density threshold between responding and non-responding patients. Fig.39 _ is a ROC curve for evaluating the discriminative performance of the double_dose_memb(mean OD)_q50 predictive efficacy score. For the predictive efficacy score of the PDX model treated with 2.5 mg / kg anti-FRα ADC, Fig.39 The optimal upper optical density threshold was shown to be optical density 39, and the area under the curve (AUC) was 0.8730. The sensitivity was 0.7778, the specificity was 0.9286, and the Youden index value was 0.7063.

[0568] Fig.40 is a graph of the median optical density of each model in multiple PDX models compared to the median optical density of human ovarian tumors. One curve shows the median optical density of cell membrane staining for 197 human ovarian tumors and includes a dashed line marking the average of all median optical density scores. Another curve shows the median optical density of cell membrane staining for 18 PDX models and includes a dashed line at the average of the median optical density scores, which is only slightly to the left of the dashed line for human ovarian tumors. This figure demonstrates that the distribution of FRα protein in ovarian cancer PDX models is comparable to that in dissected human ovarian tumors, and that the distribution of FRα in patient-derived xenografts and human ovarian cancer tumors is similar. Therefore, PDX models can be used to determine thresholds for various predictive efficacy scores.

[0569] Fig.41 The distribution of values ​​for the predicted efficacy score double_dose_memb (mean OD)_q50 in 197 tissue samples from a database of ovarian cancer patients is shown. Fig.41 The analysis was performed on stained tissue samples from human patients, rather than on Figure 34-39 The tissue of the PDX model shown in was analyzed. The bar graph with decreasing score values ​​shows the prevalence of different ranges of predictive efficacy scores in ovarian cancer patients. A quarter of the patients had an optical density score equal to 8.74 or lower, and another quarter of the patients had an optical density score equal to 99.28 or higher. The average score was an optical density of 56.21. Fig.41It is shown that the double_dose_memb (mean OD)_q50 score of tissue samples from 101 of the 197 ovarian cancer patients is 39 or higher. These ovarian cancer patients are likely to have a positive response to a lower dose of 2.5 mg / kg of anti-FRα antibody-drug conjugate. Another 8 ovarian cancer patients are likely to have a positive response to a higher dose of 5 mg / kg of anti-FRα antibody-drug conjugate. Using only the lower dose of 2.5 mg / kg, these other 8 patients are likely to fail to obtain a positive response. Although 101 patients are also likely to have a positive response to a higher dose of 5 mg / kg ADC, these patients can avoid any harmful side effects of additional ADC doses, which is not necessary for these patients to obtain a positive response. For 88 patients with a predicted efficacy score of less than 25, the recommended dose is zero. Therapy involving anti-FRα ADC is not recommended for these 88 patients.

[0570] exist Fig.33 In step 54 of method 48 of the invention, a therapy involving a recommended dose of an anti-FRα ADC is recommended to a cancer patient based on whether the predicted efficacy score of the digital image of the patient's tissue sample is below a lower optical density threshold of 25, between the lower optical density threshold of 25 and an upper optical density threshold of 39, or above the upper optical density threshold of 39. If the predicted efficacy score is between 25 and 39, a higher dose of 5 mg / kg is recommended, and if the predicted efficacy score is above 39, a lower dose of 2.5 mg / kg is recommended. If the predicted efficacy score is below 25, the recommended dose is zero.

[0571] G. Validation of the difference between cell membrane OD and cytoplasm OD

[0572] Yet another embodiment of the novel method for generating a predictive efficacy score involves determining the difference between the optical density of cell membrane staining and the optical density of cytoplasmic staining. Fig.42Flowchart of steps 56-62 of method 55 for predicting the response of cancer patients to anti-FRα ADC. Method 55 predicts the efficacy of a therapy involving anti-FRα ADC by identifying the 85% quantile of the difference between the optical density of cell membrane staining and cytoplasmic staining. In step 56, a digital image of a tissue section from a cancer patient is acquired, the tissue section having been stained with a dye attached to a diagnostic antibody that targets the FRα protein. In step 57, image analysis is performed on the digital image to generate image objects of cancer cells, cytoplasm, and cell membrane. In step 58, the average optical density (OD) of the staining of the dye in the cell membrane of each cancer cell is determined. In step 59, the average optical density of the staining of the dye in the cytoplasm of each cancer cell is determined. In step 60, for each cancer cell in the digital image, the difference between the average optical density of the staining of the cell membrane and the average optical density of the staining of the cytoplasm is determined. In step 61, the 85% quantile of the difference between the optical density of the cell membrane staining and the optical density of the cytoplasm staining is identified from all cancer cells in the digital image. In step 62, if the 85% quantile of the optical density difference exceeds a predetermined difference threshold, a therapy involving an anti-FRα ADC is recommended to the cancer patient. The difference threshold is associated with the response of a training patient cohort treated with the ADC. In this embodiment, the difference threshold is 23.

[0573] Fig.43 It is demonstrated how the differential threshold correlates with the response of tissue samples from cancer patients treated with anti-FRα ADCs. Fig.43 Correlation between actual results of PDX studies of patients with ovarian cancer, NSCLC, colon cancer, and endometrial cancer is shown. Fig.43 The tissues analyzed in the PDX study were from mice bearing tumor cells from human cancer patients. A dose of 5 mg / kg was administered in the PDX study. Fig.43 The relationship between the percentage of tumor growth (e.g., 2.5 = 250%) and the predicted efficacy score memb(mean OD)-cyto(mean OD)-q85 (85% quantile of the difference between the OD of cell membrane staining and the OD of cytoplasmic staining) is shown. Y-axis values ​​greater than 0 indicate that the tumor increased in size during the observation period, and values ​​less than 0 indicate that the tumor decreased in size during the observation period. Fig.43 A negative correlation was shown between the predicted efficacy score memb(mean OD)-cyto(mean OD)_q85 and the change in tumor size after treatment. Fig.43 In the plot of , Spearman's rho is -0.62, and p equals 1.50E-9. Therefore, a greater reduction in tumor size is associated with a higher predicted efficacy score.

[0574] Fig.44is a graph showing the predictive efficacy scores for 28 PDX models of patients with ovarian cancer, NSCLC, and endometrial cancer, sorted by increasing size. The difference threshold 23 used in step 62 of method 55 is determined based on 28 PDX models. Circles represent PDX models of ovarian cancer, squares represent PDX models of lung cancer (NSCLC), and triangles represent PDX models of endometrial cancer. The circled shapes represent PDX models that are not responsive to the ADC, and the shapes without circles represent PDX models that show a positive response to the ADC.

[0575] Fig.45 It shows how to stratify patients to determine the optimal threshold for differentiation between responding and non-responding patients. Fig.45 : is a ROC curve used to evaluate the degree of discrimination of the predictive efficacy score memb(mean OD)-cyto(mean OD)_q85 between responding and non-responding patients. Fig.45 It was shown that the optimal difference threshold was 23, and the area under the curve (AUC) was 0.8289. The sensitivity was 0.7647, the specificity was 0.9091, and the Youden index value was 0.6738.

[0576] Fig.46 The distribution of values ​​for the predictive efficacy score memb(mean OD)-cyto(mean OD)_q85 in 197 tissue samples from a database of ovarian cancer patients is shown. Fig.46 The analysis was performed on stained tissue samples from human patients, rather than on Fig.44 The tissues of the PDX models shown in were analyzed. The bar graph with decreasing score values ​​shows the prevalence of different ranges of predictive efficacy scores in ovarian cancer patients. A quarter of the patients had a score of 5.01 or lower. A quarter of the patients had a score of 41.85 or higher. The average score was 24.48. Fig.46 It was shown that just over half (104) of the tissue samples from the 197 ovarian cancer patients had a memb(mean OD)-cyto(mean OD)_q85 score of 23 or higher. These ovarian cancer patients are likely to have a positive response to therapy involving anti-FRα antibody-drug conjugates.

[0577] exist Fig.42 In step 62 of method 55 of , if the 85% quantile of the difference between the cell membrane optical density and the cytoplasm optical density for the patient's tissue sample exceeds a predetermined difference threshold, then a therapy involving an anti-FRα ADC is recommended to the cancer patient. In this example, the predetermined difference threshold is 23. Therefore, anti-FRα ADC therapy will be recommended to 104 of the 197 ovarian cancer patients.

[0578] Although the invention has been described in conjunction with certain specific embodiments for purposes of instruction, the invention is not limited thereto. Various modifications, adaptations and combinations of the various features of the described embodiments may be implemented without departing from the scope of the invention as set forth in the claims.

[0579] XII. Preferred Embodiments of the Invention

[0580] Preferred embodiments of the present invention are described below and are referred to as Embodiments E1 to E32.

[0581] E1. A method for generating a recommended dose for treating a cancer patient with an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody or an antigen-binding fragment thereof that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising:

[0582] immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0583] acquiring a digital image of the tissue sample;

[0584] detecting cancer cells in the digital image;

[0585] For each cancer cell, calculating the average optical density of the cell membrane of the cancer cell stained by the dye;

[0586] calculating the median optical density of all cancer cells in the digital image;

[0587] determining the recommended dose based on whether the median optical density falls below a lower optical density threshold, falls between the lower optical density threshold and an upper optical density threshold, or falls above the upper optical density threshold, wherein the recommended dose is zero if the median optical density falls below the lower optical density threshold, the recommended dose is a higher dose if the median optical density falls between the lower optical density threshold and the upper optical density threshold, and the recommended dose is a lower dose if the median optical density falls above the upper optical density threshold, and wherein the lower optical density threshold and the upper optical density threshold are associated with responses of a training patient cohort treated with the ADC; and

[0588] A therapy involving the ADC and the recommended dose is recommended to the cancer patient.

[0589] E2. The method as described in E1, further comprising:

[0590] The recommended dose of the therapy involving the ADC is administered to the cancer patient.

[0591] E3. A method as described in E1, wherein the lower optical density threshold is in the range of 21 to 29 on a scale with a maximum optical density of 220.

[0592] E4. A method as described in any of E1-E3, wherein the upper optical density threshold is in the range of 35 to 43 on a scale with a maximum optical density of 220.

[0593] E5. The method of any one of E1-E4, wherein the lower dose is 2.5 mg / kg or less, and wherein the higher dose is 5 mg / kg or more.

[0594] E6. A method of generating a response score to predict a cancer patient's response to an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody or antigen-binding fragment thereof that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising:

[0595] immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0596] acquiring a digital image of the tissue sample;

[0597] detecting cancer cells in the digital image;

[0598] For each cancer cell, calculating the average optical density of the cell membrane of the cancer cell stained by the dye;

[0599] calculating the median optical density of all cancer cells in the digital image;

[0600] generating a response score for the tissue sample based on the median optical density, wherein the response score is positive if the median optical density is equal to or greater than an optical density threshold and the response score is negative if the median optical density is less than the optical density threshold, and wherein the optical density threshold is associated with a response of a training patient cohort treated with the ADC; and

[0601] If the response score is positive, a therapy involving the ADC is recommended to the cancer patient.

[0602] E7. A method of generating a response score to predict a cancer patient's response to an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising:

[0603] immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0604] acquiring a digital image of the tissue sample;

[0605] detecting cancer cells in the digital image;

[0606] For each cancer cell, determining the optical density of staining by the dye in the cell membrane of the cancer cell;

[0607] Identifying each cancer cell as optical density positive if the optical density of the cancer cell is equal to or greater than an optical density threshold, or identifying each cancer cell as optical density negative if the average optical density of the cancer cells is less than an optical density threshold;

[0608] generating a response score for the tissue sample based on the percentage of cancer cells that are optical density positive in the digital image, wherein the response score is positive if the percentage of cancer cells that are optical density positive is equal to or greater than a percentage threshold, and the response score is negative if the percentage of cancer cells that are optical density positive is less than the percentage threshold, and wherein the optical density threshold and the percentage threshold are associated with responses of a training patient cohort treated with the ADC; and

[0609] If the response score is positive, a therapy involving the ADC is recommended to the cancer patient.

[0610] E8. The method of E6 or E7, further comprising:

[0611] If the response score is positive, the therapy involving the ADC is administered to the cancer patient.

[0612] E9. A method as described in any of E6-E8, wherein the percentage threshold is in the range of 80% to 90%.

[0613] E10. A method of identifying a cancer patient for treatment with an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody targeting a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising:

[0614] immunohistochemically staining a tissue sample from the cancer patient using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0615] acquiring a digital image of the tissue sample;

[0616] detecting cancer cells in the digital image;

[0617] For each cancer cell, calculating the average optical density of the cell membrane of the cancer cell stained by the dye;

[0618] Identifying each cancer cell as optical density positive if the average optical density of the cancer cells is equal to or greater than an optical density threshold, or identifying each cancer cell as optical density negative if the average optical density of the cancer cells is less than an optical density threshold;

[0619] generating a proximity score for the tissue sample, the proximity score being equal to the percentage of cancer cells in the digital image that are either optical density positive or optical density negative but within a predefined distance of an optical density positive cancer cell; and

[0620] If the proximity score exceeds a predetermined percentage threshold, the cancer patient is identified as a patient who would likely benefit from administration of the ADC.

[0621] E11. The method as described in E10, further comprising:

[0622] When the proximity score exceeds the predetermined percentage threshold, a therapy involving the ADC is recommended to the cancer patient.

[0623] E12. The method as described in E10 or E11, further comprising:

[0624] If the proximity score exceeds the predetermined percentage threshold, administering the therapy involving the ADC to the cancer patient.

[0625] E13. A method as described in any of E10-E12, wherein the percentage threshold is in the range of 95% to 100%.

[0626] E14. A method as described in any of E7-E13, wherein the optical density threshold is in the range of 10 to 15 on a scale with a maximum optical density of 220.

[0627] E15. A method of generating a response score to predict a cancer patient's response to an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody targeting a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising:

[0628] immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample;

[0629] acquiring a digital image of the tissue sample;

[0630] detecting cancer cells in the digital image;

[0631] For each cancer cell, calculating the average optical density of the cell membrane of the cancer cell stained by the dye;

[0632] calculating the median optical density of all cancer cells in the digital image;

[0633] calculating a median absolute deviation of the optical density of the cancer cell in the digital image from the median optical density of all cancer cells;

[0634] generating a response score for the tissue sample based on the median absolute deviation, wherein the response score is positive if the median absolute deviation is equal to or greater than a deviation threshold, and the response score is negative if the median absolute deviation is less than the deviation threshold, and wherein the deviation threshold is associated with a response of a training patient cohort treated with the ADC; and

[0635] If the response score is positive, a therapy involving the ADC is recommended to the cancer patient.

[0636] E16. The method as described in E15, further comprising:

[0637] If the response score is positive, the therapy involving the ADC is administered to the cancer patient.

[0638] E17. A method as described in E15 or E16, wherein the deviation threshold is in the range of 10 to 15 on an optical density scale with a maximum value of 220.

[0639] E18. The method of any one of E7-E17, wherein detection of cancer cells involves detecting pixels belonging to the cell membrane for each cancer cell using the cell center determined for each cancer cell.

[0640] E19. The method of any one of E1-E18, wherein the staining intensity of each cell membrane is calculated based on the average optical density of brown diaminobenzidine (DAB) signal in the pixels of the cell membrane.

[0641] E20. The method of any one of E1-E19, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

[0642] E21. The method of any one of E1-E20, wherein the dye is 3,3'-diaminobenzidine (DAB).

[0643] E22. A method as described in any of E1-E21, wherein the cancer patient has a cancer selected from the group consisting of: ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer and malignant pleural mesothelioma.

[0644] E23. The method of any one of E1-E22, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

[0645] E24. The method of E23, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

[0646] E25. The method of any one of E1-E24, wherein the ADC payload is a cytotoxin.

[0647] E26. The method of E25, wherein the cytotoxin is a topoisomerase I inhibitor.

[0648] E27. The method of E26, wherein the topoisomerase I inhibitor is represented by the formula:

[0649]

[0650] and salts and solvates thereof, and wherein RL represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

[0651] E28. The method of any one of E1-E27, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is selected from:

[0652]

[0653] and / or

[0654]

[0655] E29. The method of any one of E1-E28, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is

[0656]

[0657] E30. The method of any one of E1-E29, wherein the ADC has a drug to antibody ratio (DAR) falling within a range selected from the group consisting of 1 to 20, 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10.

[0658] E31. The method of any one of E1-E30, wherein the ADC has a drug to antibody ratio (DAR) selected from the group consisting of: 4 and 8.

[0659] E32. The method of any one of E1-E31, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of:

[0660] (a) heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 5 (KASGLES), and light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT);

[0661] (b) a heavy chain CDR1 of SEQ ID NO:7 (SYAMS), a heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG), a heavy chain CDR3 of SEQ ID NO:9 (EMQQLAL DY), a light chain CDR1 of SEQ ID NO:10 (RASQGISNFLA), a light chain CDR2 of SEQ ID NO:11 (AASSLQS), and a light chain CDR3 of SEQ ID NO:12 (QQ YNSYPFT);

[0662] (c) a heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN), a heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS), a heavy chain CDR3 of SEQ ID NO: 15 (G VGRFDS), a light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 17 (KASSLES), and a light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT);

[0663] (d) a heavy chain CDR1 of SEQ ID NO: 19 (SYNMN), a heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG), a heavy chain CDR3 of SEQ ID NO: 21 (GMTT LTFDY), a light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA), a light chain CDR2 of SEQ ID NO: 23 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 24 (QQYISYPLT);

[0664] (e) a heavy chain CDR1 of SEQ ID NO:25 (SYSMN), a heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO:27 (GMTTL TFDY), a light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO:29 (AASSLQS), and a light chain CDR3 of SEQ ID NO:30 (QQYNSYPLT); and

[0665] (f) heavy chain CDR1 of SEQ ID NO:31 (SDSATWN), heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO:33 (G GAPFDY), light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO:35 (KASSLES), and light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

[0666] E33. The method of any one of E1-E32, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of:

[0667] (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38;

[0668] (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39 and a VL chain comprising the amino acid sequence of SEQ ID NO: 40;

[0669] (c) a VH chain comprising the amino acid sequence of SEQ ID NO:41 and a VL chain comprising the amino acid sequence of SEQ ID NO:42;

[0670] (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43 and a VL chain comprising the amino acid sequence of SEQ ID NO: 44;

[0671] (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45 and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and

[0672] (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47 and a VL chain comprising the amino acid sequence of SEQ ID NO:48.

[0673] E34. The method of any one of E1-E33, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of:

[0674] (a) the heavy chain amino acid sequence of SEQ ID NO:49 and the light chain amino acid sequence of SEQ ID NO:50;

[0675] (b) the heavy chain amino acid sequence of SEQ ID NO:51 and the light chain amino acid sequence of SEQ ID NO:52;

[0676] (c) the heavy chain amino acid sequence of SEQ ID NO:53 and the light chain amino acid sequence of SEQ ID NO:54;

[0677] (d) the heavy chain amino acid sequence of SEQ ID NO:55 and the light chain amino acid sequence of SEQ ID NO:56;

[0678] (e) the heavy chain amino acid sequence of SEQ ID NO:57 and the light chain amino acid sequence of SEQ ID NO:58; and

[0679] (f) the heavy chain amino acid sequence of SEQ ID NO:59 and the light chain amino acid sequence of SEQ ID NO:60.

[0680] E35. The method of any one of E1-E34,

[0681] wherein the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 of SEQ ID NO: 1 (SD SATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVK S), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGL ES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT),

[0682] Wherein the ADC payload is a topoisomerase I inhibitor

[0683]

[0684] and

[0685] wherein the ADC has a drug to antibody ratio (DAR) of 8.

[0686] E36. The method of E35, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37; and a VL chain comprising the amino acid sequence of SEQ ID NO: 38.

Claims

1. A method for determining a recommended dose for treating a cancer patient with an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody or an antigen-binding fragment thereof that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising: immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; For each cancer cell, determining the average optical density of the cell membrane of the cancer cell stained by the dye; determining a median optical density of all cancer cells in the digital image; determining the recommended dose based on whether the median optical density falls below a lower optical density threshold, falls between the lower optical density threshold and an upper optical density threshold, or falls above the upper optical density threshold, wherein the recommended dose is zero if the median optical density falls below the lower optical density threshold, the recommended dose is a higher dose if the median optical density falls between the lower optical density threshold and the upper optical density threshold, and the recommended dose is a lower dose if the median optical density falls above the upper optical density threshold, and wherein the lower optical density threshold and the upper optical density threshold are associated with responses of a training patient cohort treated with the ADC; as well as A therapy involving the ADC and the recommended dose is recommended to the cancer patient.

2. The method of claim 1, further comprising: The recommended dose of the therapy involving the ADC is administered to the cancer patient.

3. The method of claim 1, wherein the lower optical density threshold is in the range of 21 to 29 on a scale with a maximum optical density of 220.

4. The method of claim 1, wherein the upper optical density threshold is in the range of 35 to 43 on a scale with a maximum optical density of 220.

5. The method of claim 1, wherein the lower dose is in the range of 2.4 to 2.6 mg / kg, and wherein the higher dose is in the range of 4.9 to 5.1 mg / kg.

6. The method of claim 1, wherein the detection of cancer cells involves detecting pixels belonging to the cell membrane for each cancer cell using the cell center determined for each cancer cell.

7. The method of claim 1, wherein the staining intensity of each cell membrane is calculated based on the average optical density of brown diaminobenzidine (DAB) signals in pixels of the cell membrane.

8. The method of claim 1, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

9. The method of claim 1, wherein the dye is 3,3'-diaminobenzidine (DAB).

10. The method of claim 1, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

11. The method of claim 1, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

12. The method of claim 11, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

13. The method of claim 1, wherein the ADC payload is a cytotoxin.

14. The method of claim 13, wherein the cytotoxin is a topoisomerase I inhibitor.

15. The method of claim 14, wherein the topoisomerase I inhibitor is represented by the formula: and salts and solvates thereof, and wherein R L Represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

16. The method of claim 1, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is selected from:

17. The method of claim 1, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is 18. The method of claim 1, wherein the ADC has a drug to antibody ratio (DAR) falling within a range selected from the group consisting of 1 to 20, 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10.

19. The method of claim 1, wherein the ADC has a drug to antibody ratio (DAR) selected from the group consisting of: 4 and 8.

20. The method of claim 1, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 5 (KASGLES), and light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT); (b) a heavy chain CDR1 of SEQ ID NO:7 (SYAMS), a heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG), a heavy chain CDR3 of SEQ ID NO:9 (EMQQLAL DY), a light chain CDR1 of SEQ ID NO:10 (RASQGISNFLA), a light chain CDR2 of SEQ ID NO:11 (AASSLQS), and a light chain CDR3 of SEQ ID NO:12 (QQ YNSYPFT); (c) a heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN), a heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS), a heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS), a light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 17 (KASSLES), and a light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT); (d) a heavy chain CDR1 of SEQ ID NO: 19 (SYNMN), a heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG), a heavy chain CDR3 of SEQ ID NO: 21 (GMTT LTFDY), a light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA), a light chain CDR2 of SEQ ID NO: 23 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 24 (QQYISYPLT); (e) a heavy chain CDR1 of SEQ ID NO:25 (SYSMN), a heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO:27 (GMTTL TFDY), a light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO:29 (AASSLQS), and a light chain CDR3 of SEQ ID NO:30 (QQYNSYPLT); and (f) heavy chain CDR1 of SEQ ID NO:31 (SDSATWN), heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO:33 (GGAPFDY), light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO:35 (KASSLES), and light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

21. The method of claim 1, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39 and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO:41 and a VL chain comprising the amino acid sequence of SEQ ID NO:42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43 and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45 and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47 and a VL chain comprising the amino acid sequence of SEQ ID NO:

48.

22. The method of claim 1, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) the heavy chain amino acid sequence of SEQ ID NO:49 and the light chain amino acid sequence of SEQ ID NO:50; (b) the heavy chain amino acid sequence of SEQ ID NO:51 and the light chain amino acid sequence of SEQ ID NO:52; (c) the heavy chain amino acid sequence of SEQ ID NO:53 and the light chain amino acid sequence of SEQ ID NO:54; (d) the heavy chain amino acid sequence of SEQ ID NO:55 and the light chain amino acid sequence of SEQ ID NO:56; (e) the heavy chain amino acid sequence of SEQ ID NO:57 and the light chain amino acid sequence of SEQ ID NO:58; and (f) the heavy chain amino acid sequence of SEQ ID NO:59 and the light chain amino acid sequence of SEQ ID NO:

60.

23. The method of claim 1, wherein the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 of SEQ ID NO: 1 (SD SATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVK S), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGL ES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT), Wherein the ADC payload is a topoisomerase I inhibitor and wherein the ADC has a drug to antibody ratio (DAR) of 8.

24. The method of claim 23, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37; and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

25. A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody or antigen binding fragment thereof that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising: immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; For each cancer cell, determining the average optical density of the cell membrane of the cancer cell stained by the dye; determining a median optical density of all cancer cells in the digital image; generating a predictive efficacy score for the tissue sample based on the median optical density, wherein the predictive efficacy score is positive if the median optical density is equal to or greater than an optical density threshold and the predictive efficacy score is negative if the median optical density is less than the optical density threshold, and wherein the optical density threshold is associated with a response of a training patient cohort treated with the ADC; as well as If the predicted efficacy score is positive, a therapy involving the ADC is recommended to the cancer patient.

26. The method of claim 25, further comprising: When the predicted efficacy score is positive, administering the therapy involving the ADC to the cancer patient.

27. The method of claim 25, wherein the detection of cancer cells involves detecting pixels belonging to the cell membrane for each cancer cell using the cell center determined for each cancer cell.

28. The method of claim 25, wherein the staining intensity of each cell membrane is calculated based on the average optical density of brown diaminobenzidine (DAB) signals in pixels of the cell membrane.

29. The method of claim 25, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

30. The method of claim 25, wherein the dye is 3,3'-diaminobenzidine (DAB).

31. The method of claim 25, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

32. The method of claim 25, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

33. The method of claim 32, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

34. The method of claim 25, wherein the ADC payload is a cytotoxin.

35. The method of claim 34, wherein the cytotoxin is a topoisomerase I inhibitor.

36. The method of claim 35, wherein the topoisomerase I inhibitor is represented by the formula: and salts and solvates thereof, and wherein R L Represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

37. The method of claim 25, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is selected from:

38. The method of claim 25, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is 39. The method of claim 25, wherein the ADC has a drug to antibody ratio (DAR) falling within a range selected from the group consisting of 1 to 20, 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10.

40. The method of claim 25, wherein the ADC has a drug to antibody ratio (DAR) selected from the group consisting of: 4 and 8.

41. The method of claim 25, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 5 (KASGLES), and light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT); (b) a heavy chain CDR1 of SEQ ID NO:7 (SYAMS), a heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG), a heavy chain CDR3 of SEQ ID NO:9 (EMQQLAL DY), a light chain CDR1 of SEQ ID NO:10 (RASQGISNFLA), a light chain CDR2 of SEQ ID NO:11 (AASSLQS), and a light chain CDR3 of SEQ ID NO:12 (QQ YNSYPFT); (c) a heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN), a heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS), a heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS), a light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 17 (KASSLES), and a light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT); (d) a heavy chain CDR1 of SEQ ID NO: 19 (SYNMN), a heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG), a heavy chain CDR3 of SEQ ID NO: 21 (GMTT LTFDY), a light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA), a light chain CDR2 of SEQ ID NO: 23 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 24 (QQYISYPLT); (e) a heavy chain CDR1 of SEQ ID NO:25 (SYSMN), a heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO:27 (GMTTL TFDY), a light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO:29 (AASSLQS), and a light chain CDR3 of SEQ ID NO:30 (QQYNSYPLT); and (f) heavy chain CDR1 of SEQ ID NO:31 (SDSATWN), heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO:33 (GGAPFDY), light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO:35 (KASSLES), and light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

42. The method of claim 25, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39 and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO:41 and a VL chain comprising the amino acid sequence of SEQ ID NO:42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO:43 and a VL chain comprising the amino acid sequence of SEQ ID NO:44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45 and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47 and a VL chain comprising the amino acid sequence of SEQ ID NO:

48.

43. The method of claim 25, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) the heavy chain amino acid sequence of SEQ ID NO:49 and the light chain amino acid sequence of SEQ ID NO:50; (b) the heavy chain amino acid sequence of SEQ ID NO:51 and the light chain amino acid sequence of SEQ ID NO:52; (c) the heavy chain amino acid sequence of SEQ ID NO:53 and the light chain amino acid sequence of SEQ ID NO:54; (d) the heavy chain amino acid sequence of SEQ ID NO:55 and the light chain amino acid sequence of SEQ ID NO:56; (e) the heavy chain amino acid sequence of SEQ ID NO:57 and the light chain amino acid sequence of SEQ ID NO:58; and (f) the heavy chain amino acid sequence of SEQ ID NO:59 and the light chain amino acid sequence of SEQ ID NO:

60.

44. The method of claim 25, wherein the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 of SEQ ID NO: 1 (SD SATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVK S), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGL ES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT), Wherein the ADC payload is a topoisomerase I inhibitor and wherein the ADC has a drug to antibody ratio (DAR) of 8.

45. The method of claim 44, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37; and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

46. ​​A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising: immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; For each cancer cell, determining the optical density of staining by the dye in the cell membrane of the cancer cell; Identifying each cancer cell as optical density positive if the optical density of the cancer cell is equal to or greater than an optical density threshold, or identifying each cancer cell as optical density negative if the average optical density of the cancer cells is less than an optical density threshold; generating a predictive efficacy score for the tissue sample based on a percentage of cancer cells that are optical density positive in the digital image, wherein the predictive efficacy score is positive if the percentage of cancer cells that are optical density positive is equal to or greater than a percentage threshold, and the predictive efficacy score is negative if the percentage of cancer cells that are optical density positive is less than the percentage threshold, and wherein the optical density threshold and the percentage threshold are correlated with responses of a training patient cohort treated with the ADC; as well as If the predicted efficacy score is positive, a therapy involving the ADC is recommended to the cancer patient.

47. The method of claim 46, further comprising: When the predicted efficacy score is positive, administering the therapy involving the ADC to the cancer patient.

48. The method of claim 46, wherein the optical density threshold is in the range of 10 to 15 on a scale with a maximum optical density of 220.

49. The method of claim 46, wherein the percentage threshold is in the range of 80% to 90%.

50. The method of claim 46, wherein the detection of cancer cells involves detecting pixels belonging to the cell membrane for each cancer cell using the cell center determined for each cancer cell.

51. The method of claim 46, wherein the staining intensity of each cell membrane is calculated based on the average optical density of brown diaminobenzidine (DAB) signal in pixels of the cell membrane.

52. The method of claim 46, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

53. The method of claim 46, wherein the dye is 3,3'-diaminobenzidine (DAB).

54. The method of claim 46, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

55. The method of claim 46, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

56. The method of claim 55, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

57. The method of claim 46, wherein the ADC payload is a cytotoxin.

58. The method of claim 57, wherein the cytotoxin is a topoisomerase I inhibitor.

59. The method of claim 58, wherein the topoisomerase I inhibitor is represented by the formula: and salts and solvates thereof, and wherein R L Represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

60. The method of claim 46, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is selected from:

61. The method of claim 46, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is (SG3932).

62. The method of claim 46, wherein the ADC has a drug to antibody ratio (DAR) falling within a range selected from the group consisting of 1 to 20, 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10.

63. The method of claim 46, wherein the ADC has a drug to antibody ratio (DAR) selected from the group consisting of: 4 and 8.

64. The method of claim 46, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) a heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGLES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT); (b) a heavy chain CDR1 of SEQ ID NO:7 (SYAMS), a heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG), a heavy chain CDR3 of SEQ ID NO:9 (EMQQLALDY), a light chain CDR1 of SEQ ID NO:10 (RASQGISNFLA), a light chain CDR2 of SEQ ID NO:11 (AASSLQS), and a light chain CDR3 of SEQ ID NO:12 (QQYNSYPFT); (c) a heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN), a heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS), a heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS), a light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 17 (KASSLES), and a light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT); (d) a heavy chain CDR1 of SEQ ID NO: 19 (SYNMN), a heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG), a heavy chain CDR3 of SEQ ID NO: 21 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA), a light chain CDR2 of SEQ ID NO: 23 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 24 (QQYISYPLT); (e) a heavy chain CDR1 of SEQ ID NO:25 (SYSMN), a heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO:27 (GMTTLTFDY), a light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO:29 (AASSLQS), and a light chain CDR3 of SEQ ID NO:30 (QQYNSYPLT); and (f) heavy chain CDR1 of SEQ ID NO:31 (SDSATWN), heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO:33 (GGAPFDY), light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO:35 (KASSLES), and light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

65. The method of claim 46, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39 and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO:41 and a VL chain comprising the amino acid sequence of SEQ ID NO:42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43 and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45 and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47 and a VL chain comprising the amino acid sequence of SEQ ID NO:

48.

66. The method of claim 46, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) the heavy chain amino acid sequence of SEQ ID NO:49 and the light chain amino acid sequence of SEQ ID NO:50; (b) the heavy chain amino acid sequence of SEQ ID NO:51 and the light chain amino acid sequence of SEQ ID NO:52; (c) the heavy chain amino acid sequence of SEQ ID NO:53 and the light chain amino acid sequence of SEQ ID NO:54; (d) the heavy chain amino acid sequence of SEQ ID NO:55 and the light chain amino acid sequence of SEQ ID NO:56; (e) the heavy chain amino acid sequence of SEQ ID NO:57 and the light chain amino acid sequence of SEQ ID NO:58; and (f) the heavy chain amino acid sequence of SEQ ID NO:59 and the light chain amino acid sequence of SEQ ID NO:

60.

67. The method of claim 46, wherein the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 of SEQ ID NO: 1 (SD SATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVK S), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGL ES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT), Wherein the ADC payload is a topoisomerase I inhibitor and wherein the ADC has a drug to antibody ratio (DAR) of 8.

68. The method of claim 67, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37; and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

69. A method of identifying a cancer patient for treatment with an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising: immunohistochemically staining a tissue sample from the cancer patient using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; For each cancer cell, determining the average optical density of the cell membrane of the cancer cell stained by the dye; Identifying each cancer cell as optical density positive if the average optical density of the cancer cells is equal to or greater than an optical density threshold, or identifying each cancer cell as optical density negative if the average optical density of the cancer cells is less than an optical density threshold; generating a proximity score for the tissue sample, the proximity score being equal to the percentage of cancer cells in the digital image that are either optical density positive or optical density negative but within a predefined distance of an optical density positive cancer cell; and If the proximity score exceeds a predetermined percentage threshold, the cancer patient is identified as a patient who would likely benefit from administration of the ADC.

70. The method of claim 69, further comprising: When the proximity score exceeds the predetermined percentage threshold, a therapy involving the ADC is recommended to the cancer patient.

71. The method of claim 69, further comprising: When the proximity score exceeds the predetermined percentage threshold, the therapy involving the ADC is administered to the cancer patient.

72. The method of claim 69, wherein the optical density threshold is in the range of 10 to 15 on a scale with a maximum optical density of 220.

73. The method of claim 69, wherein the percentage threshold is in the range of 95% to 100%.

74. The method of claim 69, wherein the detection of cancer cells involves detecting pixels belonging to the cell membrane for each cancer cell using the cell center determined for each cancer cell.

75. The method of claim 69, wherein the staining intensity of each cell membrane is calculated based on the average optical density of brown diaminobenzidine (DAB) signal in pixels of the cell membrane.

76. The method of claim 69, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

77. The method of claim 69, wherein the dye is 3,3'-diaminobenzidine (DAB).

78. The method of claim 69, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

79. The method of claim 69, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

80. The method of claim 79, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

81. The method of claim 69, wherein the ADC payload is a cytotoxin.

82. The method of claim 81, wherein the cytotoxin is a topoisomerase I inhibitor.

83. The method of claim 82, wherein the topoisomerase I inhibitor is represented by the formula: and salts and solvates thereof, and wherein R L Represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

84. The method of claim 69, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is selected from:

85. The method of claim 69, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is 86. The method of claim 69, wherein the ADC has a drug to antibody ratio (DAR) falling within a range selected from the group consisting of 1 to 20, 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10.

87. The method of claim 69, wherein the ADC has a drug to antibody ratio (DAR) selected from the group consisting of: 4 and 8.

88. The method of claim 69, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 5 (KASGLES), and light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT); (b) a heavy chain CDR1 of SEQ ID NO:7 (SYAMS), a heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG), a heavy chain CDR3 of SEQ ID NO:9 (EMQQLAL DY), a light chain CDR1 of SEQ ID NO:10 (RASQGISNFLA), a light chain CDR2 of SEQ ID NO:11 (AASSLQS), and a light chain CDR3 of SEQ ID NO:12 (QQ YNSYPFT); (c) a heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN), a heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS), a heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS), a light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 17 (KASSLES), and a light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT); (d) a heavy chain CDR1 of SEQ ID NO: 19 (SYNMN), a heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG), a heavy chain CDR3 of SEQ ID NO: 21 (GMTT LTFDY), a light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA), a light chain CDR2 of SEQ ID NO: 23 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 24 (QQYISYPLT); (e) a heavy chain CDR1 of SEQ ID NO:25 (SYSMN), a heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO:27 (GMTTL TFDY), a light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO:29 (AASSLQS), and a light chain CDR3 of SEQ ID NO:30 (QQYNSYPLT); and (f) heavy chain CDR1 of SEQ ID NO:31 (SDSATWN), heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO:33 (GGAPFDY), light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO:35 (KASSLES), and light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

89. The method of claim 69, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39 and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO:41 and a VL chain comprising the amino acid sequence of SEQ ID NO:42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO: 43 and a VL chain comprising the amino acid sequence of SEQ ID NO: 44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45 and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47 and a VL chain comprising the amino acid sequence of SEQ ID NO:

48.

90. The method of claim 69, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) the heavy chain amino acid sequence of SEQ ID NO:49 and the light chain amino acid sequence of SEQ ID NO:50; (b) the heavy chain amino acid sequence of SEQ ID NO:51 and the light chain amino acid sequence of SEQ ID NO:52; (c) the heavy chain amino acid sequence of SEQ ID NO:53 and the light chain amino acid sequence of SEQ ID NO:54; (d) the heavy chain amino acid sequence of SEQ ID NO:55 and the light chain amino acid sequence of SEQ ID NO:56; (e) the heavy chain amino acid sequence of SEQ ID NO:57 and the light chain amino acid sequence of SEQ ID NO:58; and (f) the heavy chain amino acid sequence of SEQ ID NO:59 and the light chain amino acid sequence of SEQ ID NO:

60.

91. The method of claim 69, wherein the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 of SEQ ID NO: 1 (SD SATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVK S), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGL ES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT), Wherein the ADC payload is a topoisomerase I inhibitor and wherein the ADC has a drug to antibody ratio (DAR) of 8.

92. The method of claim 91, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37; and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

93. A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising: immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; For each cancer cell, determining the average optical density of the cell membrane of the cancer cell stained by the dye; determining a median optical density of all cancer cells in the digital image; determining a median absolute deviation of the optical density of the cancer cell in the digital image from the median optical density of all cancer cells; generating a predictive efficacy score for the tissue sample based on the median absolute deviation, wherein the predictive efficacy score is positive if the median absolute deviation is equal to or greater than a deviation threshold, and the predictive efficacy score is negative if the median absolute deviation is less than the deviation threshold, and wherein the deviation threshold is associated with a response of a training patient cohort treated with the ADC; and If the predicted efficacy score is positive, a therapy involving the ADC is recommended to the cancer patient.

94. The method of claim 93, further comprising: When the predicted efficacy score is positive, administering the therapy involving the ADC to the cancer patient.

95. The method of claim 93, wherein the deviation threshold is in the range of 10 to 15 on an optical density scale with a maximum value of 220.

96. The method of claim 93, wherein the detection of cancer cells involves detecting pixels belonging to the cell membrane for each cancer cell using the cell center determined for each cancer cell.

97. The method of claim 93, wherein the staining intensity of each cell membrane is calculated based on the average optical density of brown diaminobenzidine (DAB) signal in pixels of the cell membrane.

98. The method of claim 93, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

99. The method of claim 93, wherein the dye is 3,3'-diaminobenzidine (DAB).

100. The method of claim 93, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

101. The method of claim 93, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

102. The method of claim 101, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

103. The method of claim 93, wherein the ADC payload is a cytotoxin.

104. The method of claim 103, wherein the cytotoxin is a topoisomerase I inhibitor.

105. The method of claim 104, wherein the topoisomerase I inhibitor is represented by the formula: and salts and solvates thereof, and wherein R L Represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

106. The method of claim 93, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is selected from:

107. The method of claim 93, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is (SG3932).

108. The method of claim 93, wherein the ADC has a drug to antibody ratio (DAR) falling within a range selected from the group consisting of 1 to 20, 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10.

109. The method of claim 93, wherein the ADC has a drug to antibody ratio (DAR) selected from the group consisting of: 4 and 8.

110. The method of claim 93, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 5 (KASGLES), and light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT); (b) a heavy chain CDR1 of SEQ ID NO:7 (SYAMS), a heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG), a heavy chain CDR3 of SEQ ID NO:9 (EMQQLAL DY), a light chain CDR1 of SEQ ID NO:10 (RASQGISNFLA), a light chain CDR2 of SEQ ID NO:11 (AASSLQS), and a light chain CDR3 of SEQ ID NO:12 (QQ YNSYPFT); (c) a heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN), a heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS), a heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS), a light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 17 (KASSLES), and a light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT); (d) a heavy chain CDR1 of SEQ ID NO: 19 (SYNMN), a heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG), a heavy chain CDR3 of SEQ ID NO: 21 (GMTT LTFDY), a light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA), a light chain CDR2 of SEQ ID NO: 23 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 24 (QQYISYPLT); (e) a heavy chain CDR1 of SEQ ID NO:25 (SYSMN), a heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO:27 (GMTTL TFDY), a light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO:29 (AASSLQS), and a light chain CDR3 of SEQ ID NO:30 (QQYNSYPLT); and (f) heavy chain CDR1 of SEQ ID NO:31 (SDSATWN), heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO:33 (GGAPFDY), light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO:35 (KASSLES), and light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

111. The method of claim 93, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39 and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO:41 and a VL chain comprising the amino acid sequence of SEQ ID NO:42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO:43 and a VL chain comprising the amino acid sequence of SEQ ID NO:44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45 and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47 and a VL chain comprising the amino acid sequence of SEQ ID NO:

48.

112. The method of claim 93, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) the heavy chain amino acid sequence of SEQ ID NO:49 and the light chain amino acid sequence of SEQ ID NO:50; (b) the heavy chain amino acid sequence of SEQ ID NO:51 and the light chain amino acid sequence of SEQ ID NO:52; (c) the heavy chain amino acid sequence of SEQ ID NO:53 and the light chain amino acid sequence of SEQ ID NO:54; (d) the heavy chain amino acid sequence of SEQ ID NO:55 and the light chain amino acid sequence of SEQ ID NO:56; (e) the heavy chain amino acid sequence of SEQ ID NO:57 and the light chain amino acid sequence of SEQ ID NO:58; and (f) the heavy chain amino acid sequence of SEQ ID NO:59 and the light chain amino acid sequence of SEQ ID NO:

60.

113. The method of claim 93, wherein the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 of SEQ ID NO: 1 (SD SATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVK S), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGL ES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT), Wherein the ADC payload is a topoisomerase I inhibitor and wherein the ADC has a drug to antibody ratio (DAR) of 8.

114. The method of claim 113, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37; and a VL chain comprising the amino acid sequence of SEQ ID NO:

38.

115. A method of generating a predictive efficacy score to predict a cancer patient's response to an antibody drug conjugate (ADC), the antibody drug conjugate comprising an ADC payload and an ADC antibody that targets a protein on a cancer cell, wherein the protein is folate receptor alpha (FRα), the method comprising: immunohistochemically staining the tissue sample using a dye linked to a diagnostic antibody, wherein the diagnostic antibody binds to the protein on the cancer cells in the tissue sample; acquiring a digital image of the tissue sample; detecting cancer cells in the digital image; For each cancer cell, determining the average optical density of the cell membrane of the cancer cell stained by the dye; For each cancer cell, determining the average optical density of the cytoplasm of the cancer cell stained by the dye; for each cancer cell in the digital image, determining a difference between the mean optical density of staining of the cell membrane and the mean optical density of staining of the cytoplasm; identifying, from all cancer cells in the digital image, an 85% quantile of the difference between the mean optical density of the staining of the cell membrane and the mean optical density of the staining of the cytoplasm; as well as If the 85% quantile of the difference exceeds a predetermined difference threshold, a therapy involving the ADC is recommended to the cancer patient.

116. The method of claim 115, further comprising: When the 85% quantile of the difference exceeds the predetermined difference threshold, administering the therapy involving the ADC to the cancer patient.

117. The method of claim 115, wherein the difference threshold is in the range of 21 to 25 on an optical density scale with a maximum value of 220.

118. The method of claim 115, wherein the detection of cancer cells involves detecting pixels belonging to the cell membrane and pixels belonging to the cytoplasm for each cancer cell.

119. The method of claim 115, wherein the staining intensity of each cell membrane is calculated based on the average optical density of brown diaminobenzidine (DAB) signal in pixels of the cell membrane.

120. The method of claim 115, wherein the ADC antibody is a humanized IgG1 monoclonal antibody.

121. The method of claim 115, wherein the dye is 3,3'-diaminobenzidine (DAB).

122. The method of claim 115, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, lung cancer, endometrial cancer, pancreatic cancer, gastric cancer, renal cell carcinoma (RCC), colorectal cancer, head and neck squamous cell carcinoma (HNSCC), breast cancer, cervical cancer, and malignant pleural mesothelioma.

123. The method of claim 115, wherein the cancer patient has a cancer selected from the group consisting of ovarian cancer, non-small cell lung cancer (NSCLC), and breast cancer.

124. The method of claim 123, wherein the NSCLC is selected from the group consisting of squamous NSCLC, adenocarcinoma NSCLC, and a combination of squamous NSCLC and adenocarcinoma NSCLC.

125. The method of claim 115, wherein the ADC payload is a cytotoxin.

126. The method of claim 125, wherein the cytotoxin is a topoisomerase I inhibitor.

127. The method of claim 126, wherein the topoisomerase I inhibitor is represented by the formula: and salts and solvates thereof, and wherein R L Represents a linker connected to the ADC antibody or antigen-binding fragment thereof.

128. The method of claim 115, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is selected from:

129. The method of claim 115, wherein the ADC is an anti-FRα antibody or antigen-binding fragment thereof conjugated to a topoisomerase I inhibitor, wherein the topoisomerase I inhibitor is 130. The method of claim 115, wherein the ADC has a drug to antibody ratio (DAR) falling within a range selected from the group consisting of 1 to 20, 1 to 10, 2 to 10, 2 to 8, 2 to 6, and 4 to 10.

131. The method of claim 115, wherein the ADC has a drug to antibody ratio (DAR) selected from the group consisting of: 4 and 8.

132. The method of claim 115, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) heavy chain CDR1 of SEQ ID NO: 1 (SDSATWN), heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVKS), heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), light chain CDR2 of SEQ ID NO: 5 (KASGLES), and light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT); (b) a heavy chain CDR1 of SEQ ID NO:7 (SYAMS), a heavy chain CDR2 of SEQ ID NO:8 (SISSGRSYIYYADSVKG), a heavy chain CDR3 of SEQ ID NO:9 (EMQQLAL DY), a light chain CDR1 of SEQ ID NO:10 (RASQGISNFLA), a light chain CDR2 of SEQ ID NO:11 (AASSLQS), and a light chain CDR3 of SEQ ID NO:12 (QQ YNSYPFT); (c) a heavy chain CDR1 of SEQ ID NO: 13 (SNSAAWN), a heavy chain CDR2 of SEQ ID NO: 14 (RTYYRSNWYNDYTLSVKS), a heavy chain CDR3 of SEQ ID NO: 15 (GVGRFDS), a light chain CDR1 of SEQ ID NO: 16 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 17 (KASSLES), and a light chain CDR3 of SEQ ID NO: 18 (QEYKTYSIFT); (d) a heavy chain CDR1 of SEQ ID NO: 19 (SYNMN), a heavy chain CDR2 of SEQ ID NO: 20 (SISSGSSYIYYADSMKG), a heavy chain CDR3 of SEQ ID NO: 21 (GMTT LTFDY), a light chain CDR1 of SEQ ID NO: 22 (RASQGISTFLA), a light chain CDR2 of SEQ ID NO: 23 (AASSLQS), and a light chain CDR3 of SEQ ID NO: 24 (QQYISYPLT); (e) a heavy chain CDR1 of SEQ ID NO:25 (SYSMN), a heavy chain CDR2 of SEQ ID NO:26 (SISSRSSYVYYADSVKG), a heavy chain CDR3 of SEQ ID NO:27 (GMTTL TFDY), a light chain CDR1 of SEQ ID NO:28 (RASQGISSFLA), a light chain CDR2 of SEQ ID NO:29 (AASSLQS), and a light chain CDR3 of SEQ ID NO:30 (QQYNSYPLT); and (f) heavy chain CDR1 of SEQ ID NO:31 (SDSATWN), heavy chain CDR2 of SEQ ID NO:32 (RTYYRSKWYSDYAVSVKS), heavy chain CDR3 of SEQ ID NO:33 (GGAPFDY), light chain CDR1 of SEQ ID NO:34 (RASQSISSWLA), light chain CDR2 of SEQ ID NO:35 (KASSLES), and light chain CDR3 of SEQ ID NO:36 (QQYNSYSMYT).

133. The method of claim 115, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) a VH chain comprising the amino acid sequence of SEQ ID NO: 37 and a VL chain comprising the amino acid sequence of SEQ ID NO: 38; (b) a VH chain comprising the amino acid sequence of SEQ ID NO: 39 and a VL chain comprising the amino acid sequence of SEQ ID NO: 40; (c) a VH chain comprising the amino acid sequence of SEQ ID NO:41 and a VL chain comprising the amino acid sequence of SEQ ID NO:42; (d) a VH chain comprising the amino acid sequence of SEQ ID NO:43 and a VL chain comprising the amino acid sequence of SEQ ID NO:44; (e) a VH chain comprising the amino acid sequence of SEQ ID NO: 45 and a VL chain comprising the amino acid sequence of SEQ ID NO: 46; and (f) a VH chain comprising the amino acid sequence of SEQ ID NO:47 and a VL chain comprising the amino acid sequence of SEQ ID NO:

48.

134. The method of claim 115, wherein the ADC antibody or antigen-binding fragment thereof comprises a plurality of amino acid sequences selected from the group consisting of: (a) the heavy chain amino acid sequence of SEQ ID NO:49 and the light chain amino acid sequence of SEQ ID NO:50; (b) the heavy chain amino acid sequence of SEQ ID NO:51 and the light chain amino acid sequence of SEQ ID NO:52; (c) the heavy chain amino acid sequence of SEQ ID NO:53 and the light chain amino acid sequence of SEQ ID NO:54; (d) the heavy chain amino acid sequence of SEQ ID NO:55 and the light chain amino acid sequence of SEQ ID NO:56; (e) the heavy chain amino acid sequence of SEQ ID NO:57 and the light chain amino acid sequence of SEQ ID NO:58; and (f) the heavy chain amino acid sequence of SEQ ID NO:59 and the light chain amino acid sequence of SEQ ID NO:

60.

135. The method of claim 115, wherein the ADC antibody or antigen-binding fragment thereof comprises a heavy chain CDR1 of SEQ ID NO: 1 (SD SATWN), a heavy chain CDR2 of SEQ ID NO: 2 (RTYYRSKWYNDYAVSVK S), a heavy chain CDR3 of SEQ ID NO: 3 (GVGSFDY), a light chain CDR1 of SEQ ID NO: 4 (RASQSISSWLA), a light chain CDR2 of SEQ ID NO: 5 (KASGL ES), and a light chain CDR3 of SEQ ID NO: 6 (QQYNSYSQLT), Wherein the ADC payload is a topoisomerase I inhibitor and wherein the ADC has a drug to antibody ratio (DAR) of 8.

136. The method of claim 135, wherein the ADC antibody or antigen-binding fragment thereof comprises a VH chain comprising the amino acid sequence of SEQ ID NO: 37; and a VL chain comprising the amino acid sequence of SEQ ID NO: 38.

Images