Assessment of antigen recovery and target recovery progress by vibrational spectroscopy

Abstract

The present disclosure relates to automated systems and methods for quantitatively determining the unmasking state of a biological specimen subjected to an unmasking process (e.g., an antigen retrieval process and / or a target retrieval process) using a trained unmasking state estimation engine. In some embodiments, the trained unmasking state estimation engine comprises a machine learning algorithm based on projection onto a latent structure regression model. In some embodiments, the trained unmasking state estimation engine comprises a neural network.

Description

[Technical field]

[0001] [CROSS REFERENCE TO RELATED APPLICATIONS] This application claims the benefit of the filing date of U.S. Patent Application No. 62 / 892,688, filed August 28, 2019, the disclosure of which is incorporated herein by reference in its entirety. One embodiment of the present invention relates to assessment of antigen recovery and target recovery progress, for example, by vibrational spectroscopy. [Background technology]

[0002] Immunohistochemical (IHC) slide staining is widely used to study different types of cells, such as cancerous cells and immune cells, in living tissues, as it can be utilized to identify proteins within cells in tissue sections. Therefore, IHC staining can be used in studies to understand the distribution and localization of differentially expressed biomarkers of immune cells (such as T cells and B cells) in cancerous tissues for immune response studies. For example, tumors often contain immune cell infiltrates, which may hinder tumor development or promote tumor growth.

[0003] In situ hybridization (ISH) can be used to look for the presence of genetic abnormalities or conditions, such as cancer amplifications that give rise to genes specifically in cells that appear morphologically malignant when viewed under a microscope. In situ hybridization (ISH) detects or localizes target nucleic acid target genes in a cell or tissue sample using a labeled DNA or RNA probe molecule that is antisense to the target gene sequence or transcript. ISH is performed by exposing a cell or tissue sample immobilized on a glass slide to a labeled nucleic acid probe that can specifically hybridize to a given target gene in the cell or tissue sample. By exposing the cell or tissue sample to multiple nucleic acid probes labeled with multiple different nucleic acid tags, several target genes can be analyzed simultaneously. By utilizing labels with different emission wavelengths, simultaneous multicolor analysis can be performed in a single step on a single target cell or tissue sample.

[0004] Thin tissue sections are used in histology to obtain representative information about tissue samples. The quality of the thin section should meet several characteristics in order to adequately represent the entire tissue area from which the resection of the sample was made. Although guidelines can vary depending on the type of tissue and use, the size of the thin section should generally not be less than 2 μm. Typically, tissue sections are prepared in the 2 to 5 μm range and should not vary in thickness by more than 50% over the lateral extent of the thin section to allow for proper further processing.

[0005] Fixation of tissue and cell samples is used to ensure that the morphology of the sample and the spatial distribution of biomolecules are preserved, thus helping to enable diagnosis by pathologists. Fixation of samples is routinely achieved using neutral buffered formalin (NBF). It is believed that the formaldehyde in NBF preserves the morphology of tissues and cells by forming crosslinks between reactive groups on proteins and nucleic acids in the sample, and these crosslinks can result in certain portions of the molecules being rendered undetectable. For example, formaldehyde primarily preserves or fixes tissues or cells by crosslinking primary amine groups in proteins with other nearby nitrogen atoms in the protein or DNA via -CH2- bonds. However, the process of tissue fixation often masks antigens on certain proteins whose detection is desirable for diagnostic and prognostic purposes. Thus, immunohistochemistry and in situ hybridization procedures often use an unmasking process (also known as "antigen retrieval" or "target retrieval" for IHC and ISH, respectively) that makes the protein antigen or nucleic acid target accessible to detection reagents such as antibodies or probes.

[0006] Various methods are used to reverse the effects of formalin fixation and provide access to antigens and targets in fixed biological samples. One traditional method involves treating the tissue with proteolytic enzymes, which open the tissue and allow antibodies to access the antigens. For some specific antibodies, this method is the preferred antigen retrieval procedure, but it suffers from several limitations. Because proteolytic enzymes degrade tissue proteins, there is a high risk of destroying specific epitopes in the proteins that the antibodies recognize, and antigenicity may be lost rather than retrieved. Furthermore, prolonged proteolytic treatments may cause loss of the three-dimensional structure of proteins, which is also important for antibody-antigen recognition. These factors obviously require strict control by the incubation time. However, the required incubation time may depend on the activity of the enzyme.

[0007] Another factor that affects the efficiency of proteolytic target recovery is the length of fixation of the tissue. The degree of cross-linking in the tissue is believed to increase with fixation time, and thus tissue fixed for 72 hours will require a longer proteolytic treatment than tissue fixed for 24 hours. However, in many cases, information regarding fixation time is not available, in which case corresponding adjustments cannot be made in enzyme incubation times.

[0008] A fundamentally different method involves heat-induced antigen retrieval. High temperature promotes hydrolysis of aldehyde-induced modifications, and the released aldehyde is diluted in water, thus preventing it from reacting again with proteins. Another method for retrieving the immunoreactivity of formaldehyde-crosslinked antigens involves heat treatment using heat or high-energy treatment of the sample. Such a method is described, for example, in U.S. Pat. No. 5,244,787, in which formaldehyde-fixed tissue preparations are immersed in water and subjected to microwave energy. Yet another method for retrieving antigens from formaldehyde-fixed tissues is the use of a pressure cooker in combination with microwaves or in the form of an autoclave, as described in Norton, 1994. J. Pathos. 173(4):371-9 and Taylor et al., 1996. Biotech Histochem 71(5):263-70.

[0009] In addition to pure distilled water, many reagents are routinely used for antigen retrieval, including Tris, urea, EDTA, citrate and saline buffers. Furthermore, many solutions contain detergents, and the selection includes both ionic and non-ionic detergents. Despite extensive research in the field of staining enhancement, no universal method or reagent has been identified that works for all subsequent immunohistochemical or immunocytochemical procedures. Summary of the Invention

[0010] Antigen retrieval has become central to modern immunohistochemistry, as it unmasks protein epitopes so that they can be detected by IHC-based detection. Similarly, target retrieval is important in unmasking nucleic acids for ISH-based detection. Although the exact mechanisms of protein epitope and nucleic acid unmasking remain speculative, it is a key chemical process that allows antigens and nucleic acids to be detected in formalin-fixed paraffin-embedded tissues. Unmasking is therefore of fundamental importance in detecting biomarkers, such as cancer biomarkers. Furthermore, the degree to which tissue is unmasked can significantly affect the protein and / or nucleic acid expression levels detected. To date, there are no analytical methods to measure the quality of the unmasking process or the extent to which the unmasking process is performed.

[0011] To meet this important unmet need, applicants have developed systems and methods that identify features in acquired spectral data derived from a biological specimen and predict the unmasking state of the biological specimen based on the identified features. In some embodiments, the disclosed systems and methods identify "unmasking signatures" in the acquired spectral data and calculate a predicted unmasking state based on the identified unmasking signatures. Applicants have demonstrated that changes in the molecular composition of the biological specimen (e.g., changes in tissue biostructure and / or availability of immune recognition sites based on conformational changes of their recognition sites) can appear in the acquired vibrational spectral data of the biological specimen, and the shifts in the vibrational spectral data can be utilized to predict the unmasking state of the biological specimen (e.g., unmasking time, unmasking temperature, etc.).

[0012] In view of the above, applicant provides systems and methods that advantageously enable quantitative and / or qualitative prediction of the unmasking status of a biological specimen. For example, the systems and methods facilitate quantitative determination of the degree of unmasking of a biological specimen, where the degree of unmasking can correspond to any unmasking process conditions (e.g., temperature, duration, reagents used, pressure, presence of energy introduced other than heat, or any combination thereof). Applicant herein not only demonstrates that the disclosed systems and methods can accurately predict the unmasking status of a biological specimen, but also demonstrates that the systems and methods enable accurate predictions to be made for various tissue types (e.g., breast tissue, tonsil tissue). Thus, the disclosed systems and methods are applicable to any type of tissue or cell specimen. Moreover, applicant further demonstrates that the disclosed systems and methods are applicable to both antigen retrieval processes and target retrieval processes. Overall, applicant submits that the disclosed systems enable rapid and accurate prediction of the unmasking status of a biological specimen by using certain algorithms, such as machine learning algorithms or other automated algorithms, ultimately providing improved IHC and / or ISH assay results and patient care.

[0013] In view of the above, a first aspect of the present disclosure is a system for predicting an unmasking state of a test biological specimen processed in an unmasking process, the system comprising: (i) one or more processors; and (ii) one or more memories coupled to the one or more processors, the one or more memories storing computer executable instructions that, when executed by the one or more processors, cause the system to perform operations including: (a) acquiring test vibrational spectral data from the test biological specimen, the test spectral data including vibrational spectral data derived from at least a portion of the biological specimen; (b) deriving unmasking features from the acquired test spectral data using a trained unmasking state estimation engine; and (c) predicting an unmasking state of the test biological specimen based on the derived unmasking features. In some embodiments, the unmasking process is an antigen retrieval process. In some embodiments, the unmasking process is a target retrieval process.

[0014] In some embodiments, the unmasking state includes one of a predicted unmasking duration or a predicted unmasking temperature. In some embodiments, the unmasking state includes both a predicted unmasking duration and a predicted unmasking temperature. In some embodiments, the unmasking state further includes an estimate of tissue quality. In some embodiments, the unmasking state estimation engine is trained using one or more training spectrum data sets, each training spectrum data set including a plurality of training vibration spectra derived from a plurality of separately unmasked training tissue samples, each training vibration spectrum including one or more class labels. In some embodiments, the one or more class labels are selected from a known unmasking duration, a known unmasking temperature, and a qualitative assessment of the unmasking state.

[0015] In some embodiments, each training spectral data set is derived by: (i) acquiring a training biological specimen; (ii) dividing the acquired training biological specimen into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking. In some embodiments, the different unmasking conditions include varying both the duration of unmasking and the unmasking temperature. In some embodiments, the different unmasking conditions include varying the unmasking agent used for unmasking. In some embodiments, the different unmasking conditions include varying the pressure used for unmasking.

[0016] In some embodiments, the acquired test spectral data includes an average vibration spectrum derived from a plurality of normalized and corrected vibration spectra. In some embodiments, the plurality of normalized and corrected vibration spectra are acquired by: (i) identifying a plurality of spatial regions within the test biological specimen, (ii) acquiring a vibration spectrum from each individual region of the plurality of identified regions, (iii) correcting the vibration spectrum acquired from each individual region to provide a corrected vibration spectrum for each individual region, and (iv) amplitude normalizing the corrected vibration spectrum from each individual region to a predetermined global maximum to provide an amplitude-normalized vibration spectrum for each region. In some embodiments, the vibration spectrum acquired from each individual region is corrected by: (i) compensating each acquired vibration spectrum for atmospheric effects to provide an atmospheric-corrected vibration spectrum, and (ii) compensating said atmospheric-corrected vibration spectrum for scattering.

[0017] In some embodiments, the trained unmasking state estimation engine comprises a machine learning algorithm based on dimensionality reduction. In some embodiments, the dimensionality reduction comprises projection onto a latent structure regression model. In some embodiments, the dimensionality reduction comprises principal component analysis and discriminant analysis. In some embodiments, the trained unmasking state estimation engine comprises a neural network.

[0018] In some embodiments, the system further comprises an operation for assessing whether the test biological specimen is suitable for labeling with one or more specific binding entities. In some embodiments, the acquired test spectral data comprises vibrational spectral information of at least the amide I band. In some embodiments, the acquired test spectral data comprises vibrational spectral information of at least the amide I band from about 3200 to about 3400 cm -1 In some embodiments, the acquired test spectrum data includes vibrational spectral information at wavelengths ranging from about 2800 to about 2900 cm -1 In some embodiments, the acquired test spectral data includes vibrational spectral information at wavelengths ranging from about 1020 to about 1100 cm -1 In some embodiments, the acquired test spectrum data includes vibrational spectral information at wavelengths ranging from about 1520 to about 1580 cm. -1 It contains vibrational spectral information for wavelengths in the range of

[0019] A second aspect of the present disclosure is a non-transitory computer-readable medium storing instructions for predicting an unmasking state of a test biological specimen processed in an unmasking process, the instructions including: (a) acquiring test spectral data from the test biological specimen, the test spectral data including vibrational spectral data derived from at least a portion of the biological specimen; (b) deriving unmasking features from the acquired test spectral data using a trained unmasking state estimation engine, the unmasking state estimation engine being trained using a training spectral data set acquired from a plurality of separately unmasked training biological specimens, the training spectral data set including class labels for at least unmasking duration and unmasking temperature; and (c) predicting an unmasking state of the test biological specimen processed in the unmasking process based on the derived unmasking features.

[0020] In some embodiments, the unmasking state includes at least one of a predicted unmasking duration, a predicted unmasking temperature, and an estimate of tissue quality, hi some embodiments, the unmasking state includes both a predicted unmasking duration and a predicted unmasking temperature.

[0021] In some embodiments, each training spectral data set is derived by: (i) acquiring a training biological specimen; (ii) dividing the acquired training biological specimen into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking. In some embodiments, the different unmasking conditions include varying both the duration of unmasking and the unmasking temperature. In some embodiments, the different unmasking conditions include varying the unmasking agent used for unmasking. In some embodiments, the different unmasking conditions include varying the pressure used for unmasking.

[0022] In some embodiments, the training biological specimen comprises the same tissue type as the test biological specimen. In some embodiments, the training biological specimen comprises a different tissue type than the test biological specimen. In some embodiments, the unmasking process is an antigen retrieval process. In some embodiments, the unmasking process is a target retrieval process. In some embodiments, the test biological specimen is unstained. In some embodiments, the test biological specimen is stained for the presence of one or more biomarkers.

[0023] In some embodiments, the acquired test spectral data includes vibrational spectral information of at least the amide I band. In some embodiments, the acquired test spectral data includes vibrational spectral information of at least the amide I band from about 3200 to about 3400 cm -1 In some embodiments, the acquired test spectrum data includes vibrational spectral information at wavelengths ranging from about 2800 to about 2900 cm -1In some embodiments, the acquired test spectral data includes vibrational spectral information at wavelengths ranging from about 1020 to about 1100 cm -1 In some embodiments, the acquired test spectrum data includes vibrational spectral information at wavelengths ranging from about 1520 to about 1580 cm. -1 It contains vibrational spectral information for wavelengths in the range of

[0024] A third aspect of the present disclosure is a method for predicting an unmasking state of a test biological specimen processed in an unmasking process, comprising: (a) acquiring test spectral data from the test biological specimen, the acquired test spectral data including vibrational spectral data derived from at least a portion of the biological specimen; (b) deriving unmasking features from the acquired test spectral data using a trained unmasking state estimation engine, the unmasking state estimation engine being trained using a training spectral data set acquired from a plurality of separately unmasked training biological specimens, the training spectral data set including at least unmasking duration and unmasking temperature class labels; and (c) predicting at least one of the duration or temperature of the unmasking process to which the unmasked test biological specimen was subjected based on the derived unmasking features. In some embodiments, the unmasking process is an antigen retrieval process. In some embodiments, the unmasking process is a target retrieval process.

[0025] In some embodiments, the unmasking state includes both a predicted unmasking duration and a predicted unmasking temperature. In some embodiments, the unmasking state further includes at least one of an estimate of tissue quality and an estimate of damage sustained during the unmasking process. In some embodiments, the trained unmasking state estimation engine includes a machine learning algorithm based on projection onto a latent structure regression model. In some embodiments, the trained unmasking state estimation engine includes a machine learning algorithm based on principal component analysis and discriminant analysis. In some embodiments, the trained unmasking state estimation engine includes a neural network.

[0026] In some embodiments, the acquired test spectral data includes vibrational spectral information of at least the amide I band. In some embodiments, the acquired test spectral data includes vibrational spectral information of at least the amide I band from about 3200 to about 3400 cm -1 In some embodiments, the acquired test spectrum data includes vibrational spectral information at wavelengths ranging from about 2800 to about 2900 cm -1 In some embodiments, the acquired test spectral data includes vibrational spectral information at wavelengths ranging from about 1020 to about 1100 cm -1 In some embodiments, the acquired test spectrum data includes vibrational spectral information at wavelengths ranging from about 1520 to about 1580 cm. -1 It contains vibrational spectral information for wavelengths in the range of

[0027] For a general understanding of the features of the present disclosure, reference is made to the drawings, in which like reference numerals are used throughout to identify identical elements. [Brief description of the drawings]

[0028] [Figure 1] 1 illustrates an exemplary digital pathology system including an image acquisition device and a computer system, according to one embodiment of the present disclosure. [Diagram 2]In accordance with one embodiment of the present disclosure, various modules that can be utilized within a system or digital pathology workflow to quantitatively or qualitatively predict the unmasking status of a test biological sample are described. [Diagram 3] 10 illustrates a flowchart showing various steps of quantitatively or qualitatively predicting the unmasking state of a test biological specimen using a trained unmasking state estimation engine, according to one embodiment of the present disclosure. [Figure 4A] 1 illustrates a process of obtaining multiple training tissue samples for different unmasking from two different training biological specimens according to one embodiment of the present disclosure. [Figure 4B] 1 illustrates differential unmasking of multiple training tissue samples obtained from two different training biological specimens, according to one embodiment of the present disclosure. [Figure 4C] 1 presents a flow chart illustrating various steps for acquiring vibrational spectra of training biological samples, according to one embodiment of the present disclosure. [Diagram 5] 1 illustrates a flow chart showing various steps for obtaining an average vibrational spectrum of a test biological sample, according to one embodiment of the present disclosure. [Figure 6] 10 illustrates a flow chart showing various steps for correcting, normalizing, and averaging acquired spectra from biological samples, including test and training biological samples, according to one embodiment of the present disclosure. [Figure 7A] 1 depicts a plot of the first principal component (PC1) versus the second principal component (PC2) of mid-IR spectra from differentially recovered tissue samples. The solid line represents the output from a discriminant analysis that calculates the boundary line for separating two different recovery conditions. [Figure 7B] Confusion matrices for PCA+DA with increasing numbers of principal components are presented. The best classification accuracy appears to be achieved with 5 principal components. [Figure 8]We provide an example of tonsillar tissue labeled with antiserum raised against Ki-67. Only tonsillar tissue was imaged (circled area in the left image). Connective tissue, which occasionally showed high background but was not present in other sections, was excluded. [Figure 9] 1 provides a visible image of an exemplary tissue section having multiple regions identified, and further provides examples of averaged, processed, and normalized vibrational spectra collected from the indicated regions in the visible image. [Figure 10A] 1 provides a mid-IR absorption spectrum, specifically showing protein bands within the acquired mid-IR spectrum. [Figure 10B] The peak position of the first derivative of the amide I band versus the FWHM of the band is given, making it clear that the unrecovered spectrum has a significantly different spectrum from the other recovered tissues. [Figure 11] An example is described for training an unmasking state estimation engine, specifically a PLSR machine learning algorithm. First, a model is trained on input vibration spectra with known classifications, and a model is developed that assigns a weight to each wavelength that roughly corresponds to how well each wavelength correlates (or anti-correlates) with the response (e.g., unmasking time). Finally, the model is applied to the trained vibration spectrum data to evaluate how accurately it predicts unmasking time. [Figure 12] Experimental unmasking times versus the predicted unmasking times of the unmasking state estimation engine are shown. The unmasking state estimation engine was trained on "X" data points (accuracy = 2 min) and performance was evaluated on blind spectra as shown by the purple circle (accuracy = 8 min). [Figure 13] We provide pictures of four tissues imaged by mid-IR over a time-temperature course. The unmasking state estimation engine was trained on the tissues in the bottom row and on the right (circled areas containing three tissue sections), and the predictive power of the unmasking state estimation engine was evaluated on the tissue in the top left section of the images (circled area containing only a single tissue section). [Figure 14] For the blinded tonsils only (shown in FIG. 13), we show the experimental unmasking temperatures versus the predicted unmasking temperatures (as predicted using the trained unmasking state estimation engine). The average prediction accuracy was about 1.3°C. [Figure 15] Shown are serial sections derived from tonsillar samples collected either at 98.6° C. for 60 min. The overlaid dots indicate the sampling pattern used to collect mid-IR spectra for analysis. [Figure 16] Figure 1 shows serial sections derived from tonsillar samples collected either at 140 °C for 5 min. The overlaid dots indicate the sampling pattern used to collect mid-IR spectra for analysis. [Figure 17A] It provides a graph showing how many components are needed in a PLSR model to capture the variance in the X and Y variables. [Figure 17B] We provide a graph showing how the mean squared prediction error (MSPE) of the model increases with the number of components. [Figure 17C] Graphs are provided showing the performance of the model on the training set and on the holdout validation data. The prediction error is identical for both types of data indicating a well-trained model that identified the true signature of recovery in the mid-IR spectrum. [Figure 18] Differential bands from tonsil samples collected at low and high temperatures are shown. Each graph is labeled with the possible molecular origin of mid-IR absorption in the respective mid-IR region. [Figure 19] It is shown that the trained unmasked state estimation engine was able to distinguish between low and high temperature recoveries, which also stain differently for C4d. [Figure 20]Graphs of mid-IR spectra (top row) and mid-IR first derivatives (bottom row) centered on the amide I region (left row) and amide A region (right row) where visible differences are observed between different unmasking conditions are provided. [Figure 21A] The amide I peak position of the first derivative of the amide I band versus the FWHM of the band is shown for all replicates of unmasking time (FIG. 21A) and the average variation for each unmasking temperature (FIG. 21B). [Figure 21B] The amide I peak position of the first derivative of the amide I band versus the FWHM of the band is shown for all replicates of unmasking time (FIG. 21A) and the average variation for each unmasking temperature (FIG. 21B). [Figure 22] Experimental unmasking temperatures versus model predicted unmasking temperatures are provided. The unmasking state estimation engine was trained on the "training" spectra and performance was evaluated on the "hold-out" spectra (accuracy = 2.02°C). Statistics shown are only between the hold-out data as determined by a 2-tailed rank sum test. [Figure 23A] 1 provides mid-IR absorption spectra of breast tissue after retrieval based on CC1 and CC2. [Figure 23B] FIG. 23A provides the difference in absorption spectra between CC1 and CC2 based breast tissue averages. [Figure 24A] The mid-IR first derivative (A) and the difference of the mid-IR first derivative (B) of the recovery based on CC1 and CC2 are provided. Zoom-in of bands centered near: 1200 cm-1 (C), 2880 cm-1 (D) and 1600 cm-1 (E). [Figure 24B] The mid-IR first derivative (A) and the difference of the mid-IR first derivative (B) of the recovery based on CC1 and CC2 are provided. Zoom-in of bands centered near: 1200 cm-1 (C), 2880 cm-1 (D) and 1600 cm-1 (E). [Figure 24C]The mid-IR first derivative (A) and the difference of the mid-IR first derivative (B) of the recovery based on CC1 and CC2 are provided. Zoom-in of bands centered near: 1200 cm-1 (C), 2880 cm-1 (D) and 1600 cm-1 (E). [Figure 24D] The mid-IR first derivative (A) and the difference of the mid-IR first derivative (B) of the recovery based on CC1 and CC2 are provided. Zoom-in of bands centered near: 1200 cm-1 (C), 2880 cm-1 (D) and 1600 cm-1 (E). [Figure 24E] The mid-IR first derivative (A) and the difference of the mid-IR first derivative (B) of the recovery based on CC1 and CC2 are provided. Zoom-in of bands centered near: 1200 cm-1 (C), 2880 cm-1 (D) and 1600 cm-1 (E). [Figure 25A] Zoom-in views of several bands that are well differentiated between CC1- and CC2-based recovery are shown (top row). The figure also shows the spectral distribution of CC1 and CC2 recovery at wavenumbers of (A) 1265 cm-1, (B) 1472 cm-1, (C) 1664 cm-1, and (D) 2862 cm-1 (bottom row). [Figure 25B] Zoom-in views of several bands that are well differentiated between CC1- and CC2-based recovery are shown (top row). The figure also shows the spectral distribution of CC1 and CC2 recovery at wavenumbers of (A) 1265 cm-1, (B) 1472 cm-1, (C) 1664 cm-1, and (D) 2862 cm-1 (bottom row). [Figure 25C] Zoom-in views of several bands that are well differentiated between CC1- and CC2-based recovery are shown (top row). The figure also shows the spectral distribution of CC1 and CC2 recovery at wavenumbers of (A) 1265 cm-1, (B) 1472 cm-1, (C) 1664 cm-1, and (D) 2862 cm-1 (bottom row). [Figure 25D]Zoom-in views of several bands that are well differentiated between CC1- and CC2-based recovery are shown (top row). The figure also shows the spectral distribution of CC1 and CC2 recovery at wavenumbers of (A) 1265 cm-1, (B) 1472 cm-1, (C) 1664 cm-1, and (D) 2862 cm-1 (bottom row). [Figure 26] 1 shows typical FT-IR and Raman spectra of collagen. [Figure 27A] A table showing infrared and Raman characteristic frequencies of biological samples is provided. [Figure 27B] A table showing infrared and Raman characteristic frequencies of biological samples is provided. (Continued from FIG. 27A). DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0029] It is also to be understood that, unless expressly stated to the contrary, in any method claimed herein that includes multiple steps or acts, the order of the method steps or acts is not necessarily limited to the order in which the method steps or acts are described.

[0030] References herein to "one embodiment," "an embodiment," "an illustrative embodiment," and the like indicate that the described embodiment may include a particular feature, structure, or characteristic, but that all embodiments may or may not necessarily include that particular feature, structure, or characteristic. Moreover, such phrases do not necessarily refer to the same embodiment. Moreover, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is believed to be within the knowledge of one of ordinary skill in the art to implement such feature, structure, or characteristic in connection with other embodiments, whether or not explicitly described.

[0031] As used herein, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Similarly, the word "or" is intended to include "and" unless the context clearly dictates otherwise. The term "comprising" is defined inclusively, such that "including A or B" means including A, B, or A and B.

[0032] As used herein and in the claims, "or" should be understood to have the same meaning as "and / or" as defined above. For example, when separating items in a list, "or" or "and / or" shall be interpreted as being inclusive. For example, it shall be interpreted as including at least one of a number or list of elements, but including a plurality, and optionally additional unlisted items. Only terms clearly indicated to the contrary, such as "only one of" or "exactly one of," or, when used in the claims, "consisting of," shall refer to the inclusion of exactly one element of a number or list of elements. In general, the term "or" as used herein shall only be interpreted as indicating exclusive alternatives (e.g., "one or the other but not both") when preceded by a term of exclusivity, such as "either," "one of," "only one of," or "exactly one of." "Consisting essentially of," when used in the claims, shall have its ordinary meaning as used in the field of patent law.

[0033] Terms such as "comprising," "including," "having," and the like are used interchangeably and have the same meaning. Similarly, "comprises," "includes," "has," and the like are used interchangeably and have the same meaning. Specifically, each term is defined consistent with the general U.S. patent law definition of "comprising," and therefore is to be interpreted as an open term meaning "at least the following," and not to exclude additional features, limitations, aspects, and the like. Thus, for example, "an apparatus having components a, b, and c" means that the apparatus includes at least components a, b, and c. Similarly, the phrase: "a method including steps a, b, and c" means that the method includes at least steps a, b, and c. Additionally, although steps and processes may be outlined herein in a particular order, one of ordinary skill in the art will recognize that the ordering of steps and processes may vary.

[0034] As used in the specification and claims herein, the phrase "at least one" in connection with a list of one or more elements should be understood to mean at least one element selected from any one or more of the elements in the list of elements, but not necessarily including at least one of each and every element specifically listed in the list of elements, and not excluding any combination of elements in the list of elements. This definition also allows for elements other than those specifically identified in the list of elements to which the phrase "at least one" refers can optionally be present, whether or not related to those elements specifically identified. Thus, as a non-limiting example, "at least one of A and B" (or, equivalently, "at least one of A or B," or, equivalently, "at least one of A and / or B") can refer in one embodiment to at least one, optionally including two or more A, and no B (and optionally including elements other than B); in another embodiment to at least one, optionally including two or more B, and no A (and optionally including elements other than A); in yet another embodiment to at least one, optionally including two or more A, and at least one, optionally including two or more B (and optionally including other elements); and so forth.

[0035] As used herein, the term "antigen" refers to a substance to which an antibody, antibody analog (e.g., aptamer), or antibody fragment binds. Antigens may be endogenous, whereby they are produced within a cell as a result of normal or abnormal cellular metabolism, or due to viral or intracellular bacterial infection. Endogenous antigens include xenogeneic (heterologous), self, and idiotypic or allogeneic (homologous) antigens. Antigens can also be tumor-specific antigens or can be presented by tumor cells. In this case, they are called tumor-specific antigens (TSAs), and generally result from tumor-specific mutations. Antigens can also be tumor-associated antigens (TAAs), which are presented by tumor cells and normal cells. Antigens also include CD antigens, which refer to any of several cell surface markers expressed by white blood cells and can be used to distinguish cell lineages or developmental stages. Such markers can be identified by specific monoclonal antibodies and are numbered by their differentiation clusters.

[0036] As used herein, the terms "biological specimen", "sample", or "tissue sample" refer to any sample containing biomolecules (such as proteins, peptides, nucleic acids, lipids, carbohydrates, or combinations thereof) obtained from any organism, including viruses. Other examples of organisms include mammals (such as humans, veterinary animals such as cats, dogs, horses, cows, and pigs, and laboratory animals such as mice, rats, and primates), insects, annelids, arachnids, marsupials, reptiles, amphibians, bacteria, and fungi. Biological specimens include tissue samples (such as tissue sections or needle biopsies of tissues), cell samples (such as cytological smears such as Pap smears or blood smears, or samples of cells obtained by microdissection), or cell fractions, fragments, or organelles (obtained by lysing cells and separating their components, such as by centrifugation). Other examples of biological specimens include blood, serum, urine, semen, feces, cerebrospinal fluid, interstitial fluid, mucous membranes, tears, sweat, pus, biopsy tissue (e.g., obtained by surgical or needle biopsy), nipple aspirate, earwax, milk, vaginal fluid, saliva, swabs (such as cheek swabs), or any material containing biomolecules derived from an initial biological specimen. In certain embodiments, the term "biological specimen" as used herein refers to a sample prepared from a tumor or portion thereof obtained from a subject (such as a homogenized or liquefied sample).

[0037] As used herein, the term "biomarker" or "marker" refers to a measurable indicator of some biological state or condition. In particular, a biomarker can be a protein or peptide, e.g., a surface protein, that can be specifically stained and that indicates a biological feature of a cell, e.g., a cell type or a physiological state of a cell. An immune cell marker is a biomarker that selectively indicates a characteristic related to a mammalian immune response. Biomarkers can be used to determine how well a body responds to treatment of a disease or condition, or whether a subject is predisposed to a disease or condition. In the context of cancer, a biomarker refers to a biological substance that indicates the presence of cancer in the body. A biomarker can be a molecule secreted by a tumor, or a specific response of the body to the presence of cancer. Genetic biomarkers, epigenetic biomarkers, proteomic biomarkers, glycomic biomarkers, and imaging biomarkers can be used for cancer diagnosis, prognosis, and epidemiology. Such biomarkers can be assayed in non-invasively collected biological fluids, such as blood or serum. Several gene and protein-based biomarkers are already used in patient care, including but not limited to AFP (liver cancer), BCR-ABL (chronic myeloid leukemia), BRCA1 / BRCA2 (breast / ovarian cancer), BRAF V600E (melanoma / colorectal cancer), CA-125 (ovarian cancer), CA19.9 (pancreatic cancer), CEA (colorectal cancer), EGFR (non-small cell lung cancer), HER-2 (breast cancer), KIT (gastrointestinal stromal tumor), PSA (prostate specific antigen), S100 (melanoma), and others. Biomarkers can be useful as diagnostics (to identify early cancers) and / or prognostic agents (to predict how aggressive a cancer is and / or how a subject will respond to a particular treatment and / or how likely the cancer is to recur). Other examples of biomarkers are disclosed herein.

[0038] As used herein, the term "cytological sample" refers to a cell sample in which the cells of the sample are partially or completely disaggregated, such that the sample no longer reflects the spatial relationship of the cells when they were present in the subject from which the cell sample was obtained. Examples of cytological samples include tissue scrapings (such as cervical scrapings), fine needle aspirates, samples obtained by lavage of a subject, and the like.

[0039] As used herein, the term "immunohistochemistry" refers to a method of determining the presence or distribution of an antigen in a sample by detecting the interaction of the antigen with a specific binding agent such as an antibody. The sample is contacted with the antibody under conditions that allow antibody-antigen binding. Antibody-antigen binding can be detected by a detectable label conjugated to the antibody (direct detection) or by a detectable label conjugated to a secondary antibody that specifically binds to the primary antibody (indirect detection). In some instances, indirect detection can include a third or more antibodies that serve to further increase the detectability of the antigen. Examples of detectable labels include enzymes, fluorophores and haptens, which, in the case of enzymes, can be used with chromogenic or fluorogenic substrates.

[0040] As used herein, the term "slide" refers to any substrate of any suitable dimensions (e.g., substrates made entirely or partially of glass, quartz, plastic, silicon, etc.) on which a biological specimen is placed for analysis, more specifically, a "microscope slide" such as a standard 3" x 1" microscope slide or a standard 75mm x 25mm microscope slide. Examples of biological specimens that can be placed on a slide include, but are not limited to, cytological smears, thin tissue sections (such as from a biopsy), and arrays of biological specimens, e.g., tissue arrays, cell arrays, DNA arrays, RNA arrays, protein arrays, or any combination thereof. Thus, in one embodiment, tissue sections, DNA samples, RNA samples, and / or proteins are placed at specific locations on the slide. In some embodiments, the term slide can refer to SELDI and MALDI chips, as well as silicon wafers.

[0041] As used herein, the term "specific binding entity" refers to a member of a specific binding pair. A specific binding pair is a pair of molecules that are characterized by binding to one another to the substantial exclusion of binding to other molecules (e.g., a specific binding pair has a binding constant that is at least 10 times higher than the binding constant of either of the two members of the binding pair with other molecules in a biological sample). 3 M -1 Large, 10 4 M -1 Greater than or equal to 10 5 M -1 (Can have large binding constants). Particular examples of specific binding moieties include specific binding proteins (e.g., antibodies, lectins, avidins such as streptavidin, and protein A). Specific binding moieties can also include molecules (or portions thereof) that are specifically bound by such specific binding proteins.

[0042] As used herein, the term "spectral data" encompasses raw image spectral data obtained from a biological specimen or any portion thereof, such as with a spectrometer.

[0043] As used herein, the term "spectrum" refers to information (absorption, transmission, reflection) obtained "at" or within a particular wavelength or wavenumber range of electromagnetic radiation. -1 It can be as small as 0.01cm -1 It should be noted that measurements at a so-called "single laser wavelength" typically cover a small spectral range (e.g., the laser linewidth) and are therefore included whenever the term "spectrum" is used throughout this manuscript. For example, a transmission measurement at a fixed wavelength setting of a quantum cascade laser is herein intended to be included in the term spectrum throughout this application.

[0044] As used herein, the terms "staining," "staining," and the like generally refer to any treatment of a biological specimen that detects and / or distinguishes the presence, location, and / or amount (e.g., concentration) of a particular molecule (e.g., lipid, protein, or nucleic acid) or a particular structure (e.g., normal or malignant cells, cytosol, nucleus, Golgi apparatus, or cytoskeleton) in the biological specimen. For example, staining can provide contrast between a particular molecule or particular cellular structure and the surrounding parts of the biological specimen, and the intensity of the staining can provide a measure of the amount of a particular molecule in the specimen. Staining can be used to aid in the observation of molecules, cellular structures, and organisms using bright field microscopes as well as other observation tools such as phase contrast microscopes, electron microscopes, and fluorescence microscopes. Some staining performed by the system can be used to visualize the outline of a cell. Other staining performed by the system can depend on the particular cellular component (e.g., molecule or structure) being stained, with no or relatively little staining of other cellular components. Examples of types of staining methods performed by the system include, but are not limited to, histochemical methods, immunohistochemical methods, and other methods based on reactions between molecules (including non-covalent interactions), such as hybridization reactions between nucleic acid molecules. Specific staining methods include, but are not limited to, primary staining methods (e.g., H&E staining, Pap staining, etc.), enzyme-linked immunohistochemical methods, and in situ RNA and DNA hybridization methods, such as fluorescent in situ hybridization (FISH).

[0045] As used herein, the term "target" refers to any molecule whose presence, location, and / or concentration is or can be determined. Examples of target molecules include proteins, epitopes, nucleic acid sequences, and haptens, such as haptens covalently bound to proteins. Target molecules are usually detected using one or more conjugates of a specific binding molecule and a detectable label.

[0046] As used herein, the term "tissue sample" refers to a cell sample that preserves the cross-sectional spatial relationships between cells that were present within the subject from which the sample was obtained. "Tissue sample" is intended to encompass both primary tissue samples (e.g., cells and tissues produced by the subject) and xenografts (e.g., foreign cell samples implanted into a subject).

[0047] As used herein, the term "unmasking" or "unmasking" is commonly used in the field of tissue diagnostics to refer to retrieving antigens or targets and / or improving the detection of antigens, amino acids, peptides, proteins, nucleic acids, and / or other targets in fixed tissue. For example, it is believed that antigenic sites that may otherwise go undetected can be revealed, for example, by disrupting some of the protein bridges surrounding the antigen during unmasking. In some embodiments, the antigen and / or other target is unmasked by application of one or more unmasking agents, heat, and / or pressure. In some embodiments, only one or more unmasking agents are applied to the specimen to achieve unmasking (e.g., no heat or pressure is required). In other embodiments, only heat is applied to achieve unmasking. In some embodiments, unmasking may be performed only in the presence of water and additional heat. Unmasking is described in U.S. Patent Application Publication Nos. 2009 / 0170152 and 2009 / 0104654, the disclosures of which are incorporated herein by reference in their entireties. Examples of suitable unmasking agents are described herein.

[0048] overview

[0049] The present disclosure describes systems and methods for predicting the unmasking state of a biological specimen. As used herein, the term "unmasking state" refers to the unmasking state of a biological specimen that has been subjected to one or more unmasking processes. The unmasking state is believed to be highly dependent on the conditions to which the biological specimen was subjected during the unmasking process, such conditions including temperature, time (e.g., duration of the unmasking process), pressure, one or more unmasking agents utilized to achieve unmasking, whether external energy (UV, microwave) was provided to the specimen, etc. Thus, the unmasking state provides insight into the extent and / or quality of the unmasking process, and any unmasking state "output" can include, for example, the predicted duration of the unmasking process or the temperature at which the unmasking process was performed.

[0050] In view of the above, and as further described herein, in some embodiments, the predicted unmasking state can include a quantitative estimate of the temperature and / or duration of the unmasking process that the biological specimen underwent. In other embodiments, the predicted unmasking state can include a qualitative estimate of the extent or quality of the unmasking process, such as, for example, whether the antigen or target was not unmasked, partially unmasked, acceptably unmasked, or completely unmasked. In some embodiments, the qualitative estimate can be binary, for example, unmasked or not unmasked. By way of example, with reference to FIG. 15B, a data point of about 98.6° C. / about 100° C. may be considered a data point that is “unrecovered” and a data point of about 140° C. / about 140° C. may be considered a data point that is “overrecovered”. Such data points can be used to train an unmasking state estimation engine.

[0051] In some embodiments, the unmasking state of the biological analyte is predicted based on identifying unmasking features of the spectral data acquired from the biological analyte. In some embodiments, the unmasking features in the spectral data acquired from the biological analyte are identified using a trained unmasking state estimation engine. The unmasking state can be derived based on these identified unmasking features. In some embodiments, the identified unmasking features correspond to conformational changes in the composition of the biological analyte as a result of the unmasking process. For example, unmasking features corresponding to the anatomy of the tissue and / or the availability of immune recognition sites within the tissue can be identified. The availability of immune recognition sites is believed to be related to the conformational state of the protein or nucleic acid as the conformational state of the protein or nucleic acid changes during the unmasking process.

[0052] The present disclosure also describes systems and methods for training a demasking state estimation engine to enable quantitative and / or qualitative determination of the demasking state based on ground truth data, such as training spectral data including class labels, as further described herein. In some embodiments, the training spectral data includes differentially demasked biological analytes.

[0053] system

[0054] At least some embodiments of the present disclosure relate to computer systems and methods for analyzing spectral data obtained from a biological specimen that has been at least partially subjected to an unmasking process, which can be either an antigen retrieval process (a process utilized to render an antigen site recognizable by a specific binding entity such as an antibody probe) or a target retrieval process (a process utilized to render a nucleic acid recognizable by a specific binding entity such as a nucleic acid probe). According to the present disclosure, a trained unmasking state estimation engine can be used to provide a quantitative and / or qualitative estimate of the unmasking state of a test biological specimen, for example, a system of the present disclosure can receive input test spectral data from a test biological specimen and then provide as output an unmasking state, for example, a quantitative estimate of the unmasking temperature, a quantitative estimate of the unmasking duration, and / or a qualitative estimate of the unmasking degree or quality.

[0055] In other embodiments, the output provided by the disclosed system and method can be used to evaluate whether a particular test biological specimen is sufficiently unmasked before application of one or more specific binding entities or before performing any further downstream processing steps. It is believed that different biomarkers can show different responses to increasing unmasking treatments. For example, as described in the examples and figures herein, Ki-67 increases in staining intensity and number of labeled cells, followed by a decrease in intensity and positivity. Conversely, C4d continues to increase in intensity and positivity throughout the duration of the applied unmasking process, even under unmasking conditions that would otherwise damage the biological specimen. Thus, in some embodiments, the disclosed system allows for the determination of whether the unmasking process applied to the biological specimen provides a sufficiently unmasked specimen before downstream operations, such as staining in an IHC or ISH process. Indeed, in the context of staining with Ki-67 as described above, one skilled in the art would be able to interpret the results of actual Ki-67 staining taking into account the predicted unmasking state of a biological specimen based on the actual unmasking state of the specimen (regardless of the unmasking conditions (time and duration) reported by the laboratory).

[0056] In some embodiments, the output, e.g., the predicted unmasking state, can be used to verify the reported unmasking state of the sample, e.g., if a test biological specimen is received at a laboratory with a report indicating that the test biological specimen was subjected to an unmasking process at about 120° C. for about 30 minutes, the output from the systems and methods of the present disclosure can be used to verify the reported unmasking state.

[0057] In some embodiments, the output may be in the form of a generated report. In other embodiments, the output may be an overlay superimposed on an image of the test biological specimen. In still other embodiments, any output may be stored in a memory coupled to the system (e.g., storage system 240), and the output may be associated with the test biological specimen and / or other patient data.

[0058] A system 200 for acquiring spectral data, e.g., vibrational spectral data, and analyzing biological specimens (including test and training biological specimens) is shown in Figures 1 and 2. The system can include a spectral acquisition device 12, such as one configured to acquire a vibrational spectrum (e.g., a mid-IR spectrum or a Raman spectrum) of a biological specimen (or any portion thereof), and a computer 14, whereby the spectral acquisition device 12 and the computer can be communicatively coupled to each other (e.g., directly or indirectly via a network 20). The computer system 14 can include a desktop computer, laptop computer, tablet, etc., digital electronic circuitry, firmware, hardware, memory 201, a computer storage medium (240), a computer program or set of instructions (e.g., a program stored in the memory or storage medium), one or more processors (209) (including a programmed processor), and any other hardware, software, firmware modules or combinations thereof (as further described herein). For example, the system 14 shown in Figure 1 can include a computer with a display device 16 and an enclosure 18. The computer system may store the acquired spectral data locally, such as in a memory, a server, or another networked device.

[0059] Vibrational spectroscopy involves transitions due to the absorption or emission of electromagnetic radiation. These transitions range from 102 to 104 cm -1and are believed to originate from the vibration of the nuclei that make up the molecules in any given sample. It is believed that the chemical bonds within a molecule can vibrate in many ways, each vibration being called a vibrational mode. There are two types of molecular vibrations: stretching and bending. Stretching vibrations are characterized by a movement along the bond axis with an increase or decrease in the interatomic distance, while bending vibrations consist of a change in the bond angle relative to the rest of the molecule. Two widely used spectroscopic techniques based on vibrational energy are Raman spectroscopy and infrared spectroscopy. Both methods give complementary information and are based on the fact that within any molecule the atoms vibrate at some distinct well-defined frequencies characteristic of that molecule. A sample, when cleared against a beam of incident radiation, absorbs energy at frequencies characteristic of the frequencies of vibration of the chemical bonds present in the molecule. This absorption of energy by the vibration of the chemical bonds results in an infrared spectrum.

[0060] Although IR and Raman spectroscopy measure the vibrational energy of molecules, both methods rely on different selection rules, e.g., absorption processes and scattering effects. Their contrast mechanisms are different and each methodology has its own advantages and disadvantages, but the spectra obtained from each modality are often correlated (see, e.g., Figures 26 and 27).

[0061] IR spectroscopy is based on the absorption of electromagnetic radiation, whereas Raman spectroscopy relies on the inelastic scattering of electromagnetic radiation. Infrared spectroscopy offers a number of analytical tools, ranging from absorption to reflection and dispersion techniques, extending over a wide range of wavenumbers and including the near, mid and far infrared regions, where the different bonds present in the sample molecules provide a number of common and characteristic bands suitable for use for both qualitative and quantitative purposes. The sample is illuminated by IR light in IR spectroscopy, and vibrations induced by the electric dipole moment are detected.

[0062] Raman spectroscopy is a scattering phenomenon that occurs due to the difference between the incident and scattered radiation frequencies. It utilizes scattered light to gain knowledge about molecular vibrations, which can provide information about the structure, symmetry, electronic environment, and bonding of molecules. In Raman spectroscopy, a sample is illuminated by monochromatic visible or near-IR light from a laser source, and its vibrations during changes in electric polarizability are determined.

[0063] Any spectral acquisition device can be utilized in the system of the present disclosure. Examples of spectral acquisition devices or components of such devices suitable for use in acquiring mid-infrared spectra are described in US Patent Application Publication No. 2018 / 0109078a and US Patent Application Publication No. 2016 / 0091704, as well as US Patent No. 10,041,832, US Patent No. 8,036,252, US Patent No. 9,046,650, US Patent No. 6,972,409, and US Patent No. 7,280,576, the disclosures of which are incorporated herein by reference in their entirety.

[0064] Any method suitable for generating a representative mid-IR spectrum of a sample can be used. Fourier transform infrared spectroscopy and its biomedical applications are described, for example, in P. Lasch, J. Kneipp (Eds.) Biomedical Vibrational Spectroscopy'' 2008 (John Wiley & Sons). More recently, however, tunable quantum cascade lasers have enabled rapid spectroscopy and microscopy of biomedical specimens due to their high spectral power density (see N. Kroger et al.: Biomedical Vibrational Spectroscopy VI: Advances in Research and Industry, edited by A. Mahadevan-Jansen, W. Petrich, Proc. of SPIE Vol. 8939, 89390Z; N. Kroger et al., J. Biomed. Opt. 19 (2014) 111607; N. Kroger-Lui et al., Analyst 140 (2015) 2086). The contents of each of these publications are incorporated herein by reference in their entirety. This work is believed to constitute a major breakthrough for applicability (compared to the aforementioned infrared microscope setups) in that the investigations are much faster (e.g., 5 min instead of 18 h), do not require liquid nitrogen cooling, and offer more pixels per image at substantially lower cost. One particular advantage of QCL-based microscopy in the context of quality assessment of unstained tissues is the larger field of view (compared to FT-IR imaging), enabled by, for example, a 640x480 pixel microbolometer array detector.

[0065] In some embodiments, the spectrum can be obtained over a broad wavelength range, one or more narrow wavelength ranges, or even just a single wavelength, or a combination thereof. For example, a spectrum of the amide I band and the amide II band can be obtained. As another example, a spectrum can be obtained from about 3200 to about 3400 cm -1 , about 2800 to about 2900 cm -1 , about 1020 to about 1100 cm -1, and / or from about 1520 to about 1580 cm -1 In some embodiments, the spectrum may be acquired over a range of wavelengths from about 3200 to about 3400 cm -1 In some embodiments, the spectrum can be acquired over a range of wavelengths from about 2800 to about 2900 cm -1 In some embodiments, the spectrum can be acquired over a range of wavelengths from about 1020 to about 1100 cm -1 In some embodiments, the spectrum can be acquired over a range of wavelengths from about 1520 to about 1580 cm -1 A narrower spectral range is usually considered advantageous in terms of acquisition speed, especially when using quantum cascade lasers. In some embodiments, a single tunable laser is tuned to each wavelength in turn. Alternatively, a set of fixed-frequency non-tunable lasers can be used, such that wavelength selection is performed by switching on and off the laser required for measurement at a particular frequency.

[0066] The spectrum can be obtained, for example, using transmission or reflection measurements. For transmission measurements, barium fluoride, calcium fluoride, silicon, thin polymer films, or zinc selenide are typically used as substrates. For reflection measurements, gold or silver plated substrates are common, as well as standard microscope glass slides, or glass slides coated with mid-IR reflective coatings (e.g., multi-layer dielectric coatings or thin sliver coatings). In addition, means for using surface enhancements (e.g., SEIRS), such as structured surfaces like nanoantennas, may be implemented.

[0067] Those skilled in the art will appreciate that other computer devices or systems can be utilized, and that the computer systems described herein can be communicatively coupled to additional components, such as, for example, microscopes, imaging devices, scanners, other imaging systems, automated slide preparation devices, etc. Some of these additional components, as well as the various computers, networks, etc. that can be utilized, are described further herein.

[0068] For example, in some embodiments, the system 200 may further include an imaging device, and any images captured from the imaging device may be stored in a binary format, such as locally or on a server. In some embodiments, the captured image may be stored, such as in the storage subsystem 240, along with the unmasked state estimate and / or any patient data. The captured digital image may also be divided into a matrix of pixels. A pixel may include a digital value of one or more bits defined by a bit depth. In general, an imaging device (or other image source, including pre-scanned images stored in memory) may include, but is not limited to, one or more image capture devices. Image capture devices include, but are not limited to, cameras (e.g., analog cameras, digital cameras, etc.), optical systems (e.g., one or more lenses, a sensor focus lens group, a microscope objective lens, etc.), imaging sensors (e.g., charge-coupled devices (CCDs), complementary metal oxide semiconductor (CMOS) image sensors, etc.), photographic film, etc. In digital embodiments, the image capture devices may include multiple lenses that cooperate to demonstrate on-the-fly focusing. An image sensor, such as a CCD sensor, may capture a digital image of the specimen.

[0069] In some embodiments, the imaging device is a bright field imaging system, a multispectral imaging (MSI) system, or a fluorescence microscope system. The digitized tissue data can be generated by an image scanning system, such as, for example, a VENTANA DP200 scanner by VENTANA MEDICAL SYSTEMS, Inc. (Tucson, AZ) or other suitable imaging device. Additional imaging devices and systems are further described herein. Those skilled in the art will understand that the digital color image acquired by the imaging device is conventionally composed of basic color pixels. Each colored pixel can be coded with three digital components, each containing the same number of bits, and each component generally corresponds to a primary color, red, green, or blue, also referred to by the term "RGB" components.

[0070] 2 provides an overview of the disclosed system 200 and various modules utilized therein. In some embodiments, the system 200 employs a computing device or computer-implemented method having one or more processors 209 and one or more memories 201 that store non-transitory computer-readable instructions for execution by the one or more processors to cause the one or more processors to perform certain instructions described herein.

[0071] In some embodiments, as described above, the system includes a spectral acquisition module 202 for acquiring a vibrational spectrum, such as a mid-IR spectrum or a Raman spectrum, of the acquired biological specimen (e.g., see step 310 of FIG. 3) or any portion thereof (e.g., see step 320 of FIG. 3). In some embodiments, the system 200 further includes a spectral processing module 212 adapted to process the acquired spectral data. In some embodiments, the spectral processing module 212 is configured to pre-process the spectral data. In some embodiments, the spectral processing module 212 corrects and / or normalizes the acquired spectrum or converts the acquired transmission spectrum to an absorption spectrum. In other embodiments, the spectral processing module 212 is configured to average multiple spectra acquired from a single biological specimen. In still other embodiments, the spectral processing module 212 is configured to further process the acquired spectrum, such as to calculate first derivatives, second derivatives, etc. of the acquired spectrum.

[0072] In some embodiments, the system 200 further includes a training module 211 adapted to receive training spectral data and train the unmasking state estimation engine 210 using the received training spectral data.

[0073] In some embodiments, the system 200 includes an unmasking state estimation engine 210 trained to detect unmasking features in the test spectral data (e.g., see step 340 of FIG. 3 ) and provide an estimate of the unmasking state of the biological analyte based on the detected unmasking features (e.g., see step 350 of FIG. 3 ). In some embodiments, the unmasking state estimation engine 210 includes one or more machine learning algorithms. In some embodiments, the one or more machine learning algorithms are based on dimensionality reduction as further described herein. In some embodiments, the dimensionality reduction utilized principal component analysis, such as principal component analysis with discriminant analysis. In other embodiments, the dimensionality reduction is a projection onto latent structure regression. In some embodiments, the unmasking state estimation engine 210 includes a neural network. In other embodiments, the unmasking state estimation engine 210 includes a classifier, such as a support vector machine.

[0074] One of ordinary skill in the art will also appreciate that additional modules can be incorporated into the workflow or system 200. In some embodiments, an image acquisition module is executed to acquire a digital image of the biological specimen or any portion thereof. In other embodiments, automated image analysis algorithms can be executed such that cells can be detected, classified, and / or scored (see, e.g., U.S. Patent Application Publication No. 2017 / 0372117, the disclosure of which is incorporated herein by reference in its entirety).

[0075] Spectroscopic acquisition module and acquired spectroscopic data

[0076] 2, in some embodiments, the system 200 executes a spectrum acquisition module 202 to acquire (e.g., using a spectral imaging device 12 such as any of those described above) a vibrational spectrum from at least a portion of a biological specimen (e.g., a test biological specimen or a training biological specimen). In some embodiments, the biological specimen is unstained, e.g., the biological specimen does not include a stain indicative of the presence of a biomarker. In other embodiments, the biological specimen may also include one or more stains, e.g., a primary stain, or a stain indicative of the presence of one or more biomarkers (e.g., proteins, nucleic acids). Once a spectrum is acquired using the spectrum acquisition module 202, the acquired spectrum can be stored in a storage module 240 (e.g., a local storage module or a network storage module).

[0077] In some embodiments, the vibrational spectrum can be obtained from a portion of the biological specimen (as described further herein, this is regardless of whether the specimen is a training biological specimen or a test biological specimen). In such cases, the spectrum acquisition module 202 may be programmed to obtain a vibrational spectrum from a predetermined portion of the sample, for example by random sampling or by sampling at regular intervals across a grid covering the entire sample. This can also be useful when only certain regions of the sample are relevant for analysis. For example, a region of interest may contain a particular type of tissue, or a relatively high population of a particular type of cell compared to another region of interest. For example, a region of interest may be selected that contains tonsillar tissue but not connective tissue. In such cases, the spectrum acquisition module 202 may be programmed to obtain a vibrational spectrum from a predetermined portion of the sample, for example by random sampling or by sampling at regular intervals across a grid covering the entire sample. In embodiments where the sample contains one or more stains, vibrational spectra can be obtained from regions of interest that contain no stain or that contain relatively less stain than other regions.

[0078] In some embodiments, at least two regions of the biological specimen are sampled and a vibrational spectrum is obtained for each of the at least two regions (and this is regardless of whether the specimen is a training biological specimen or a test biological specimen). In other embodiments, at least 10 regions of the biological specimen are sampled and a vibrational spectrum is obtained for each of the at least 10 regions. In still other embodiments, at least 30 regions of the biological specimen are sampled and a vibrational spectrum is obtained for each of the at least 30 regions. In further embodiments, at least 60 regions of the biological specimen are sampled and a vibrational spectrum is obtained for each of the at least 60 regions. In still further embodiments, at least 90 regions of the biological specimen are sampled and a vibrational spectrum is obtained for each of the at least 90 regions. In still further embodiments, between about 30 and about 150 regions of the biological specimen are sampled and a vibrational spectrum is obtained for each region.

[0079] In some embodiments, a single vibrational spectrum is acquired per region of the biological specimen, in other embodiments, at least two vibrational spectra are acquired per region of the biological specimen, and in yet other embodiments, at least three vibrational spectra are acquired per region of the biological specimen.

[0080] In some embodiments, the acquired spectra or acquired spectral data (used interchangeably herein) stored in the storage module 240 comprises “training spectral data.” In some embodiments, the training spectral data is derived from a training biological specimen, which may be a histological specimen, a cytological specimen, or any combination thereof. In some embodiments, the training spectral data is used to train the unmasking state estimation engine 210, such as by use of the training module 211 described herein. In some embodiments, the training spectral data includes class labels, such as unmasking time, unmasking duration, relative quality information (e.g., unrecovered, fully recovered, partially recovered), etc. In some embodiments, the training spectral data includes multiple class labels.

[0081] In some embodiments, the training biological specimen is not differentially masked. Differential unmasking refers to the process where a single training biological specimen is subjected to multiple masking steps (e.g., a first training tissue sample, a second training tissue sample, and 第 n training tissue samples) (see, e.g., FIG. 4A) and each portion of the plurality of portions is subjected to a different unmasking treatment (see, e.g., FIG. 4B). For example, a single tonsil tissue sample can be divided into 10 or more portions, and each portion can be unmasked for a predetermined time and at a predetermined temperature. In some embodiments, the sample can be partitioned into three or more portions, and each portion can be unmasked under a different set of unmasking conditions. In other embodiments, the sample can be partitioned into five or more portions, and each portion can be unmasked under a different set of unmasking conditions. In yet other embodiments, the sample can be partitioned into seven or more portions, and each portion can be unmasked under a different set of unmasking conditions. In further embodiments, the sample can be partitioned into nine or more portions, and each portion can be unmasked under a different set of unmasking conditions. This process is illustrated in FIG. 4A.

[0082] In some embodiments, multiple training samples derived from a single training biological specimen are each unmasked at the same temperature for different durations. For example, each training tissue sample derived from a single training biological specimen can be unmasked at the same temperature (e.g., about 98.6° C.), but the duration of unmasking can vary (about 5 minutes, about 30 minutes, about 60 minutes, etc.).

[0083] In other embodiments, multiple training samples derived from a single biological specimen are each unmasked for the same duration but at different temperatures. For example, each training sample can be unmasked for the same duration (e.g., 10 minutes) but at varying unmasking temperatures (about 98.6°C, about 110°C, about 120°C, about 130°C, etc.).

[0084] Of course, one skilled in the art will recognize that both the unmasking time and temperature can be varied. As in the embodiment described above, a first set of training samples can be unmasked at a first temperature but for different durations to provide a first set of training samples. A second and third set of training samples can be unmasked at a second and third temperature, respectively, and for different durations to provide a second and third set of training samples. This process is illustrated in FIG. 4B. This process can be repeated for multiple biological specimens, such as biological specimens having the same or different tissue types, to build a database (such as a database stored in the storage module 240) that includes multiple training spectral data sets that can be used in training the unmasking state estimation engine. One skilled in the art will also appreciate that temperature and duration are not the only unmasking conditions that can be altered. Indeed, various unmasking reagents, pressures, etc. can be utilized in any unmasking process, and such conditions can be altered to provide training spectral data sets that take such process conditions into account.

[0085] The process of differential unmasking and obtaining spectral data from the differentially unmasked samples is further illustrated in FIG. 4C. As described above, first, one or more training biological samples are obtained (step 410). Then, each of the one or more training biological samples is divided into at least two parts (step 420). In this manner, each of the one or more training biological samples provides at least two "training samples". Each of these training samples may be unmasked separately, e.g., each may be unmasked for a different predetermined amount of time or for a different predetermined duration (step 430). Following the differential unmasking of the at least two training samples, a plurality of regions in each of the at least two training samples are identified (step 440).

[0086] Next, at least one vibration spectrum is obtained for each of the identified regions of the plurality of identified regions (step 450). An average of each vibration spectrum obtained from each identified region (or a further processed version thereof as further described below) is calculated to provide an average vibration spectrum for that training sample (step 460). Steps 400 to 460 may be repeated for a plurality of different training biological specimens (see dotted line 470). In some embodiments, the average vibration spectrum from all training samples from all training biological specimens "training spectrum data" is stored (step 480), such as in the storage module 240. In this manner, the training spectrum data can be retrieved from the storage module 240 by the training module 211 for training the unmasking state estimation engine 210. In addition to storing the average vibration spectrum from all training samples, the storage module 240 is also adapted to store any class label (e.g., unmasking temperature, unmasking duration, etc.) associated with the average vibration spectrum.

[0087] In some embodiments, each of the training samples is stained for the presence of one or more biomarkers so that functional staining intensity can be assessed for each training sample. In some embodiments, each training sample is stained for the presence of a single biomarker, and then images of the sample are captured and analyzed using an imaging device (e.g., for staining intensity and / or positivity). In other embodiments, each training sample is stained for the presence of two or more biomarkers, and then images of the sample are captured and analyzed using an imaging device (so that the staining intensity and / or positivity of each of the two or more biomarkers are analyzed independently). In some embodiments, multiple biomarkers are tested for functional staining information because each biomarker can have a different response to the unmasking process. For example, Ki-67 and C4d biomarkers can behave differently when introduced into unmasked samples under different conditions. For example, Ki-67 is believed to increase in staining intensity as the degree of unmasking increases, but only up to a certain point, after which the intensity and positivity decrease. Conversely, C4d continues to increase in intensity and positivity through unmasking conditions that would otherwise damage the sample.

[0088] The above-described process for preparing training biological specimens and acquiring spectral data from such specimens may be repeated for multiple different training biological specimens, each of which may be of the same tissue type or different tissue types (e.g., tonsil tissue or breast tissue). The Examples section herein further describes methods for preparing training biological specimens and acquiring spectral data for use in training the unmasking state estimation engine 210, such as for predicting the unmasking state of a test biological specimen derived from a pair of tonsil tissue and breast tissue. Additionally, the above-described process may be repeated for different unmasking reagents, or for antigen retrieval and target retrieval processes (the examples provided herein show that the systems and methods of the present disclosure are equally applicable to antigen retrieval and target retrieval processes, providing accurate results in both use cases).

[0089] In some embodiments, the acquired spectral data stored in storage module 240 includes “test spectral data.” In some embodiments, the test spectral data is derived from a test biological specimen, such as a specimen derived from a subject (e.g., a human patient), which can be a histological specimen, a cytological specimen, or any combination thereof.

[0090] With reference to FIG. 5, a test biological specimen can be acquired (step 510) and then a number of spatial regions within the test biological specimen can be identified (step 520). At least one vibrational spectrum can be acquired for each identified region (step 530). The vibrational spectra acquired from all regions can then be corrected, normalized, and averaged to provide an average vibrational spectrum of the test biological specimen ("test spectral data"). As described further herein, the test spectral data can be provided to a trained demasking state estimation engine 210 to predict the demasking state of the test biological specimen. The predicted demasking state can then be used in downstream processes or downstream decision making, for example, to determine whether the test specimen was properly demasked prior to deposition of one or more specific binding entities.

[0091] As discussed above, whether the spectral data is acquired from a training or test biological specimen, multiple vibrational spectra are acquired for each biological specimen, e.g., to account for spatial heterogeneity of the sample. In some embodiments, the spectral processing module 212 is first utilized to convert each acquired vibrational transmission spectrum into a vibrational absorption spectrum. In some embodiments, the transmission and absorbance spectra are directly related via the formula absorbance=ln(blank transmission / transmission through tissue), and thus, the acquired transmission spectra can be converted into an absorption spectrum. Once all of the vibrational spectra have been converted from transmission to absorbance, in some embodiments, the spectral processing module 212 averages all of the spectra acquired from all of the various regions, which is the averaged vibrational spectrum that is used for downstream analysis, e.g., training or predicting the unmasked state. In some embodiments, with reference to FIG. 6, the vibrational spectra acquired from each of the multiple spatial regions are first normalized and / or corrected prior to their averaging. In some embodiments, the vibrational spectrum from each region is individually corrected (step 620) to provide a corrected vibrational spectrum. For example, the correction may include correcting each acquired vibration spectrum for atmospheric effects (step 630), and then correcting each atmospherically corrected vibration spectrum for scattering (step 640). Each corrected vibration spectrum is then normalized (step 650), e.g., to a maximum of 2, to mitigate differences in specimen thickness and tissue density. The set of amplitude-normalized spectra is then averaged (step 660).

[0092] Unmasking state estimation engine

[0093] The disclosed systems and methods mine spectral data using machine learning techniques. In the case of an unmasking state estimation engine in training mode, the unmasking state estimation engine can learn features from a plurality of acquired and processed training spectra (such as training spectra stored in the storage module 240) and correlate those learned features with class labels associated with the training spectra (e.g., known unmasking temperature, known unmasking duration, tissue quality, etc.). In the case of a trained unmasking state estimation engine, the trained unmasking state engine can derive unmasking features from the test biological specimen and, based on the learned dataset, predict the unmasking state of the test biological specimen based on the derived unmasking features.

[0094] Machine learning can generally be defined as a type of artificial intelligence (AI) that provides computers with the ability to learn without being explicitly programmed. Machine learning focuses on developing computer programs that can teach themselves to grow and change when exposed to new data. In other words, machine learning can be defined as a subfield of computer science that gives computers the ability to learn without being explicitly programmed. Machine learning explores the study and construction of algorithms that can learn from and make predictions on data, and such algorithms overcome strictly following static program instructions by making data-driven predictions or decisions by building models from sample inputs. The machine learning described herein can be further performed as described in "Introduction to Statistical Machine Learning" by Sugiyama, Morgan Kaufmann, 2016, p. 534; "Discriminative, Generative, and Imitative Learning" by Jebara, MIT Thesis, 2002, p. 212; and "Principles of Data Mining (Adaptive Computation and Machine Learning)" by Hand et al., MIT Press, 2001, p. 578, which are incorporated by reference as if fully set forth herein. The embodiments described herein can be further configured as described in these references.

[0095] In some embodiments, the unmasked state estimation engine 210 uses "supervised learning" for the task of predicting the unmasked state of a test spectrum derived from a test biological specimen. Supervised learning is a machine learning task that learns a function that maps inputs to outputs based on example input-output pairs. It infers the function from labeled training data (where the unmasked state is the label associated with the training spectrum data) consisting of a set of training examples (here, training spectra). In supervised learning, each example is a pair consisting of an input object (typically a vector) and a desired output value (also called a supervisory signal). The supervised learning algorithm analyzes the training data and generates an inference function that can be used to map new examples. The optimal scenario allows the algorithm to correctly determine the class label of unseen instances.

[0096] The unmasking state estimation engine 210 may include any type of machine learning algorithm known to those skilled in the art. Suitable machine learning algorithms include regression algorithms, similarity-based algorithms, feature selection algorithms, regularization method-based algorithms, decision tree algorithms, Bayesian models, kernel-based algorithms (e.g., support vector machines), clustering-based methods, artificial neural networks, deep learning networks, ensemble methods, and dimensionality reduction methods. Examples of suitable dimensionality reduction methods include principal component analysis (such as principal component analysis and discriminant analysis) and projection onto latent structure regression.

[0097] In some embodiments, the unmasking state estimation engine 210 utilizes principal component analysis. The main idea of ​​principal component analysis (PCA) is to reduce the dimensionality of a dataset consisting of many mutually correlated variables while retaining to the maximum extent the variation present in the dataset. The same is done by transforming the variables into a new set of variables, known as principal components (or simply PCs), which are ordered orthogonally so that the retention of the variation present in the original variables decreases as they move down the order. In this way, the first principal component retains the maximum variation present in the original components. Principal components are orthogonal because they are eigenvectors of the covariance matrix. Principal component analysis and methods of using it are described in US Patent Application Publication No. 2005 / 0123202 and US Pat. Nos. 6,894,639 and 8,565,488, the disclosures of which are incorporated herein by reference in their entirety. PCA and linear discriminant analysis are further described by Khan et al., “Principal Component Analysis-Linear Discriminant Analysis Feature Extractor for Pattern Recognition,” IJCSI International Journal of Computer Sciences Issues, Vol. 8, Issue 6, No. 2, November 2011, the disclosure of which is incorporated herein by reference in its entirety.

[0098] In some embodiments, the unmasking state estimation engine 210 utilizes projection on latent structure regression (PLSR). PLSR is a recent technique that combines and generalizes features of PCA and multiple linear regression. Its goal is to predict a set of dependent variables from a set of independent variables or predictors. This prediction is achieved by extracting a set of orthogonal factors, called latent variables, from the predictors that have the best predictive power. These latent variables can be used to create displays similar to PCA displays. The quality of predictions obtained from PLS regression models is evaluated with cross-validation techniques such as bootstrap and jackknife. There are two main variants of PLS ​​regression. The most common separates the roles of dependent and independent variables. The second gives the same role to dependent and independent variables. PLSR is further described by Abdi, "Partial Least Squares Regression and Projection on Latent Structure Regression (PLS Regression)", WIREs Computational Statistics, John Wiley & Sons, Inc., 2010, the disclosure of which is incorporated herein by reference in its entirety. The Examples section provided herein describes a PLSR-based trained demasking state estimation engine and shows that the PLSR-based trained demasking state estimation engine 210 can be used to provide at least a quantitative estimate of the demasking temperature and / or the demasking duration.

[0099] In some embodiments, the unmasking state estimation engine 210 utilizes reinforcement learning. Reinforcement learning (RL) is a type of machine learning method in which an agent receives a delayed reward in the next time step to evaluate its previous action. In other words, RL is a model-free machine learning paradigm on how a software agent should take action in an environment to maximize the notion of cumulative reward. Typically, an RL setting consists of two components: an agent and an environment. The environment refers to the object on which the agent is acting, and the agent represents the RL algorithm. The environment starts by sending a state to the agent, and then the agent performs an action in response to that state based on its knowledge. The environment then sends back the next state and reward pair to the agent. The agent updates its knowledge with the reward returned by the environment to evaluate its final action. The loop continues until the environment sends a terminal state, which ends with the episode. Reinforcement learning algorithms are further described in U.S. Pat. Nos. 10,279,474 and 7,395,252, the disclosures of which are incorporated herein by reference in their entireties.

[0100] In some embodiments, the machine learning algorithm is a support vector machine ("SVM"). In general, SVM is a classification technique based on statistical learning theory in which a nonlinear input data set is transformed into a high-dimensional linear feature space through a kernel in the nonlinear case. A support vector machine projects a set of training data E representing two different classes into a high-dimensional space by a kernel function K. In this transformed data space, the nonlinear data is transformed so that a flat line can be generated to separate the classes so as to maximize class separation (discrimination hyperplane). The test data is then projected into the high-dimensional space through K, and the test data (such as the features or metrics listed below) are classified based on where they are located relative to the hyperplane. The kernel function K defines how to project the data into the high-dimensional space.

[0101] In some embodiments, the unmasking state estimation engine 210 includes a neural network. In some embodiments, the neural network is configured as a deep learning network. Generally speaking, "deep learning" is a branch of machine learning based on a set of algorithms that attempt to model high-level abstractions in data. Deep learning is part of a broad family of machine learning methods based on learning representations of data. Observations can be represented in many ways, such as vectors of intensity values ​​per pixel, or in more abstract ways as sets of edges, regions of particular shapes, etc. Some representations are better than others in simplifying the learning task. One of the promises of deep learning is replacing handcrafted features with efficient algorithms for unsupervised or semi-supervised feature learning and hierarchical feature extraction.

[0102] In some embodiments, the neural network is a generative network. A "generative" network can be generally defined as a model that is probabilistic in nature. In other words, a "generative" network does not perform forward simulation or rule-based techniques. Instead, a generative network can be trained (in that it can learn its parameters) based on a set of appropriate training data (e.g., multiple training spectral data sets). In some embodiments, the neural network is configured as a deep generative network. For example, the network can be configured to have a deep learning architecture in that the network can include multiple layers that perform multiple algorithms or transformations.

[0103] In some embodiments, the neural network includes an autoencoder. An autoencoder neural network is an unsupervised learning algorithm that applies backpropagation and sets a target value equal to the input. The goal of an autoencoder is to learn a representation (encoding) of a dataset, typically for dimensionality reduction, by training the network to ignore signal "noise". Along with the reduction side, a reconstruction side is trained, and the autoencoder attempts to generate a representation from the reduced encoding that is as close as possible to its original input. Additional information regarding autoencoders can be found at http: / / ufldl.stanford.edu / tutorial / unsupervised / Autoencoders / , the disclosure of which is incorporated herein by reference in its entirety.

[0104] In some embodiments, the neural network may be a deep neural network with a set of weights that model the world according to the data fed to train it. A neural network typically consists of multiple layers, with signal paths traversing back and forth between the layers. Any neural network can be implemented for this purpose. Suitable neural networks include LeNet, AlexNet, ZFnet, GoogLeNet, VGGNet, VGG16, DenseNet, and ResNet. In some embodiments, a fully convolutional neural network is utilized, as described in Long et al., "Fully Convolutional Networks for Semantic Segmentation," Computer Vision and Pattern Recognition (CVPR), 2015 IEEE Conference, June 2015 (INSPEC Accession No.: 15524435), the disclosure of which is incorporated herein by reference.

[0105] In some embodiments, the neural network is configured as AlexNet. For example, the classification network structure can be AlexNet. The term "classification network" is used herein to refer to a CNN that includes one or more fully connected layers. In general, AlexNet includes several convolutional layers (e.g., 5) followed by several fully connected layers (e.g., 3) that are combinatorially configured and trained to classify data.

[0106] In another embodiment, the neural network is configured as GoogleNet. The GoogleNet architecture can include a relatively large number of layers (especially compared to some other neural networks described herein), but some of the layers can operate in parallel, and a group of layers that work in parallel with each other is generally referred to as an initiation module. Other layers may operate sequentially. Thus, GoogleNet differs from other neural networks described herein in that not all of the layers are arranged in a sequential structure. An example of a neural network configured as GoogleNet is described in "Going Deeper with Convolutions" by Szegedy et al., CVPR 2015, which is incorporated by reference as if fully set forth herein.

[0107] In another embodiment, the neural network is configured as a VGG network. For example, the classification network structure can be VGG. The VGG network was created by increasing the number of convolutional layers while holding other parameters of the architecture fixed. Increasing the depth by adding convolutional layers is made possible by using substantially small convolutional filters in all layers.

[0108] In other embodiments, the neural network is configured as a deep residual network. For example, the classification network structure can be a Deep Residual Net or ResNet. As with some other networks described herein, a deep residual network can include a convolutional layer followed by a fully connected layer, which are combined and configured and trained for detection and / or classification. In a deep residual network, the layers are configured to learn a residual function with reference to the layer inputs, instead of learning an unreferenced function. In particular, instead of hoping that each of several stacked layers directly conforms to a desired underlying mapping, these layers are allowed to explicitly conform to a residual mapping realized by a feedforward neural network with shortcut connections. A shortcut connection is a connection that skips one or more layers. A deep residual net can be created by taking a plain neural network structure including a convolutional layer and inserting shortcut connections, thereby taking the plain neural network and making it a residual learning partner. An example of a deep residual net is described in "Deep Residual Learning for Image Recognition" by He et al., NIPS 2015, which is incorporated by reference as if fully set forth herein. The neural networks described herein can be further configured as described in this reference.

[0109] Training the unmasking state estimation engine

[0110] In some embodiments, the unmasking state estimation engine 210 is adapted to operate in a training mode. In some embodiments, the training module 211 can be operable to provide training spectral data to the unmasking state estimation engine 210 according to any suitable training algorithm and operate the unmasking state estimation engine 210 in the training mode. In some embodiments, the training module 211 is in communication with the unmasking state estimation engine 210 and is configured to receive training spectral data (or further processed variants of the training absorbance spectral data, such as first or second derivatives of the training spectral data, magnitudes of individual bands in the training spectral data, integrals of bands in the training spectral data, ratios of two or more band intensities in the training spectral data, ratios from second and third derivatives of the training spectral data, etc.) and provide the training spectral data to the unmasking state estimation engine 210. In some embodiments, the training module 211 is also adapted to provide class labels associated with the training spectral data.

[0111] In some embodiments, the training algorithm utilizes a set of known training spectral data (as described herein) and a corresponding set of known output class labels (e.g., unmasking conditions), and is configured to modify internal connections within the unmasking state estimation engine 210 such that processing of the input training spectral data provides the desired corresponding class labels.

[0112] The unmasking state estimation engine 210 can be trained according to any method known to one of skill in the art. Additional examples of indicators are disclosed in U.S. Patent Application Publication Nos. 2015 / 0268255, 2019 / 0102675, 2015 / 0356461, 2013 / 0132786, 2013 / 0240010, 2014 / 0134616, and 2019 / 0108344, the disclosures of which are incorporated herein by reference in their entireties.

[0113] In some embodiments, the unmasking state estimation engine 210 is trained using a cross-validation method. Cross-validation is a technique that can be used to assist in model selection and / or parameter tuning when developing a classifier. Cross-validation uses one or more subsets of cases from the set of labeled cases as a test set. For example, in k-fold cross-validation, the set of labeled cases is equally divided into k "folds", e.g., K-fold cross-validation is a resampling procedure used to evaluate machine learning models. A series of training-testing cycles are performed, iterating through the k folds such that in each cycle a different fold is used as the test set and the remaining folds are used as the training set. Since each fold is used as a test set at some point, the non-random selection of cases in the set of labeled cases appears to bias the cross-validation. For example, in a 5-fold cross-validation (k=5) scenario, the dataset is divided 5-fold. In the first iteration, the first fold is used to test the model and the remaining are used to train the model. In the second iteration, the second fold is used as a test set and the rest serve as training sets. This process is repeated until each fold of the five folds has been used as a test set. Additional methods of thresholding are described in WO2014 / 0279734 and WO2005 / 0234753, the disclosures of which are incorporated herein by reference in their entirety.

[0114] In the context of the unmasking state estimation engine 210 utilizing a PLSR-based machine learning algorithm, FIG. 7 and FIG. 11 show how a PLSR model is trained to mine vibrational spectra for features in the training spectra, such as features related to the conformational state of the immune recognition site, and how these features change as the unmasking conditions or unmasking process progresses. In some embodiments, the PLSR model is also trained to recognize changes in these features for different types of tissues and / or different types of molecules (proteins, nucleic acids). In some embodiments, the PLSR algorithm takes vibrational spectrum data (e.g., absorption spectra, first derivatives, second derivatives) and creates a model that is used to determine which features (wavelengths) are most predictive of the response variables (unmasking time / duration, etc.). In some embodiments, the generated model can be further evaluated for performance using the same unknown vibrational spectrum data for performance evaluation and optimization.

[0115] In the context of the unmasking state estimation engine 210 utilizing a machine learning algorithm based on principal component analysis, PCA is performed on an initial training dataset of a default sample size to generate a PCA transformation matrix. A second PCA is performed on a combined dataset including the initial training dataset and a test dataset. The number of samples in the initial training dataset is then incremented to generate an extended training dataset. A PCA of the extended training dataset is performed to determine whether the PCA number of the extended training dataset is the same as that of the initial training dataset. If so, the error between the initial test dataset and the extended test dataset is evaluated based on the PCA signal and the PCA transformation matrix to estimate a final solution error. The PCA matrix of the combined dataset is transformed back to the initial training dataset domain (e.g., the spectral domain) using the transformation matrix from the first PCA to generate a test dataset estimate. The method iterates by expanding the size of the training matrix until the PCA number converges and a final error target is achieved. When the error target is reached, the training dataset of the identified size adequately represents the training target feature information contained in the specified input parameter range. A machine learning system (e.g., the unmasking state estimation engine 210) can then be trained with a training matrix of the identified size. Additional aspects of training using PCA are described in U.S. Pat. Nos. 8,452,718 and 7,734,087, the disclosures of which are incorporated herein by reference in their entireties.

[0116] In embodiments in which the unmasking state estimation engine 210 includes a neural network, a backpropagation algorithm may be used to train the unmasking state estimation engine 210. Backpropagation is an iterative process in which the connections between network nodes are given some random initial values ​​and the network is operated to calculate the corresponding output vectors for a set of input vectors (the training spectral data set). The output vectors are compared to the desired outputs of the training spectral data set and the error between the desired and actual outputs is calculated. The calculated error is backpropagated from the output nodes to the input nodes and used to modify the values ​​of the network connection weights to reduce the error. After each such iteration, the training module 211 may calculate the total error for the entire training set, and then the training module 211 may repeat the process with another iteration. The training of the unmasking state estimation engine 210 is complete when the total error reaches a minimum value. If the minimum value of the total error is not reached after a predetermined number of iterations and if the total error is not constant, the training module 211 may consider the training process to not converge.

[0117] As described above, in the context of training with acquired spectral data acquired from training biological specimens that have been differentially unmasked (e.g., unmasked for different durations and / or at different temperatures), each acquired training spectrum is associated with a known unmasking condition, e.g., the duration of the unmasking process and the temperature at which the unmasking process was performed. For example, the two training spectral data sets shown in FIG. 4B (see dotted boxes) can be provided to the training module 211 for training the unmasking state estimation engine 210, along with any associated class labels. In other embodiments, the training spectral data can further include class labels, such as unrecovered, fully recovered, or partially recovered.

[0118] Once training of the demasking state estimation engine 210 is complete, the system 200 is ready to operate to detect features in the test spectral data and, based on the detected features, estimate the demasking state of the test biological specimen, e.g., the demasking conditions including the temperature and / or duration of the applied demasking process. In some embodiments, the demasking state estimation engine 210 can be periodically retrained to adapt to variations in the input data.

[0119] Estimation of the unmasking state

[0120] As described above, once the unmasking state estimation engine 210 is properly trained, it can be used to detect features in the test vibration spectral data and, based on the detected features, predict the unmasking state, e.g., duration and / or temperature of the unmasking process. In some embodiments, referring to FIG. 3, a test biological specimen is obtained (step 310) (e.g., from a subject suspected of having or known to have a particular disease), and then test vibration spectral data is obtained (step 320) from the test biological specimen (see also FIG. 5). In some embodiments, the test vibration spectral data includes an absorbance spectrum, first and / or second derivatives of the absorbance spectrum, magnitudes of individual bands in the training spectral data, integrals of bands in the training spectral data, ratios of two or more band intensities in the training spectral data, ratios from second and third derivatives of the training spectral data, etc.

[0121] Once the vibration test spectrum data and / or variants thereof described above have been acquired and processed, demasking features can be derived from the test spectrum data using the trained demasking state estimation engine 210 (step 340). In some embodiments, the derived demasking features are a mapping of how each wavenumber relates to a prediction of the recovery state. For example, with particular reference to a graph of PLSR coefficients, as shown in FIG. 11, a large positive coefficient means that the wavenumber is positively correlated with the search condition. Meanwhile, a large negative number means that the wavenumber is negatively correlated with the recovery condition. Values ​​close to 0 have little significance. In some embodiments, demasking features that can be detected include peak amplitudes, peak positions, peak ratios, sums of spectral values ​​(e.g., integrals over a particular spectral range), one or more changes in slope (first derivative) or curvature (second derivative), etc. Based on the derived demasking features, a demasking state estimate can be calculated (step 350).

[0122] In embodiments where the unmasking state estimation engine 210 is trained using training spectral data including known unmasking temperature and duration class labels, the unmasking state estimation engine 210 can provide as output a predicted temperature and / or duration of the unmasking process undergone by the test biological specimen. This concept is further described in Example 1 herein. By way of example, FIG. 12 and FIG. 14 show results achieved using the trained unmasking state estimation engine 210. FIG. 12 and FIG. 14 show predicted unmasking durations and temperatures compared to known (experimental) unmasking durations and temperatures of differentially unmasked test biological specimens. As shown, the unmasking state estimation engine 210 can accurately predict unmasking times and / or unmasking durations across individually unmasked specimens.

[0123] One skilled in the art will appreciate that the information provided as output may be used in further downstream processes and may be used to make decisions regarding whether to treat a test biological sample with one or more specific binding entities.

[0124] Example 1 - Estimating Unmasked States Using a Trained Unmasked State Estimation Engine

[0125] summary

[0126] This project utilized mid-infrared (mid-IR) spectroscopy to investigate the vibrational state of molecules in histological tissue sections. The testable hypothesis is that the vibrational state of tissue, as judged by the mid-IR signal, is indicative of conformational changes in tissue composition that occur during the antigen retrieval (AR) process.

[0127] A measurement has been developed that correlates changes in the mid-IR spectrum with the antigen retrieval state of the tissue. This measurement was used to examine the differences between low and high temperature retrieval. The mid-IR results were supported by staining results showing different staining intensities resulting from low and high temperature antigen retrieval, indicating that these treatments do not leave tissues in an equivalent state.

[0128] In this study, changes in mid-IR spectra resulting from differentially retrieved tonsillar tissues were studied. In particular, the two largest variables affecting antigen retrieval, time and temperature, were independently evaluated. The identified shifts in mid-IR spectra were correlated with immunohistochemical (IHC) staining for Ki-67 and C4d proteins. Results were validated in tonsillar and breast tissue, indicating that this measurement is generalizable to different types of tissue. Furthermore, it was shown that tissues are in dramatically different states after CC1-based antigen retrieval and CC2-based target retrieval. Thus, the presented method is applicable to multiple retrieval methods.

[0129] preface

[0130] Mid-infrared spectroscopy (MID-IR) is a powerful optical technique that probes the vibrational states of individual molecules within tissues and is extremely sensitive to the conformational state of proteins. This extreme sensitivity makes mid-IR spectroscopy ideally suited for microscopy applications, as the presence and even conformational state of endogenous and exogenous materials are revealed through changes in the mid-IR absorption profile of biological specimens. Vibrational spectroscopy is also used in diagnostic applications, for example to distinguish between healthy and cancerous tissue.

[0131] Although the exact mechanism of retrieval is not fully understood, epitope unmasking requires significant changes in the biological composition of the tissue. Therefore, it was sought to determine whether changes in the molecular composition of the biological specimen would be manifested in changes in the mid-IR spectrum. The ultimate goal was to develop a measurement that could be used to accurately determine antigen retrieval of tissue samples using mid-IR spectra. Such a novel capability would enable researchers to assess the retrieval status of tissue samples by standardized and objective measurements.

[0132] Antigen retrieval is central to modern immunohistochemistry because it unmasks protein epitopes so that they can be detected by IHC-based detection. Although the exact mechanism of antigen retrieval remains speculative, it is the key chemical process that allows IHC to be detected on FFPE tissue. Thus, retrieval is fundamentally important to detect current cancer biomarkers, and even the degree to which tissue is retrieved can significantly affect the protein expression levels detected. Despite its importance, the exact mechanism of antigen retrieval remains unclear, and there are no analytical methods to measure the quality of antigen retrieval.

[0133] Methods and Materials

[0134] Acquisition Procedure

[0135] MirrIR microscope slides for reflectance infrared studies (Kevley Technologies, Chesterland, OH) were used for mid-IR spectral measurements. Four-micron serial sections of formalin-fixed paraffin-embedded (FFPE) tonsillar tissue were placed on pretreated MirrIR slides. Deparaffinization of tonsillar tissue was performed manually according to OP2100-025. Briefly, after the xylene step, slides were hydrated through descending grades of ethanol and then transferred to a rapid antigen retrieval (RAR) test bed in VENTANA Cell Conditioning 1 (CC1) solution.

[0136] The antigen retrieval step was performed in CC1 solution in a RAR chamber that was pre-pressurized to 30 psi before turning on the heater. The total heating time for any given experiment included a 90 second ramp-up time and a 2 minute cool-down time. After the antigen retrieval step, the slides were gently washed with deionized water and air-dried at room temperature. The dried slides with intact tonsil tissue were used for mid-IR measurements.

[0137] IHC staining and quantification

[0138] Immunoreactivity data was collected for all samples and processing was analyzed by mid-IR spectroscopy. Briefly, samples were processed using a hybrid procedure with manual deparaffinization and antigen retrieval. Depar was performed using xylene, followed by rehydration through a graded alcohol series. Samples were then placed in CC1 (cat. no. 950-124). After antigen retrieval, samples were transferred in reaction buffer (cat. no. 950-300) to the BenchMark UTLRA instrument for subsequent processing steps from peroxide inhibitor to counterstaining.

[0139] In the study presented here, tonsillar samples were labeled with antisera raised against Ki-67 (30-9) or C4d (SP91). These markers were selected because they show differential responses to increasing antigen retrieval treatments. Ki-67 increases in staining intensity and number of labeled cells, followed by a decrease in intensity and positivity. Conversely, C4d continues to increase in intensity and positivity through antigen retrieval conditions that would otherwise be damaging to the sample. C4d was further selected because it performs poorly when processed with current retrieval methods, but shows clear immunoreactivity when processed with high temperature antigen retrieval (this behavior is described in detail in an addendum to D081973 entitled Improving Staining Quality from Rapid Antigen Retrieval).

[0140] Specimen slides were scanned using a Leica Aperio AT2 (Leica Biosystems, Nussloch, Germany) slide scanner, and the intensity of immunoreactivity and the percentage of stained tissue were quantified using the "Positive Pixels v9" algorithm supplied with the Aperio Imagescope software. For each tissue, a region of interest (ROI) was selected to include tonsillar tissue where staining was expected. As shown in Figure 8, connective tissue was excluded, which showed high background in some staining treatments but was absent in others.

[0141] This quantification method generates intensity units that are reproducible between samples and comparable within experiments. However, no attempt was made to map or match the intensity measurements, or the percentage of positive pixels reported, to pathologist scores.

[0142] Collecting mid-IR data

[0143] Mid-IR spectra were collected with a Fourier transform infrared (FTIR) microscope (Bruker Hyperion 3000, Bruker Optics, Billerica, MA) equipped with an attached optical interferometer (Vertex 70). Serial sections from tonsil blocks were sectioned onto mid-IR reflecting slides (Kevley Technologies, MirrIR) at 4 micrometer thickness, differentially recovered, and imaged with the mid-IR microscope. Tonsil tissue sections recovered under different experimental conditions were placed on the FTIR microscope and the entire tissue section was imaged with a visible objective by raster scanning the field of view (FOV). Bruker software OPUS was used to randomly select areas of tissue where mid-IR spectra were collected using a mercury-cadmium-telluride (MCT) detector. Typically, about 20 to about 80 spectra were collected from each tissue sample. Absorption spectra were collected over a 4 cm -1 A resolution of 100x was used to collect the spectra, with each selected ROI sampled 64 times, and these spectra were averaged together to obtain the final spectrum for a given location. An exemplary tissue image, sampling pattern, and the average spectrum obtained for a single ROI are shown below in Figure 9. All spectra were collected using a 15x IR objective lens, which produces a FOV of approximately 200 μm x 200 μm.

[0144] Preprocessing of mid-IR data

[0145] The collected spectra were preprocessed to remove artifacts, standardize the spectral format, and isolate the mid-IR absorption of the tissue. The microscope directly measures mid-IR transmission. To convert the transmission spectrum to an absorption spectrum, a reference transmission spectrum was collected at a spatial location outside the sample and used to divide the spectrum collected through the tissue. This calculation provides the amount of light attenuated (absorbed and scattered) by the tissue. Atmospheric absorption, primarily from water vapor and carbon dioxide, was then removed using an algorithm in the OPUS software. A baseline correction was then used to correct for tissue scattering using a concave rubber band correction (8 iterations, 64 baseline points). The resulting spectrum represents the absorption by the sample tissue. Finally, all spectra were normalized to a maximum value of 2 to mitigate differences in section thickness and tissue density.

[0146] Experimental design and results

[0147] Variation in antigen retrieval time at a constant antigen retrieval temperature

[0148] In this experiment, antigen retrieval was performed on tonsillar tissue at about 98.6° C. for either about 0, about 10, about 30, about 60, or about 120 minutes. Each treatment was performed on duplicate samples. Mid-IR spectra show a significant shift in the primary protein band, called the amide I band, that loosely correlates with the antigen retrieval treatment. An example of this amide I shift is shown in FIG. 10A. Quantification of the peak wavelength vs. full width at half maximum (FWHM) of the amide I band allows antigen retrieval treatments to be differentiated over time into unrecovered, partial, complete, and over-recovered (FIG. 10B).

[0149] Multiple other metrics were evaluated throughout this project, including principal component analysis, integration of the amide I band, normalization to several bands to correct for scattering, and quantification of the methyl and methylene peaks. Unfortunately, none of these other metrics were able to improve the level of stratification of the antigen retrieval status of the tissues. Finally, a supervised machine learning model was established to take advantage of non-obvious signatures in the mid-IR spectra that are indicative of the antigen retrieval status of the tissues. These small differences in the spectra were identified using the Projection to Latent Structure Regression (PLSR) method. This algorithm takes the mid-IR signal (e.g., absorption spectrum, first derivative, second derivative) and creates a model that is used to determine which features (wavelengths) are most predictive of the response variables (antigen retrieval status, target retrieval status, etc.). The generated model was then evaluated for performance using the same unknown mid-IR data for performance evaluation and optimization. Figure 11 shows how the PLSR model is trained to obtain mid-IR spectra for antigen retrieval signatures. In this experiment, the accuracy of the model was 3 min.

[0150] These studies demonstrate that supervised machine learning models can be used to mine data and develop methods that can be used to determine antigen retrieval times in tonsillar samples. To further validate that the model is recognizing true antigen retrieval signatures, the model was given a set of spectra that the algorithm was not trained on to determine how well it could make blind predictions. As can be seen in Figure 12, the model was able to predict the antigen retrieval time of unknown spectra (purple dots) to within 8 minutes, demonstrating that the model found a robust antigen retrieval signature in the mid-IR spectrum.

[0151] Varying antigen retrieval time and temperature

[0152] In this study, we investigated how well mid-IR spectra combined with a machine learning model could be used to estimate antigen retrieval times and temperatures. Five multi-tissue slides with four separate tonsil tissues were retrieved for 5 minutes at temperatures ranging from about 98.6°C to about 140°C. Mid-IR spectra from three tonsil tissues (Figure 13, circled area containing three tissue specimens) were used to train a PLSR model. This model was then used to infer antigen retrieval conditions in "unknown" tonsil tissue (Figure 13).

[0153] The performance of the model on blind (FIG. 13, circled area containing only a single tissue specimen) tonsillar samples was further tested with the same temperature course run at approximately 30 and 60 minutes of antigen retrieval (see FIG. 14).

[0154] The results from Figure 9 demonstrate that, at least in tonsillar tissue, mid-IR spectroscopy coupled with PLSR can accurately quantitate the extent to which an unknown sample is recovered across all times and temperatures. This is extremely important since time and temperature are the two most important variables affecting antigen recovery.

[0155] Comparison of two extreme antigen retrieval conditions

[0156] In this experiment, two completely different antigen retrieval conditions were compared using mid-IR spectra and IHC staining of C4d protein. The antigen retrieval step in tonsillar tissue was performed in a RAR test stand. The first antigen retrieval treatment was at about 98.6°C for 60 minutes, which mimics the antigen retrieval conditions typically used on the Ventana BenchMark ULTRA platform. The second antigen retrieval condition tested was at 140°C for 5 minutes. In this study, each sample was imaged approximately 128 times in a grid pattern near the center of each tissue, as shown in Figures 17A and 17B.

[0157] Visual inspection of the mid-IR spectra from the two differently treated samples reveals significant differences (Figure 18). For example, the 1060 cm-1 The mid-IR absorption near 1600 to 1700 cm (Figure 18, left) shows a distinct absorption band for the low temperature / long time antigen retrieval sample (dashed trace) that is not present in the high temperature / short time sample (solid trace). This indicates that the tissue nucleic acid is probably in a different conformational state, allowing the low temperature nucleic acid to absorb mid-IR energy. Similarly, the mid-IR absorption near 1600 to 1700 cm -1 The primary protein band located at (Figure 18, center) also has a large difference with a pronounced deformation in the low temperature / long time tissue that is not present in the high temperature / short time antigen retrieval sample. Finally, from about 2800 to about 2950 cm -1 Substantial differences are observed in the methyl and methylene absorption bands between (Figure 18, right). These bands are associated with mid-IR absorption of proteins and lipids. Further studies are required to understand the molecular basis of these different signals. However, this data supports the theory that tissue specimens are in very different conformational states after low and high temperature antigen retrieval.

[0158] Furthermore, these two antigen retrieval protocols were distinguishable using machine learning algorithms on both the training and blinded spectra. The clear difference in the mid-IR spectra is reflected by distinct IHC staining for C4d (Figure 19). Samples treated with antigen retrieval at about 140°C for about 5 minutes show significant C4d immunoreactivity in germinal centers, whereas samples retrieved at about 98.6°C for 1 hour show little immunoreactivity.

[0159] Breast tissue antigen retrieval

[0160] This section expands on the results achieved in tonsil tissue by applying the same method to breast tissue. Four different breast tissue pieces were differentially retrieved at temperatures of about 98.6°C, about 110°C, about 120°C, about 130°C, or about 140°C, all for a fixed time period of 5 minutes. Each tissue was cut in quadruplicate such that four average spectra were obtained for each antigen retrieval temperature.

[0161] As can be seen in Figure 20, visible differences are seen in the mid-IR absorbance spectra as well as the first derivative spectra of differentially recovered tissues. Figures 21A and 21B present a quantitative plot of the first derivative peak position vs. FWHM of the amide I band. There is a clear trend toward narrower bands and lower peak wavenumbers as antigen retrieval progresses. This method did not work for antigen retrieval time, but was able to crudely resolve antigen retrieval temperature. Several other analytical methods were also investigated that were not able to reliably and accurately distinguish samples based on antigen retrieval status, including looking at the magnitudes of individual bands, integrals of bands, ratios of about two or more band intensities, and magnitudes and ratios from second and third derivative spectra.

[0162] To achieve more accurate results needed for a practical embodiment of this technology, a PLSR model was trained and tested on blinded spectra, the results of which are shown in Figure 22. Looking only at the blinded spectra, the trained model was able to determine the antigen retrieval temperature of the tissue to within about 2°C.

[0163] Targeted Recovery in Breast Tissue

[0164] This section presents proof-of-principle data that tissue is in a significantly different configuration after CC1-based antigen retrieval and CC2-based target retrieval. Serial sections of breast tissue were retrieved using either CC1 or CC2. The average absorbance spectra are plotted in Figure 23A and Figure 23B, and clear differences are observed across the spectrum. Figure 24 shows the first derivative spectra in several different bands. Statistical analysis confirms that the spectra are significantly different across bands across the electromagnetic range. This indicates that tissue is in a dramatically different state after CC1 and CC2 retrieval, and that mid-IR spectroscopy is able to detect and resolve these spectral differences. Figure 25 provides a further illustration of the differences between the use of CC1 and CC2.

[0165] Summary and Discussion

[0166] This study establishes proof of concept for the hypothesis that mid-IR spectroscopy can be used to quantify the antigen retrieval process when driven by time or temperature. It also established an independent measurement supported by staining results showing that specific antibodies have differential immune recognition of targets subjected to low and high temperature retrieval conditions. Thus, mid-IR spectroscopy provided evidence that histological tissue sections are not in identical conditions after low and high temperature retrieval. Findings were validated in tonsil and breast tissue.

[0167] Example 2 - Training the unmasking state estimation engine

[0168] In one embodiment, a Principal Component Analysis and Discriminant Analysis (PCDA) algorithm is used to determine the recovery status of the tissue. Mid-IR spectra were collected from approximately 100 locations throughout the tissue sample, so the average spectrum represents the average for the tissue. All spectra were atmospherically corrected to remove CO2 contamination, baseline corrected using concave rubber band correction with 10 iterations and 64 baseline points, and then the acquired spectra were amplitude normalized. Finally, all spectra from each tissue were averaged. A PCDA model was then used to classify the group to which a spectrum belonged. The two main variables in this algorithm were the number of principal components used for classification purposes and the type of discriminant analysis (linear or quadratic). An example is shown in Figure 7A, where the first two principal components are plotted for a series of slide recoveries for different amounts of time. Thus, the discriminant analysis determined the optimal break point and classified each spectrum from one group of recovery conditions. Those skilled in the art will appreciate that as the number of principal components increases, this approach may become a viable option for determining the recovery conditions of a slide based on its mid-IR spectrum. The confusion matrix is ​​shown in Figure 7B, which shows that classification performance can be further improved as more components are utilized during training.

[0169] Other System Components

[0170] The system 200 of the present disclosure can be coupled to a specimen processing device that can perform one or more preparation processes on a tissue specimen. Preparation processes can include, but are not limited to, deparaffinizing the specimen, conditioning the specimen (e.g., cell conditioning), staining the specimen, performing antigen retrieval, performing immunohistochemical staining (including labeling) or other reactions, and / or performing in situ hybridization (e.g., SISH, FISH, etc.) staining (including labeling) or other reactions, as well as other processes to prepare the specimen for microscopy, microanalysis, mass spectrometry, or other analytical methods.

[0171] The processor can apply fixative to the specimen. Fixatives can include crosslinkers (e.g., aldehydes such as formaldehyde, paraformaldehyde, and glutaraldehyde, as well as non-aldehyde crosslinkers), oxidants (e.g., metal ions and complexes such as osmium tetroxide and chromate), protein denaturants (e.g., acetic acid, methanol, ethanol), fixatives of unknown mechanism (e.g., mercuric chloride, acetone, and picric acid), compounding reagents (e.g., Carnoy's fixative, methacarn, Bouin's solution, B5 fixative), Rossmann's liquid, and Gendre's liquid, microwave, and other fixatives (e.g., excluding volume fixation and vapor fixation).

[0172] If the specimen is a paraffin-embedded sample, the specimen can be deparaffinized using an appropriate deparaffinization liquid. After removing the paraffin, any number of substances can be applied to the specimen in succession. The substances can be for pretreatment (e.g., reversing protein crosslinks, exposing nucleic acids, etc.), denaturation, hybridization, washing (e.g., stringent washing), detection (e.g., linking visual or marker molecules to probes), amplification (e.g., amplification of proteins, genes, etc.), counterstaining, cover slipping, etc.

[0173] The specimen processor can apply a wide range of substances to the specimen. Substances include, but are not limited to, stains, probes, reagents, rinses, and / or conditioners. Substances can be fluids (e.g., gases, liquids, or gas / liquid mixtures), etc. Liquids can be solvents (e.g., polar solvents, non-polar solvents, etc.), solutions (e.g., aqueous or other types of solutions), etc. Reagents can include, but are not limited to, stains, wetting agents, antibodies (e.g., monoclonal antibodies, polyclonal antibodies, etc.), antigen retrieval solutions (e.g., aqueous or non-aqueous based antigen retrieval solutions, antigen retrieval buffers, etc.). Probes can be isolated nucleic acids or isolated synthetic oligonucleotides attached to a detectable label or reporter molecule. Labels can include radioisotopes, enzyme substrates, cofactors, ligands, chemiluminescent or fluorescent agents, haptens, and enzymes.

[0174] Staining can be performed using a histochemical staining module or a separate platform (such as an automated IHC / ISH slide stainer). An automated IHC / ISH slide stainer typically includes at least storage containers for the various reagents used in the staining protocol, a reagent dispensing unit in fluid communication with the storage containers for dispensing the reagents onto the slides, a waste removal system for removing used reagents and other waste from the slides, and a control system that coordinates the operation of the reagent dispensing unit and the waste removal system. In addition to performing the staining steps, many automated slide stainers also perform (or are compatible with another system that performs) steps incidental to staining, such as slide baking (to adhere the sample to the slide), degreasing (also called deparaffinization), antigen retrieval, counterstaining, dehydration and clearing, coverslipping, etc. Prichard, Overview of Automated Immunohistochemistry, Arch Pathol Lab Med., Vol. 138, pp. 1578-1582 (2014), which is incorporated herein in its entirety, describes several specific examples of automated IHC / ISH slide stainers and their various features, including intelliPATH (Biocare Medical), WAVE (Celerus Diagnostics), DAKO OMNIS and DAKO AUTOSTAINER LINK 48 (Agilent Technologies), BENCHMARK (Ventana Medical Systems, Inc.), Leica BOND, and Lab Vision Autostainer (Thermo Scientific) automated slide stainers.Additionally, Ventana Medical Systems, Inc. is the assignee of several U.S. patents disclosing systems and methods for performing automated analysis, including U.S. Pat. Nos. 5,650,327, 5,654,200, 6,296,809, 6,352,861, 6,827,901, and 6,943,029, as well as U.S. Patent Application Publication Nos. 20030211630 and 20040052685, each of which is incorporated herein by reference in its entirety.

[0175] Commercially available staining units usually operate on one of the following principles: (1) open individual slide staining, where the slide is placed horizontally and the reagents are dispensed as a puddle on the surface of the slide containing the tissue sample (as implemented e.g. in the DAKO AUTOSTAINER Link 48 (Agilent Technologies) and intelliPATH (Biocare Medical) stainers); (2) liquid overlay technique, where the reagents are covered by or dispensed through an inert fluid layer deposited on the sample (as implemented e.g. in the VENTANA BenchMark and DISCOVERY stainers); (3) capillary gap staining, where the slide surface is placed close to another surface (such as another slide or a cover plate) creating a narrow gap through which capillary forces draw in and keep the liquid reagents in contact with the sample (as the staining principle used in the DAKO TECHMATE, Leica BOND, and DAKO OMNIS stainers). Even after multiple repetitions of capillary gap staining, the fluids in the gap do not mix (e.g., in DAKO TECHMATE and Leica BOND). A variation of capillary gap staining, called dynamic gap staining, utilizes capillary forces to apply the sample to the slide and then translates the parallel surfaces relative to each other during incubation to agitate and mix the reagents (e.g., as implemented in the DAKO OMNIS slide stainer (Agilent)). In translational gap staining, a translatable head is positioned above the slide. The lower surface of the head is spaced from the slide by a first gap that is small enough to allow a meniscus of liquid to form from the liquid on the slide during translation of the slide. A mixing extension having a lateral dimension smaller than the width of the slide extends from the lower surface of the translatable head to define a second gap between the mixing extension and the slide that is smaller than the first gap. During translation of the head, the lateral dimension of the mixing extension is sufficient to cause a lateral movement in the liquid on the slide generally in a direction from the second gap to the first gap, see WO2011-139978.Recently, it has been proposed to use inkjet technology to deposit reagents onto slides, see WO 2016-170008. This list of staining techniques is not intended to be comprehensive, and any fully or semi-automated system for performing biomarker staining can be incorporated into the histochemical staining platform.

[0176] If morphologically stained samples are also desired, an automated H&E staining platform can be used. Automated systems for performing H&E staining typically operate on one of two staining principles: batch staining (also called "dip'n dunk") or individual slide staining. Batch stainers generally use a reagent vat or bath into which many slides are immersed simultaneously. Individual slide stainers, on the other hand, apply reagent directly to each slide, with no two slides sharing the same aliquot of reagent. Examples of commercially available H&E stainers include Roche's VENTANA SYMPHONY (individual slide stainer) and VENTANA HE 600 (individual slide stainer) series H&E stainers; Agilent Technologies' Dako CoverStainer (batch stainer); Leica Biosystems Nussloch GmbH's Leica ST4020 Small Linear Stainer (batch stainer), Leica ST5020 Multistainer (batch stainer), and Leica ST5010 Autostainer XL series (batch stainer) H&E stainers.

[0177] After the specimen is stained, the stained sample can be manually analyzed in a microscope and / or digital images of the stained sample can be taken for archival and / or digital analysis. Digital images can be captured via a scanning platform such as a slide scanner that can scan stained slides at 20x, 40x, or other magnifications to generate high-resolution whole-slide digital images. At a basic level, a typical slide scanner includes at least: (1) a microscope with a lens objective; (2) a light source (such as halogen, light-emitting diode, white light, and / or multispectral light source, depending on the dye); (3) robotics to move the glass slide and / or move the optical element around the slide; (4) one or more digital cameras for image capture; and (5) a computer and associated software to control the robotics and to manipulate, manage, and display the digital slides. Digital data for multiple different XY locations (and sometimes multiple Z planes) on the slide are captured by the camera's charge-coupled device (CCD), and these images are combined to form a composite image of the entire scanned surface. Common methods for achieving this include:

[0178] (1) Tile-based scanning, in which the slide stage or optics is moved in very small increments to capture square image frames that slightly overlap adjacent squares. The captured squares are then automatically matched together to create a composite image; and

[0179] (2) Line-based scanning, in which the slide stage moves in a single axis during acquisition to capture multiple composite image "strips," which can then be matched together to form a larger composite image.

[0180] A detailed overview of the various scanners (both fluorescent and brightfield) can be found in Farahani et al., Whole slide imaging in pathology: advantages, limitations, and emerging perspectives, Pathology and Laboratory Medicine Int'l, Vol. 7, p. 23-33 (June 2015), the contents of which are incorporated by reference in their entirety. Examples of commercially available slide scanners include: 3DHistech PANNORAMIC SCAN II; DigiPath PATHSCOPE; Hamamatsu NANOZOOMER RS, HT, and XR; Huron TISSUESCOPE 4000, 4000XT, and HS; Leica SCANSCOPE AT, AT2, CS, FL, and SCN400; Mikroscan D2; Olympus VS120-SL; Omnyx VL4, and VL120; PerkinElmer LAMINA; Philips ULTRA-FAST SCANNER; Sakura Finetek VISIONTEK; Unic PRECICE 500, and PRECICE 600x; VENTANA ISCAN COREO and ISCAN HT; and Zeiss AXIO SCAN.Z1. Other exemplary systems and features can be found, for example, in International Publication No. WO 2011-049608, or U.S. patent application Ser. No. 61 / 533,114, filed Sep. 9, 2011, entitled IMAGING SYSTEMS, CASSETTES, AND METHODS OF USING THE SAME, the contents of which are incorporated by reference in their entireties.

[0181] In some embodiments, any imaging can be accomplished using any of the systems disclosed in U.S. Patent Nos. 10,317,666 and 10,313,606, the disclosures of which are incorporated herein by reference in their entireties. The imaging device may be a brightfield imager, such as the iScan Coreo™ brightfield scanner or DP 200 scanner sold by Ventana Medical Systems, Inc.

[0182] In some cases, the images can be analyzed in an image analysis system. The image analysis system can include one or more computing devices, such as a desktop computer, a laptop, a tablet, a smartphone, a server, a special-purpose computing device, or any other type of electronic device, capable of performing the techniques and operations described herein. In some embodiments, the image analysis system can be implemented as a single device. In other embodiments, the image analysis system can be implemented as a combination of two or more devices that together implement various functionalities discussed herein. For example, the image analysis system can include one or more server computers and one or more client computers communicatively connected to each other via one or more local area networks and / or wide area networks (such as the Internet). In some embodiments, the image analysis system typically includes at least a memory, a processor, and a display. The memory can include any combination of any type of volatile or non-volatile memory, such as random access memory (RAM), read-only memory such as electrically erasable programmable read-only memory (EEPROM), flash memory, hard drive, solid state drive, optical disk, etc. It is understood that the memory can be included in a single device or distributed across two or more devices. The processor may include one or more processors of any type, such as a central processing unit (CPU), a graphics processing unit (GPU), a dedicated signal or image processor, a field programmable gate array (FPGA), a tensor processing unit (TPU), etc. It is understood that the processor may be included in a single device or distributed across two or more devices. The display may be implemented using any suitable technology, such as LCD, LED, OLED, TFT, plasma, etc. In some implementations, the display may be a touch-sensitive display (touch screen).Image analysis systems also typically include a software system stored on the memory with a set of instructions executable on the processor, the instructions including various image analysis tasks such as object identification, staining intensity quantification, etc. Exemplary commercially available software packages useful for implementing the modules disclosed herein include VENTANA VIRTUOSO; Definiens TISSUE STUDIO, DEVELOPER XD, and IMAGE MINER; and Visopharm BIOTOPIX, ONCOTOPIX, and STEREOTOPIX software packages.

[0183] After the specimen is processed, the user can transport the slide containing the specimen to an imaging device. In some embodiments, the imaging device is a brightfield imager slide scanner. One brightfield imager is the iScan Coreo brightfield scanner sold by Ventana Medical Systems, Inc. In an automated embodiment, the imaging device is a digital pathology device as disclosed in International Patent Application No. PCT / US2010 / 002772 (Patent Publication No. WO / 2011 / 049608), entitled "IMAGING SYSTEMS AND TECHNIQUES," or as disclosed in U.S. Patent Application No. 61 / 533,114, filed September 9, 2011, entitled "IMAGING SYSTEMS, CASSETTES, AND METHODS OF USING THE SAME." International Patent Application No. PCT / US2010 / 002772 and U.S. Patent Application No. 61 / 533,114 are incorporated by reference in their entireties.

[0184] Embodiments of the subject matter and operations described herein can be implemented in digital electronic circuitry, or computer software, firmware, or hardware, including the structures disclosed herein and their structural equivalents, or one or more combinations thereof. Embodiments of the subject matter described herein can be implemented as one or more computer programs, e.g., one or more modules of computer program instructions encoded on a computer storage medium for execution by or to control the operation of a data processing device. Any of the modules described herein can include logic executed by a processor. As used herein, "logic" refers to any information having the form of instruction signals and / or data that can be applied to affect the operation of a processor. Software is an example of logic.

[0185] A computer storage medium can be or can be included in a computer readable storage device, a computer readable storage substrate, a random or serial access memory array or device, or a combination of one or more thereof. Additionally, a computer storage medium is not a propagating signal, but a computer storage medium can be a source or destination of computer program instructions encoded in an artificially generated propagated signal. A computer storage medium can also be or can be included in one or more separate physical components or media (e.g., multiple CDs, disks, or other storage devices). The operations described herein can be implemented as operations performed by a data processing device on data stored in one or more computer readable storage devices or received from other sources.

[0186] The term "programmed processor" encompasses all kinds of devices, apparatus, and machines for processing data, including, for example, a programmable microprocessor, a computer, a system on a chip, or a combination of the above. An apparatus may include special purpose logic circuitry, such as an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit). An apparatus may also include, in addition to hardware, code that creates an execution environment for the computer program, such as code that constitutes a processor firmware, a protocol stack, a database management system, an operating system, a cross-platform runtime environment, a virtual machine, or one or more combinations thereof. The apparatus and execution environment may implement a variety of different computing model infrastructures, such as web services, distributed computing, and grid computing infrastructures.

[0187] A computer program (also called a program, software, software application, script, or code) can be written in any form of programming language, including compiled or interpreted, declarative or procedural, and can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, object, or other unit suitable for use in a computing environment. A computer program can, but need not, correspond to a file in a file system. A program can be stored as part of a file that holds other programs or data (e.g., one or more scripts stored in a markup language document), in a single file dedicated to the program, or in multiple coordinated files (e.g., files that store one or more modules, subprograms, or portions of code). A computer program can be deployed to be executed on one computer, or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communications network.

[0188] The processes and logic flows described herein may be performed by one or more programmable processors executing one or more computer programs to perform actions by manipulating input data to generate output. The processes and logic flows may also be performed by, and apparatus may be implemented as, special purpose logic circuitry, such as an FPGA (field programmable gate array) or an ASIC (application specific integrated circuit).

[0189] Processors suitable for executing a computer program include, by way of example, both general purpose and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor receives instructions and data from a read-only memory or a random access memory, or both. The essential elements of a computer are a processor for performing actions according to the instructions, and one or more memory devices for storing instructions and data. Generally, a computer also includes one or more mass storage devices, such as magnetic disks, magneto-optical disks, or optical disks, for storing data, or is operatively coupled to receive data, transfer data, or both. However, a computer does not require such devices. Furthermore, a computer can be incorporated into another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device (e.g., a universal serial bus (USB) flash drive), to name just a few. Suitable devices for storing computer program instructions and data include, by way of example, all forms of non-volatile memory, media and memory devices, including semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices, magnetic disks, e.g., internal hard disks or removable disks, magneto-optical disks, and CD-ROM and DVD-ROM disks. The processor and memory may be supplemented by, or incorporated in, special purpose logic circuitry.

[0190] To provide for interaction with a user, embodiments of the subject matter described herein can be implemented on a computer having a display device, e.g., an LCD (liquid crystal display), LED (light emitting diode) display, or OLED (organic light emitting diode) display, for displaying information to the user, and a keyboard and pointing device, e.g., a mouse or trackball, for allowing the user to provide input to the computer. In some implementations, a touch screen can be used to display information and receive input from the user. Other types of devices can also be used to provide interaction with the user. For example, feedback provided to the user can be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback. Also, input from the user can be received in any form, including acoustic, speech, or tactile input. Additionally, the computer can interact with the user by sending documents to and receiving documents from the device used by the user, such as, for example, by sending a web page to a web browser on the user's client device in response to a request received from the web browser.

[0191] An embodiment of the subject matter described herein can be implemented in a computing system that includes a back-end component, e.g., as a data server, or includes a middleware component, such as an application server, or includes a front-end component, e.g., a client computer having a graphical user interface or a web browser through which a user can interact with an implementation of the subject matter described herein, or includes any combination of one or more such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communications network. Examples of communications networks include local area networks ("LANs") and wide area networks ("WANs"), inter-networks (e.g., the Internet), and peer-to-peer networks (e.g., ad-hoc peer-to-peer networks). For example, network 20 of FIG. 1 can include one or more local area networks.

[0192] A computing system may include any number of clients and servers. Clients and servers are typically remote from each other and typically interact through a communications network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. In some embodiments, a server sends data (e.g., HTML pages) to a client device (e.g., for the purpose of displaying the data and receiving user input from a user interacting with the client device). Data generated at the client device (e.g., the result of a user's operation) can be received from the client device at the server.

[0193] Examples of biomarkers

[0194] Non-limiting examples of biomarkers that can be unmasked are identified herein. Certain markers are properties of certain cells, while others are identified as being associated with certain diseases or conditions. Examples of known prognostic markers include enzyme markers, such as galactosyltransferase II, neuron-specific enolase, proton ATPase-2, and acid phosphatase. Hormones or hormone receptor markers include human chorionic gonadotropin (HCG), adrenocorticotropic hormone, carcinoembryonic antigen (CEA), prostate-specific antigen (PSA), estrogen receptor, progesterone receptor, androgen receptor, gC1q-R / p33 complement receptor, IL-2 receptor, p75 neurotrophin receptor, PTH receptor, thyroid hormone receptor, and insulin receptor.

[0195] Lymphoid markers include alpha-1-antichymotrypsin, alpha-1-antitrypsin, B cell markers, bcl-2, bcl-6, B lymphocyte antigen 36kD, BM1 (myeloid marker), BM2 (myeloid marker), galectin-3, granzyme B, HLA class I antigens, HLA class II (DP) antigens, HLA class II (DQ) antigens, HLA class II (DR) antigens, human neutrophil defensins, immunoglobulin A, immunoglobulin D, immunoglobulin G, immunoglobulin M, kappa light chains, kappa light chains, lambda light chains, lymphocyte / histiocyte antigens, macrophage markers, muramidase (lysozyme), p80 anaplastic lymphoma kinase, plasma cell markers, secretory leukocyte protease inhibitor, T cell antigen receptor (JOVI 1), T cell antigen receptor (JOVI 3), including termidonucleotidyl transferase, a marker for non-clustered B cells.

[0196] Tumor markers include alpha-fetoprotein, apolipoprotein D, BAG-1 (RAP46 protein), CA19-9 (sialyl Lewis), CA50 (carcinoma-associated mucin antigen), CA125 (ovarian cancer antigen), CA242 (tumor-associated mucin antigen), chromogranin A, clusterin (apolipoprotein J), epithelial membrane antigen, epithelial-associated antigen, epithelial-specific antigen, epidermal growth factor receptor, estrogen receptor (ER), macroscopic cystic disease fluid protein-15, hepatocyte-specific antigen, HER2, heregulin, human gastric mucin, human milk fat globule, MAGE-1, matrix metalloproteinase, melan A, melanoma marker (HMB45), mesothelin, metallothionein, microphthalmic transcription factor (MITF), Muc-1 core glycoprotein, Muc-1 glycotan. Protein, Muc-2 glycoprotein, Muc-5AC glycoprotein, Muc-6 glycoprotein, myeloperoxidase, Myf-3 (rhabdomyosarcoma marker), Myf-4 (rhabdomyosarcoma marker), MyoD1 (rhabdomyosarcoma marker), myoglobin, nm23 protein, placental alkaline phosphatase, prealbumin, progesterone receptor, prostate-specific antigen, prostatic acid phosphatase, prostatic inhibin peptide, PTEN, renal cell carcinoma marker, small intestinal mucous antigen, tetranectin, thyroid transcription factor-1, tissue inhibitor of matrix metalloproteinase 1, tissue inhibitor of matrix metalloproteinase 2, tyrosinase, tyrosinase-related protein-1, villin, von Willebrand factor, CD34, CD34 class II, CD51 Ab-1, CD63, CD69, Chk1, Chk2, Claspin C-met, COX6C, CREB, Cyclin D1, Cytokeratin, Cytokeratin 8, DAPI, Desmin, DHP (1-6 diphenyl-1,3,5-hexatriene), E-cadherin, EEA1, EGFR, EGFRvIII, EMA (epithelial membrane antigen), ER, ERB3, ERCC1, ERK, E-selectin, FAK, Fibronectin, FOXP3, gamma-H2AX, GB3, GFAP, Giantin, GM130, Golgin 97, GRB2, GRP78BiP, GSK3 beta, HER-2, Histone 3, Histone 3_K14-Ace [Anti-acetyl-histone H3 (Lys14)], Histone3_K18-Ace [Histone H3-Acetyl Lys 18), Histone3_K27-TriMe, [Histone H3(Trimethyl K27)], Histone3_K4-diMe [Anti-Dimethyl-Histone H3(Lys 4)], Histone3_K9-Ace [Acetyl-Histone H3(Lys 9)], Histone3_K9-triMe [Histone 3-Trimethyl Lys9], Histone3_S10-Phos [Anti-Phospho Histone H3(Ser 10), mitotic marker], Histone4, HistoneH2A.X-5139-Phos [Phospho Histone H2A.X(Ser139) antibody], HistoneH2B, HistoneH3_DiMethyl K4, HistoneH4_TriMethyl K20-Chip grad, HSP70, urokinase, VEGF R1, ICAM-1, IGF-1, IGF-1R, IGF-1 receptor beta, IGF-II, IGF-IIR, IKB-Alpha IKKE, IL6, IL8, integrin alpha V beta 3, integrin alpha V beta 6, integrin alpha V / CD51, integrin B5, integrin B6, integrin B8, integrin beta 1 (CD 29), integrin beta 3, integrin beta 5 integrin B6, IRS-1, Jagged 1, anti-protein kinase C Beta 2, LAMP-1, light chain Ab-4 (cocktail), lambda light chain, kappa light chain, M6P, Mach 2, MAPKAPK-2, MEK 1, MEK 1 / 2 (Ps222), MEK 2, MEK1 / 2(47E6), MEK1 / 2 blocking peptide, MET / HGFR, MGMT, mitochondrial antigen, Mitotracker Green FM, MMP-2, MMP9, E-cadherin, mTOR, ATPase, N-cadherin, nephrin, NFKB, NFKB p105 / p50, NF-KB P65, Notch1, Notch2, Notch3, OxPhos complex IV, p130Cas, p38 MAPK, p44 / 42 MAPK antibody, P504S, P53, P70, P70S6K, Pan-cadherin, Paxillin, P-cadherin, PDI, pEGFR, phospho-AKT, phospho-CREB, phospho-EGF receptor, phospho-GSK3 beta, phospho-H3, phospho-HSP-70, phospho-MAPKAPK-2, phospho-MEK1 / 2, phospho-p38 MAP kinase, phospho-p44 / 42 MAPK, phospho-p53, phospho-PKC, phospho-S6 ribosomal protein, phospho-Src, phospho-Akt, phospho-Bad, phospho-IKB-a, phospho-mTOR, phospho-NF-kappaB P65, phospho-p38, phospho-p44 / 42 MAPK, phospho-p70 These include S6 kinase, Phospho-Rb, Phospho-Smad2, PIM1, PIM2, PKCβ, podocalyxin, PR, PTEN, R1, Rb 4H1, R-cadherin, ribonucleotide reductase, RRM1, RRM11, SLC7A5, NDRG, HTF9C, HTF9C, CEACAM, p33, S6 ribosomal protein, Src, survivin, synapopodin, syndecan-4, talin, tensin, thymidylate synthase, tuberin, VCAM-1, VEGF, vimentin, agglutinin, YES, ZAP-70, and ZEB.

[0197] Cell cycle-related markers include apoptotic protease effector-1, bcl-w, bcl-x, bromodeoxyuridine, CAK (cdk-acting kinase), cellular apoptosis-susceptibility protein (CAS), caspase 2, caspase 8, CPP32 (caspase-3), cyclin-dependent kinase, cyclin A, cyclin B1, cyclin D1, cyclin D2, cyclin D3, cyclin E, cyclin G, DNA fragmentation factor (N-terminal), and Fa s (CD95), Fas-associated death domain protein, Fas ligand, Fen-1, IPO-38, Mc1-1, minichromosome maintenance protein, mismatch repair protein (MSH2), poly(ADP-ribose) polymerase, proliferating cell nuclear antigen, p16 protein, p27 protein, p34cdc2, p57 protein (Kip2), p105 protein, Stat1α, topoisomerase I, topoisomerase IIα, topoisomerase IIIα, and topoisomerase IIβ.

[0198] Neural tissue and tumor markers include alpha B-crystallin, alpha-internexin, alpha synuclein, amyloid precursor protein, beta amyloid, calbindin, choline acetyltransferase, excitatory amino acid transporter 1, GAP43, glial fibrillary acidic protein, glutamate receptor 2, myelin basic protein, nerve growth factor receptor (gp75), neuroblastoma marker, neurofibrillary 68 kD, neurofibrillary 160 kD, neurofibrillary 200 kD, neuron-specific enolase, nicotinic acetylcholine receptor alpha 4, nicotinic acetylcholine receptor beta 2, peripherin, protein gene product 9, S-100 protein, serotonin, SNAP-25, synapsin I, synaptophysin, tau, tryptophan hydroxylase, tyrosine hydroxylase, and ubiquitin.

[0199] Cluster differentiation markers are CD1a, CD1b, CD1c, CD1d, CD1e, CD2, CD3delta, CD3epsilon, CD3gamma, CD4, CD5, CD6, CD7, CD8alpha, CD8beta, CD9, CD10, CD11a, CD11b, CD11c, CDw12, CD13, CD14, CD15, CD15s, CD16a, CD16b, CDw17, CD18, CD19, CD20, CD21, CD22, CD23, CD24, CD25, CD26, CD27, CD28, CD29, CD30, CD31, CD32, CD33, CD34, CD35, CD36, CD37, CD38, CD39, CD40, CD41, CD42a, CD42b, CD42c, CD42d, CD43, CD44, CD44R, CD45, CD46, CD47, CD48, CD49a, CD49b, CD49c, CD49d, CD49e, CD49f, CD50, CD51, CD52, CD53, CD54, CD55, CD56, CD57, CD58, CD59, CDw60, CD61, CD62E, CD62L, CD62P, CD63, CD64, CD65, CD65s, CD66a, CD66b, CD66c, CD66d, CD66e, CD66f, CD68, CD69, CD70, CD71, CD72, CD73, CD74, CDw75, CDw76, CD77, CD79a, CD79b, CD80, CD81, CD82, CD83, CD84, CD85, CD86, CD87, CD88, CD89, CD90, CD91, CDw92, CDw93, CD94, CD95, CD96, CD97, CD98, CD99, CD100, CD101, CD102, CD103, CD104, CD105, CD106, CD107a, CD107b, CDw108, CD109, CD114, CD115, CD116, CD117, CDw119, CD120a, CD120b, CD121a, CDw121b, CD122, CD123, CD124, CDw125, CD126, CD127, CDw128a, CDw128b, CD130, CDw131, CD132, CD134, CD135, CDw136, CDw137, CD138, CD139, CD140a, CD140b, CD141, CD142, CD143, CD144, CDw145, CD146, CD147, CD148These include CDw149, CDw150, CD151, CD152, CD153, CD154, CD155, CD156, CD157, CD158a, CD158b, CD161, CD162, CD163, CD164, CD165, CD166, and TCR zeta.

[0200] Other cellular markers include centromere protein-F (CENP-F), dianthin, involucrin, lamins A&C [XB 10], LAP-70, mucins, nuclear pore complex proteins, p180 lamellar body protein, ran, r, cathepsin D, Ps2 protein, Her2-neu, P53, S100, epithelial marker antigen (EMA), TdT, MB2, MB3, PCNA and Ki67.

[0201] Suitable demasking agent

[0202] As used herein, the term "unmasking agent" refers to any liquid, including solutions and mixtures, that is dispensed to a specimen to aid in unmasking. In some embodiments, the unmasking agent includes multiple components, such as those listed below. In some embodiments, the unmasking agent is a buffer. In some embodiments, the buffer has a pH in the range of about 5 to about 10. In other embodiments, the buffer has a pH in the range of about 7 to about 9. In other embodiments, the buffer has a pH in the range of about 7.5 to about 11. Non-limiting examples of buffering agents include citric acid, potassium dihydrogen phosphate, boric acid, diethylbarbituric acid, piperazine-N,N'-bis(2-ethanesulfonic acid), dimethylarsinic acid, 2-(N-morpholino)ethanesulfonic acid, tris(hydroxymethyl)methylamine (TRIS), 2-(N-morpholino)ethanesulfonic acid (TAPS), N,N-bis(2-hydroxyethyl)glycine (bicine), N-tris(hydroxymethyl)methylglycine (tricine), 4-2-hydroxyethyl-1-piperazineethanesulfonic acid (HEPES), 2-{[tris(hydroxymethyl)methyl]amino}ethanesulfonic acid (TES), and combinations thereof. In some embodiments, the unmasking agent is water. In other embodiments, the buffer can be comprised of tris(hydroxymethyl)methylamine (TRIS), 2-(N-morpholino)ethanesulfonic acid (TAPS), N,N-bis(2-hydroxyethyl)glycine (bicine), N-tris(hydroxymethyl)methylglycine (tricine), 4-2-hydroxyethyl-1-piperazineethanesulfonic acid (HEPES), 2-{[tris(hydroxymethyl)methyl]amino}ethanesulfonic acid (TES), or combinations thereof.

[0203] In some embodiments, the unmasking agent comprises a TRIS-based buffer having a basic pH. In some embodiments, the TRIS-based buffer has a pH of about 10 (such as at elevated temperature). In yet other embodiments, the buffer can be a citrate sodium phosphate buffer having a pH of about 6.0 at elevated temperature. In other embodiments, the unmasking agent comprises about 0.05% citraconic anhydride. In other embodiments, the unmasking agent comprises about 100 mM TRIS and has a pH of about 8 to about 10. In other embodiments, the unmasking agent comprises about 10 mM citrate, about 2 mM EDTA, and about 0.05% Tween 20 and has a pH of about 6.2. In other embodiments, the unmasking agent comprises about 0.01 M citrate buffer and has a pH of about 6.0.

[0204] In some embodiments, the unmasking agent comprises a component that reacts with the released fixative to prevent it from reacting again with the sample. Examples of such unmasking agents include purging agents or dimedone. Alternatively or additionally, the unmasking agent can comprise a component that reversibly reacts with the free amino groups of proteins, thus protecting them from reacting with any available formaldehyde, such as citraconic anhydride (CCA).

[0205] In other embodiments, the unmasking agent comprises a chelating agent. Examples of chelating agents include, but are not limited to, ethylenediaminetetraacetic acid (EDTA), ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), EGTA / AM (EGTA, tetra(acetoxymethyl ester)), (1,2-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid) (BAPTA), BAPTA / AM (EGTA tetra(acetoxymethyl ester)), 5,5'-dimethyl-BAPTA-AM, 1,2-bis(2-amino-5-methylphenoxy)ethane (MAPTAM), N,N,N',N'-tetrakis(2-pyridylmethyl)ethane-1,2-diamine (TPEN), citrate, or an ionophore such as ionomycin or calcimycin, or any combination or mixture thereof.

[0206] In other embodiments, the unmasking agent comprises an enzyme. Non-limiting examples of enzymes include proteinases (e.g., trypsin, chymotrypsin, pepsin, or proteinase K), nucleases, glycanases, and hyaluronidases.

[0207] In other embodiments, the unmasking agent comprises a chaotropic agent, non-limiting examples of which include butanol, ethanol, guanidinium salts (e.g., guanidinium hydrochloride or guanidinium thiocyanate), lithium perchlorate, lithium acetate, magnesium chloride, phenol, propanol, thiourea, and urea.

[0208] In other embodiments, the unmasking agent comprises a nucleophile, e.g., a chemical species that donates an electron pair to an electrophile. Non-limiting examples of nucleophiles include ammonia, primary amines, secondary amines, and tertiary amines. Other examples of nucleophiles include hydrazines, alcohols, and halides.

[0209] In other embodiments, the unmasking agent comprises a Lewis acid. Non-limiting examples of Lewis acids include metal ions such as iron(III) ions, aluminum ions, magnesium ions, or other electron-deficient compounds such as toluenesulfonic acid, boric acid, boron trifluoride, and tartaric acid.

[0210] In other embodiments, the unmasking agent may include a surfactant. As used herein, "surfactants" are classified as anionic, cationic, or nonionic, depending on their chemical mode of action. Generally, surfactants reduce the interfacial tension between two liquids. An example of a surfactant is sodium dodecyl sulfate. Examples of suitable nonionic surfactants are polyethylene glycol monohexadecyl ether, cetostearyl alcohol, cetyl alcohol, cocamide diethanolamine, cocamide monoethanolamine, decyl glucoside, octylphenoxypolyethoxyethanol, polyethylene glycol monoisohexadecyl ether, lauryl glucoside, nonylphenoxypolyethoxyethanol, 4-nonylphenyl-polyethylene glycol, 1-(4-nonylphenyl)-1,4,7,10,13,16,19,22,25 nonaoxaheptacosane-27- These include ol, nonoxynol, monolaurin, octaethylene glycol monododecyl ether, oleyl alcohol, polyethylene-polypropylene glycol, polyglycerol polyricinoleate, polysorbate, sorbitan monostearate, sorbitan tristearate, stearyl alcohol, polyethylene glycol p-(1,1,3,3-tetramethylbutyl)-phenyl ether, and polyoxyethylene (20) sorbitan monooleate, octyl-, decyl-, dodecyl-glucopyranoside, -maltoside or deoxycholic acid. Exemplary surfactants are sold under the following names: Brij® 35, TWEEN®, Tergitol™, Triton™, Ecosurf™, Dowfax™, Polysorbate 80™, BigCHAP, Deoxy BigCHAP, IGEPAL®, Saponin, Thesit®, Nonidet®, Pluronic F-68, digitonin, deoxycholate, etc. Certain disclosed embodiments relate to using a surfactant selected from Brij® 35, TWEEN®, Tergitol™, Triton™.Additional antigen unmasking agents are disclosed in U.S. Pat. No. 8,486,335, the disclosure of which is incorporated herein by reference in its entirety.

[0211] In some embodiments, the unmasking agent comprises deionized water, TRIS, and a chelating agent. In some embodiments, the unmasking agent comprises deionized water, TRIS, and a chelating agent and has a pH ranging from about 7 to about 9.5. In some embodiments, the unmasking agent comprises deionized water, TRIS, a chelating agent, and a preservative. In some embodiments, the unmasking agent is CC1 available from Ventana Medical Systems, Inc., Tucson, Arizona, USA. Additional unmasking agents are further described herein and in U.S. Patent Application Publication No. 2009 / 0170152, the disclosure of which is incorporated herein by reference in its entirety.

[0212] Additional Embodiments

[0213] Another aspect of the present disclosure is a system for predicting an unmasking state of a test biological specimen processed in an unmasking process, the system comprising: (i) one or more processors; and (ii) one or more memories coupled to the one or more processors that store computer-executable instructions that, when executed by the one or more processors, cause the system to perform operations including: acquiring test spectral data from the test biological specimen, the test spectral data including vibrational spectral data derived from at least a portion of the biological specimen; deriving unmasking features from the acquired test spectral data using a trained unmasking state estimation engine, the unmasking state estimation engine being trained using a training spectral dataset acquired from a plurality of separately unmasked training biological specimens, the training spectral dataset including at least class labels for unmasking duration and unmasking temperature; and predicting an unmasking state of the test biological specimen based on the derived unmasking features.

[0214] In some embodiments, the unmasking state includes at least one of a predicted unmasking duration, a predicted unmasking temperature, and an estimate of tissue quality. In some embodiments, the unmasking state includes both a predicted unmasking duration and a predicted unmasking temperature. In some embodiments, each training spectral data set is derived by: (i) acquiring a training biological specimen; (ii) dividing the acquired training biological specimen into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the unmasking duration. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the unmasking duration. In some embodiments, the training biological specimen includes the same tissue type as the test biological specimen. In some embodiments, the training biological specimen includes a different tissue type than the test biological specimen. In some embodiments, the unmasking process is an antigen retrieval process. In some embodiments, the unmasking process is a target recovery process.

[0215] Another aspect of the present disclosure is a method of predicting an unmasking state of a test biological specimen processed in an unmasking process, the method comprising: acquiring test spectral data from the test biological specimen, the test spectral data including vibrational spectral data derived from at least a portion of the biological specimen; deriving unmasking features from the acquired test spectral data using a trained unmasking state estimation engine; and predicting an unmasking state of the test biological specimen based on the derived unmasking features. In some embodiments, the unmasking state includes one of a predicted unmasking duration or a predicted unmasking temperature. In some embodiments, the unmasking state includes both a predicted unmasking duration and a predicted unmasking temperature. In some embodiments, the unmasking state further includes an estimate of tissue quality. In some embodiments, the unmasking state estimation engine is trained using one or more training spectral datasets, each training spectral dataset including a plurality of training vibrational spectra derived from a plurality of separately unmasked training tissue samples, each training vibrational spectrum including one or more class labels. In some embodiments, the one or more class labels are selected from a known unmasking duration, a known unmasking temperature, and a qualitative assessment of the unmasking condition. In some embodiments, each training spectral data set is derived by: (i) acquiring a training biological specimen; (ii) dividing the acquired training biological specimen into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking.In some embodiments, the test spectrum data includes an average vibration spectrum derived from a plurality of normalized and corrected vibration spectra.

[0216] Another aspect of the present disclosure is a non-transitory computer readable medium storing instructions for predicting an unmasking state of a test biological specimen processed in an unmasking process, the instructions including: acquiring test spectral data from the test biological specimen, the test spectral data including vibrational spectral data derived from at least a portion of the biological specimen; deriving unmasking features from the acquired test spectral data using a trained unmasking state estimation engine; and predicting an unmasking state of the test biological specimen based on the derived unmasking features. In some embodiments, the unmasking state includes one of a predicted unmasking duration or a predicted unmasking temperature. In some embodiments, the unmasking state includes both a predicted unmasking duration and a predicted unmasking temperature. In some embodiments, the unmasking state further includes an estimate of tissue quality. In some embodiments, the unmasking state estimation engine is trained using one or more training spectrum data sets, each training spectrum data set including a plurality of training vibration spectra derived from a plurality of separately unmasked training tissue samples, each training vibration spectrum including one or more class labels. In some embodiments, the one or more class labels are selected from a known unmasking duration, a known unmasking temperature, and a qualitative assessment of the unmasking state. In some embodiments, each training spectrum data set is derived by: (i) acquiring a training biological specimen; (ii) dividing the acquired training biological specimen into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking. In some embodiments, the different unmasking conditions include holding the unmasking temperature constant while varying the duration of unmasking.In some embodiments, the test spectrum data includes an average vibration spectrum derived from a plurality of normalized and corrected vibration spectra.

[0217] All U.S. patents, U.S. patent application publications, U.S. patent applications, foreign patents, foreign patent applications, and non-patent publications referred to herein and / or listed in the Application Data Sheet are incorporated herein by reference in their entirety. Aspects of the embodiments can be modified, if necessary, to use concepts from the various patents, applications, and publications to provide further embodiments.

[0218] Although the present disclosure has been described with reference to certain exemplary embodiments, it should be understood that many other modifications and embodiments can be devised by those skilled in the art that would fall within the spirit and scope of the principles of the present disclosure. More specifically, reasonable variations and modifications are possible in the components and / or arrangements of the combined configuration of the subject matter within the scope of the foregoing disclosure, the drawings, and the appended claims without departing from the spirit of the present disclosure. In addition to modifications and changes in the components and / or arrangements, alternative uses will also be apparent to those skilled in the art.

Claims

1. 1. A system (200) for predicting an unmasking state of a test biological specimen processed in an unmasking process, comprising: (i) one or more processors (209); and (ii) one or more memories (201) coupled to the one or more processors (209), the one or more memories (201) storing computer executable instructions that, when executed by the one or more processors (209), cause the system (200) to: a. acquiring test spectral data from the test biological specimen, the test spectral data comprising vibrational spectral data derived from at least a portion of the biological specimen; b. deriving demasking features from the acquired test spectrum data using a trained demasking state estimation engine (210); c. predicting the unmasking status of the test biological specimen based on the derived unmasking features.

2. The system of claim 1 , wherein the demasking condition comprises one of a predicted demasking duration or a predicted demasking temperature.

3. The system of claim 2 , wherein the demasking conditions include both a predicted demasking duration and a predicted demasking temperature.

4. The system of claim 2 , wherein the unmasking state further comprises an estimate of tissue quality.

5. 2. The system of claim 1, wherein the unmasked state estimation engine is trained using one or more training spectral datasets, each training spectral dataset of the one or more training spectral datasets comprising a plurality of training vibration spectra derived from a plurality of separately unmasked training tissue samples, each training spectral dataset comprising one or more class labels.

6. The system of claim 5 , wherein the one or more class labels are selected from a known unmasking duration, a known unmasking temperature, and a qualitative assessment of the unmasking condition.

7. 7. The system of claim 5 or 6, wherein each training spectral data set is derived by: (i) acquiring a training biological sample; (ii) splitting the acquired training biological sample into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions.

8. The system of claim 7 , wherein the different demasking conditions include holding the demasking temperature constant while varying the duration of the demasking.

9. The system of claim 7 , wherein the different demasking conditions include holding the demasking duration constant while varying the demasking temperature.

10. The system of claim 1 , wherein the acquired test spectrum data comprises an average vibration spectrum derived from a plurality of normalized and corrected vibration spectra.

11. 11. The system of claim 10, wherein the plurality of normalized and corrected vibrational spectra are obtained by: (i) identifying a plurality of spatial regions within the test biological specimen; (ii) obtaining a vibrational spectrum from each individual region of the plurality of identified regions; (iii) correcting the vibrational spectrum obtained from each individual region to provide a corrected vibrational spectrum for each individual region; and (iv) amplitude normalizing the corrected vibrational spectrum from each individual region to a predetermined global maximum to provide an amplitude-normalized vibrational spectrum for each region.

12. 12. The system of claim 11, wherein the vibrational spectrum obtained from each individual region is corrected by: (i) compensating each obtained vibrational spectrum for atmospheric effects to provide an atmospherically corrected vibrational spectrum; and (ii) compensating the atmospherically corrected vibrational spectrum for scattering.

13. The system of claim 1 , wherein the trained unmasking state estimation engine comprises a machine learning algorithm based on dimensionality reduction.

14. The system of claim 13 , wherein the dimensionality reduction comprises projection into a latent structure regression model.

15. The system of claim 13 , wherein the dimensionality reduction comprises principal component analysis and discriminant analysis.

16. The system of claim 1 , wherein the trained unmasking state estimation engine comprises a neural network.

17. 17. The system of claim 1, further comprising an operation for evaluating whether the test biological sample is suitable for labeling with one or more specific binding entities.

18. 18. The system of claim 1, wherein the acquired test spectral data includes vibrational spectral information of at least an amide I band.

19. The test spectral data obtained ranges from about 3200 to about 3400 cm －１ , about 2800 to about 2900 cm －１ , about 1020 to about 1100 cm －１ and / or from about 1520 to about 1580 cm －１ 18. The system of claim 1 , further comprising vibrational spectral information for wavelengths in the range of

20. 1. A non-transitory computer readable medium storing instructions for predicting an unmasking state of a test biological specimen processed in an unmasking process, the instructions comprising: (a) obtaining test spectral data from the test biological specimen, the test spectral data comprising vibrational spectral data derived from at least a portion of the biological specimen; (b) deriving demasking features from the acquired test spectral data using a trained demasking state estimation engine (210), the demasking state estimation engine being trained using a training spectral data set acquired from a plurality of independently demasked training biological specimens, the training spectral data set deriving demasking features including at least class labels for demasking duration and demasking temperature; (c) predicting the unmasking status of the test biological specimen based on the derived unmasking characteristics.

21. 21. The non-transitory computer readable medium of claim 20, wherein the demasking state includes at least one of a predicted demasking duration, a predicted demasking temperature, and an estimate of tissue quality.

22. 21. The non-transitory computer readable medium of claim 20, wherein the demasking condition includes both a predicted demasking duration and a predicted demasking temperature.

23. 23. The non-transitory computer-readable medium of claim 20, wherein each training spectral data set is derived by: (i) acquiring a training biological sample; (ii) splitting the acquired training biological sample into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions.

24. 24. The non-transitory computer readable medium of claim 23, wherein the different demasking conditions include holding a demasking temperature constant while varying a duration of demasking.

25. 24. The non-transitory computer readable medium of claim 23, wherein the different demasking conditions include holding the demasking duration constant while varying the demasking temperature.

26. 26. The non-transitory computer readable medium of any one of claims 20 to 25, wherein the training biological specimens comprise the same tissue type as the test biological specimen.

27. 26. The non-transitory computer readable medium of any one of claims 20 to 25, wherein the training biological specimens comprise a different tissue type than the test biological specimen.

28. 28. The non-transitory computer readable medium of any one of claims 20 to 27, wherein the unmasking process is an antigen retrieval process.

29. 28. The non-transitory computer readable medium of any one of claims 20 to 27, wherein the unmasking process is a target recovery process.

30. 1. A computer-implemented method for predicting an unmasking state of a test biological specimen processed in an unmasking process, comprising: a. acquiring, by a computer, test spectral data from the test biological specimen (320), the test spectral data comprising vibrational spectral data derived from at least a portion of the biological specimen; b. deriving, by the computer, demasking features from the acquired test spectral data using a trained demasking state estimation engine (340), wherein the demasking state estimation engine is trained using a training spectral data set acquired from a plurality of separately demasked training biological specimens, the training spectral data set deriving demasking features including at least a class label for demasking duration and demasking temperature (340); and c. predicting (350), by the computer, at least one of a duration or a temperature of the unmasking process to which the test biological specimen was subjected based on the derived unmasking characteristics.

31. 31. The method of claim 30, wherein the unmasking process is an antigen retrieval process.

32. 31. The method of claim 30, wherein the unmasking process is a target recovery process.

33. 33. The method of any one of claims 30 to 32, wherein the demasking conditions include both a predicted demasking duration and a predicted demasking temperature.

34. 34. The method of any one of claims 30 to 33, further comprising estimating at least one of tissue quality and damage sustained during the unmasking process.

35. 35. The method of any one of claims 30 to 34, wherein the trained unmasking state estimation engine comprises a machine learning algorithm based on projection onto a latent structure regression model.

36. 35. The method of any one of claims 30 to 34, wherein the trained unmasking state estimation engine comprises a machine learning algorithm based on principal component analysis and discriminant analysis.

37. 35. The method of any one of claims 30 to 34, wherein the trained unmasking state estimation engine comprises a neural network.

38. 31. The method of claim 30, wherein each training spectral data set is derived by: (i) acquiring a training biological sample; (ii) splitting the acquired training biological sample into a plurality of training tissue samples; and (iii) unmasking each training tissue sample of the plurality of training tissue samples under different unmasking conditions.

39. 39. The method of claim 38, wherein the different demasking conditions include holding the demasking temperature constant while varying the duration of the demasking.

40. 39. The method of claim 38, wherein the different demasking conditions include holding the demasking duration constant while varying the demasking temperature.

41. 41. The method of any one of claims 30 to 40, wherein the training biological specimen comprises the same tissue type as the test biological specimen.

42. 41. The method of any one of claims 30 to 40, wherein the training biological specimen comprises a different tissue type than the test biological specimen.

43. 43. The method of any one of claims 30 to 42, wherein the test biological specimen is unstained.

44. 43. The method of any one of claims 30 to 42, wherein the test biological specimen is stained for the presence of one or more biomarkers.

45. The test spectrum data is from about 3200 to about 3400 cm －１ , about 2800 to about 2900 cm －１ , about 1020 to about 1100 cm －１ and / or from about 1520 to about 1580 cm －１ 45. The method of any one of claims 30 to 44, comprising vibrational spectral information in the wavelength range of

Images