Colorectal cancer detection

Abstract

A method of determining or detecting whether a subject has an adenoma includes performing vibration (e.g., IR) spectroscopic analysis on a blood sample from the subject to produce a spectral feature specific to the blood sample, where the spectral feature of the blood sample is analyzed for representative features from a previous subject having and not having an adenoma to determine whether the subject has an adenoma. And detecting whether the subject suffers from the adenoma based on the spectral characteristics obtained from the subject.

Description

Technical Field

[0001] This disclosure relates to a method for detecting pre-cancerous conditions (e.g., colorectal adenomas) using spectral techniques. Background Technology

[0002] Early detection of cancer is crucial for improving patient outcomes and reducing mortality. In 2020, cancer caused ten million deaths, representing one of the leading causes of death globally. Early diagnosis can inhibit disease progression before tumors multiply to later stages. Therefore, with earlier detection, many patients can be cured solely through surgical intervention, meaning they will not have to endure aggressive systemic treatments such as radiation or chemotherapy—reportedly, 70% of early-stage tumors (stage I) are treated surgically, compared to only about 13% of stage IV cancers. If the disease is not quickly identified and the cancerous tumor is allowed to grow and metastasize to distant secondary sites, the patient's prospects are often bleak, as current treatments are rarely effective for advanced disease.

[0003] Recent research has yielded a wealth of studies on liquid biopsy technologies, demonstrating their potential to transform cancer diagnosis. Many of these tests are based on genomic approaches, utilizing genetic material such as circulating tumor DNA (ctDNA) and / or cell-free DNA (cfDNA). Notably, approximately 150,000 scientific papers document thousands of biomarkers with clear clinical applicability, yet only about 1% of known biomarkers are routinely used in clinical practice. Currently, only a handful of commercially available liquid biopsies are available, primarily for single-cancer detection. For example, the SelectMDx test for prostate cancer, a urine-based test, has been successfully launched by MDxHealth in the US and Europe. The ExoDx test (Exosome Diagnostics), also used for prostate cancer, provides patients with an individual risk score to determine whether they should be referred for a biopsy. The commercial use of these single-cancer liquid biopsies demonstrates the potential of early detection strategies within the healthcare system.

[0004] Colorectal cancer (CRC) is the third most common cancer worldwide, with nearly two million new cases reported in 2020. Early detection of CRC reduces mortality by removing precancerous lesions and treating early-stage cancer. The mean 5-year survival rate after diagnosis has decreased from 91% for early-stage CRC to as low as 15% for stage IV CRC. The National Institute of Health and Care Excellence (NICE) recommends screening for average-risk adults aged 50–75 years with fecal immunochemical testing (FIT). Patients with elevated fecal hemoglobin (f-Hb) in their stool are referred for further investigation via colonoscopy and / or medical imaging. However, FIT is not without its limitations. While highly sensitive for CRC, false positives are often caused by other causes of gastrointestinal or rectal bleeding, such as inflammation or hemorrhoids, resulting in the presence of f-Hb in the stool. Crucially, FIT cannot reliably identify precancerous adenomas, such as advanced adenomas (AA). The sensitivity of the AA test is only 24%, which limits the benefit of using FIT alone as a single screening tool. Furthermore, current CRC screening procedures have extremely poor adherence; only 43% of age-appropriate adults undergo FIT testing. This low screening adherence results in more than half of patients being diagnosed after their disease has progressed, highlighting the need for alternative testing strategies.

[0005] The most advanced technology in the CRC field is probably Cologuard. ® (Exact Sciences, USA) The Cologuard® test is a multi-target stool test. It aims to detect f-Hb in stool and utilizes quantitative molecular assays of KRAS mutations, aberrant NDRG4 and BMP3 methylation, and β-actin, as well as hemoglobin immunoassay. For CRC detection, the sensitivity and specificity are 92% and 87%, respectively. Although Cologuard's sensitivity is higher than FIT, it still only achieves an AA sensitivity of 42%, indicating that its ability to detect precancerous lesions needs improvement.

[0006] Liquid biopsy has great potential to complement stool-based tests, and various blood tests are currently being developed for CRC detection. Because the unpleasant nature of providing samples for stool-based tests can affect screening compliance, simpler blood tests may be a more convenient screening option. Blood tests are reportedly the preferred screening method for CRC; more than two-thirds of patients prefer to provide a blood sample rather than a stool sample. Therefore, liquid biopsy could improve compliance with CRC screening.

[0007] One promising cancer detection method using liquid biopsy is based on spectral features obtained using attenuated total reflectance infrared (IR) spectroscopy to distinguish between cancer and non-cancerous conditions in a subject (see WO2017 / 221027; Baker et al.). However, this method does not address the need in the art to develop methods capable of detecting earlier (particularly precancerous) features to determine the precancerous status of an individual, particularly those associated with colorectal cancer. Therefore, developing a robust test capable of detecting early and precancerous conditions associated with colorectal cancer, such as adenomas, would be transformative in the field of diagnostics.

[0008] Many references disclose methods for detecting cancer using infrared spectroscopy, such as US 2020 / 02000685 (Kapelushnik et al.), US 2017 / 0285030 (Baker et al.), WO 2016 / 097996 (Grasso et al.), WO2022 / 051753 (Khammanivong et al.), and US 2022 / 0308058 (Baker et al.). However, none of these documents disclose methods for detecting precancerous conditions such as adenomas.

[0009] US2006 / 0269972 (Smith et al.) discloses a method for detecting colorectal cancer. However, this method is based on the analysis of stool samples, which is associated with the drawbacks highlighted above.

[0010] One of the purposes of this disclosure is to develop an early detection and diagnostic system that will mitigate one or more of the aforementioned drawbacks of existing screening methods. Summary of the Invention

[0011] This invention is partly based on research using vibrational (e.g., infrared (IR)) spectroscopy to identify precancerous conditions, such as adenomas, using liquid blood samples. This method is unique because it does not search for single biomarkers, but rather probes a broad range of biological characteristics and produces unique features that represent a significant portion or the entire biochemical profile of the sample. These unique spectral features contain molecular information from both the adenoma and the host response (e.g., the immune response). IR spectroscopy methods, such as attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy, are particularly suitable for clinical use because they are minimally invasive, cost-effective, require little or no sample preparation, and produce reproducible results within minutes due to their rapid analysis speed.

[0012] According to the first aspect, a method is provided for determining or detecting whether a subject suspected of having a type of cancer has an adenoma or cancer regardless of stage, said method comprising:

[0013] Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples. These spectral features are then analyzed against representative features from previous subjects who have and do not have cancer and who have and do not have adenomas, in order to detect whether a subject has cancer or adenomas based on the spectral features obtained from the subject.

[0014] Advantageously, this method allows not only the detection of specific types of cancer, but also early and precancerous conditions, such as adenomas associated with that type of cancer. This can facilitate early diagnosis of such precancerous conditions, potentially allowing treatment at an earlier stage than usual, thus reducing the scale of treatment and significantly increasing the survival chances of subjects.

[0015] Furthermore, this method allows users to differentiate between adenomas and cancer in subjects suspected of having either. The spectral characteristics of blood samples can be analyzed against representative features from previous subjects who had and did not have cancer, and who had and did not have adenomas, to identify whether a subject has cancer or an adenoma based on the spectral characteristics obtained from the subject.

[0016] This method allows users to distinguish between different types of adenomas, such as advanced and non-advanced adenomas. Detection of specific molecular vibrational modes and peaks in specific wavenumber regions based on spectral characteristics can also be used to differentiate between different types of adenomas, such as advanced and non-advanced adenomas.

[0017] According to the second aspect, a method is provided for determining or detecting whether a subject suspected of having a type of cancer has an adenoma, the method comprising:

[0018] Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples, wherein the spectral features of the blood samples are analyzed against representative features from previous subjects with and without adenomas in order to detect whether a subject has an adenoma based on the spectral features obtained from the subject.

[0019] Advantageously, this method allows for the detection of precancerous conditions, such as adenomas associated with specific types of cancer (e.g., CRC). This can facilitate early diagnosis of such precancerous conditions, potentially allowing treatment to be initiated at an earlier stage than usual, thus reducing the scale of treatment and significantly increasing the chances of patient survival.

[0020] This method allows users to distinguish between different types of adenomas, such as advanced and non-advanced adenomas. Detection of specific molecular vibrational modes and peaks in specific wavenumber regions based on spectral characteristics can also be used to differentiate between different types of adenomas, such as advanced and non-advanced adenomas.

[0021] According to the third aspect, a method is provided for determining or detecting whether a subject suspected of having a type of cancer has advanced adenoma or cancer regardless of stage, said method comprising:

[0022] Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples. These spectral features are then analyzed against representative features from previous subjects who have and do not have cancer, and who have and do not have advanced adenomas, in order to detect whether a subject has cancer or advanced adenomas based on the spectral features obtained from the subject.

[0023] Advantageously, this method allows for the detection not only of specific types of cancer but also of early and precancerous conditions, such as advanced adenomas associated with that type of cancer. This can aid in the early diagnosis of such precancerous conditions while ruling out non-advanced adenomas to avoid unnecessary colonoscopies or treatments. It can also allow for targeted therapy at potentially earlier stages than usual, thus reducing the scale of treatment and significantly increasing the chances of patient survival.

[0024] This method allows users to differentiate between advanced adenomas and cancer in subjects suspected of having advanced adenomas or cancer. The spectral characteristics of blood samples can be analyzed against representative features from previous subjects who had and did not have cancer, and who had and did not have advanced adenomas, to identify whether a subject has cancer or an advanced adenoma based on the spectral characteristics obtained from the subject.

[0025] Advantageously, the methods disclosed herein can provide a more reliable way to detect early and / or precancerous conditions, such as adenomas associated with specific types of cancer (e.g., CRC). This can facilitate early diagnosis of such precancerous conditions, allowing treatment to potentially be initiated at an earlier stage than usual, thereby reducing the scale of treatment and significantly increasing the chances of subject survival.

[0026] The diagnosis of colorectal cancer typically relies on confirmation via colonoscopy. Therefore, the method disclosed herein may allow for one or more of the following:

[0027] - Providing additional gate-keeping tests improves the reliability of screening tests and reduces the number of patients requiring colonoscopy, thereby lowering healthcare system costs and avoiding unnecessary invasive procedures for patients; and / or

[0028] - Provides an additional test with improved reliability compared to routine screening tests (such as FIT), with positive results identified as high-probability patients and prompt colonoscopy.

[0029] In this document, the term “subject” refers to the individual on whom the test is performed, while the term “previous subject” refers to the individual from whom reference spectral features (e.g., a database or dataset of reference spectral features) are obtained, which can be used for subsequent comparison and / or correlation with the spectral features of the “subject’s” blood sample for the purpose of the test, as explained in more detail below.

[0030] As used herein, the term "blood" refers to whole blood or a portion thereof. In one embodiment, the sample from the subject for vibrational (e.g., IR) spectral analysis is obtained from whole blood or a portion thereof (e.g., serum or plasma). In one embodiment, the sample from the subject for spectral analysis is serum.

[0031] The term "cancer" refers to a physiological condition in which cells exhibit abnormal and uncontrolled growth. Unless otherwise stated, as used herein, "cancer" generally refers to any type of cancer, regardless of its stage. Examples of cancer include, but are not limited to, bile duct cancer, bladder cancer, brain and central nervous system cancer, breast cancer, cervical cancer, colorectal cancer, kidney cancer, gallbladder cancer, leukemia, liver cancer, lung cancer, stomach cancer, Hodgkin's lymphoma and non-Hodgkin's lymphoma, melanoma, multiple myeloma, mesothelioma, osteosarcoma, oral (including lip and salivary gland) cancer, laryngeal and oral, nasal, and hypopharyngeal cancer, ovarian cancer, pancreatic cancer, prostate cancer, sarcoma, thyroid cancer, uterine cancer and vaginal cancer, as well as testicular cancer and penile cancer. Advantageously, in this disclosure, cancer can be selected from any type of cancer that can be associated with precancerous conditions such as adenoma. Therefore, cancer can be selected from a group consisting of various cancers (such as colorectal cancer, stomach cancer, cervical cancer, esophageal cancer, bladder cancer, breast cancer, liver cancer, lung cancer, prostate cancer, and vaginal cancer).

[0032] The term "adenoma" in this article refers to a benign tumor that is generally considered a precancerous condition. Adenomas can comprise several subtypes and can be classified as "advanced adenoma" (AA) or "non-advanced adenoma" (NAA). Advanced adenomas can comprise or may consist of adenomas of the following types: (i) highly dysplastic or villous growth patterns (any size) or (ii) adenomas / serrated lesions >10 mm in size. Non-advanced adenomas can comprise or may consist of adenomas <10 mm in size.

[0033] The type of cancer can be colorectal cancer (CRC) and / or the type of adenoma can be colorectal adenoma.

[0034] Subjects may be suspected of having colorectal cancer.

[0035] Therefore, in an implementation of the first aspect, a method is provided for determining or detecting whether a subject suspected of having colorectal cancer has colorectal adenoma or colorectal cancer regardless of stage, the method comprising:

[0036] Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples. These spectral features are then analyzed against representative features from previous subjects who have and do not have colorectal cancer and colorectal adenomas, in order to detect whether a subject has colorectal adenoma or colorectal cancer based on the spectral features obtained from the subject.

[0037] In the second aspect of the implementation, a method is provided for determining or detecting whether a subject suspected of having colorectal cancer has a colorectal adenoma, the method comprising:

[0038] Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples, wherein the spectral features of the blood samples are analyzed against representative features from previous subjects who have and do not have colorectal adenomas, so as to detect whether a subject has a colorectal adenoma based on the spectral features obtained from the subject.

[0039] In the third aspect of the implementation, a method is provided for determining or detecting whether a subject suspected of having colorectal cancer has advanced colorectal adenoma or colorectal cancer regardless of stage, said method comprising:

[0040] Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples. These spectral features are then analyzed against representative features from previous subjects who have and do not have colorectal cancer and who have and do not have advanced colorectal adenomas, in order to detect whether a subject has colorectal cancer or advanced colorectal adenoma based on the spectral features obtained from the subject.

[0041] Typically, vibrational spectral analysis may include or consist of IR spectral analysis, wherein the IR spectral analysis includes 400-4000 cm⁻¹. -1 The wavelengths between [a certain range]. IR spectral analysis can include or may consist of ATR-IR spectral analysis.

[0042] Previous participants with one type of cancer (such as colorectal cancer) may include participants with different stages of cancer (such as colorectal cancer).

[0043] As used herein, “cancer stage” refers to the size of the tumor and signs of its spread from the tissue of origin, enabling clinicians to assess cancer progression. Based on Overall Stage Grouping, also known as Roman Numeral Staging, most cancers can be classified into four stages. Stage I refers to when the cancer is small and located within the organ where it originated. Stage II refers to when the cancer has grown larger than Stage I but has not yet spread to surrounding tissues. In some types of cancer, Stage II may include cases where the cancer has spread to nearby lymph nodes. Early-stage cancer is typically referred to as Stage I or II cancer. Stage III refers to when the cancer is large and may have spread to surrounding tissues (e.g., nearby tissues and / or lymph nodes). Stage IV refers to when the cancer has spread to other organs or has grown to a considerable size in organ-limited tumors (e.g., brain cancer) (also known as advanced or metastatic cancer). In one implementation, the methods described herein are particularly suitable for the detection of early-stage cancer, typically Stages I–III or I–II. In another implementation, the methods described herein are particularly suitable for detecting cancer regardless of its stage, and / or the methods described herein are particularly suitable for detecting cancer at any stage from I to IV.

[0044] As used herein, the term "detection" is widely used and can be used to facilitate aspects of diagnosis and / or prognosis in subjects. Detection of cancer in a subject may involve analyzing representative features obtained from previous subjects. This may involve comparing and / or correlating the spectral characteristics of a blood sample (or its components) with one or more reference spectral characteristics previously obtained through statistical analysis from previous subjects who had and did not have cancer, and who had and did not have adenomas. The comparison and / or correlation may be performed using suitable computational processes or procedures. The method may include the use of machine learning or deep learning processes for comparison and / or correlation.

[0045] Machine learning or deep learning processes can be or may include neural networks. Neural networks can be or may include convolutional neural networks or recurrent neural networks. Convolutional neural networks may include one or more of the following: at least one one-dimensional convolutional layer, at least one downsampling layer, and / or at least one batch normalization layer. Recurrent neural networks may include one or more Long Short-Term Memory (LSTM) layers. Other exemplary algorithms that may be used include: (skewed) random forests with partial least squares, logistic regression, support vector machines, or other machine learning algorithms at each node in each decision tree; other linear models, such as LDA. Machine learning or deep learning models can be single models or ensembles of multiple models. Ensemble models can be built on different samples of data using one or more of guided aggregation (“bagging”) or random sampling.

[0046] The data may be, may include, may originate from, or may represent directly measurable spectral data, including or generated from spectral measurements obtained from blood samples. The data may be or may include measurements performed over a defined wavelength region, or may be or may include measurements obtained from two or more sub-regions, as will be discussed further herein.

[0047] According to examples of this disclosure, a computer-readable medium is provided carrying a computer program that includes computer-readable instructions configured to cause a computer to perform the methods described herein.

[0048] According to examples of this disclosure, a computer device is provided, such as a computer device embodied within or connected to a spectrometer, the computer device including or configured to access: a memory storing processor-readable instructions, and a processor configured to read and execute the instructions stored in the memory, wherein the processor-readable instructions include instructions configured to control the computer to perform the methods described herein.

[0049] Examples of the subject matter and functional operations described in this specification can be implemented in digital electronic circuits, tangibly embodied computer software or firmware, computer hardware (including the structures disclosed herein and their equivalents), or combinations thereof. Examples of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more computer program instruction modules encoded on a tangible, non-transitory program carrier, for execution by a data processing apparatus or for controlling the operation of a data processing apparatus. Optionally or additionally, the program instructions can be encoded on artificially generated propagation signals, such as machine-generated electrical, optical, or electromagnetic signals, which are generated to encode information for transmission to a suitable receiving device for execution by the data processing apparatus. The computer storage medium can be a machine-readable storage device, a machine-readable storage matrix, a random or serial access memory device, or combinations thereof.

[0050] The term "data processing apparatus" encompasses all kinds of devices, apparatuses, and machines used for processing data, including programmable processors, computers, or multiple processors or computers. The apparatus can include special-purpose logic circuitry, such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits). In addition to hardware, the apparatus can also include code that creates an execution environment for the computer program in question, such as code constituting processor firmware, protocol stacks, database management systems, operating systems, or combinations thereof.

[0051] A computer program (also referred to or described as a program, software, software application, module, software module, script, or code) can be written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and can be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program may, but is not required to, correspond to a file in a file system. A program may 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), as a single file dedicated to the program in question, or as multiple coordinated files (e.g., a file storing one or more modules, subroutines, or portions of code). A computer program can be deployed to execute on a single computer, or on multiple computers located at a site or distributed across multiple sites and interconnected through a communication network.

[0052] The processes and logic flows described in this specification can be executed by one or more programmable computers, which execute one or more computer programs to perform functions by manipulating input data and generating output. The processes and logic flows can also be executed by dedicated logic circuits, and the devices can also be implemented as dedicated logic circuits, such as FPGAs (Field-Programmable Gate Arrays) or ASICs (Application-Specific Integrated Circuits).

[0053] Computers suitable for executing computer programs include, for example, those based on general-purpose or special-purpose microprocessors or both, or any other type of central processing unit. Typically, the central processing unit receives instructions and data from read-only memory or random access memory or both. The basic components of a computer are the central processing unit for executing or running instructions and one or more storage devices for storing instructions and data. Typically, a computer will also include one or more mass storage devices (e.g., disks, magneto-optical disks, or optical disks) for storing data, or operatively coupled to receive data from or transfer data to one or more mass storage devices for storing data, or both. However, a computer does not necessarily have such devices. Furthermore, computers can be embedded in other devices, such as mobile phones, personal digital assistants (PDAs), mobile audio or video players, game consoles, GPS receivers, or portable storage devices (e.g., Universal Serial Bus (USB) flash drives), to name just a few.

[0054] Computer-readable media suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and storage devices, such as semiconductor storage devices like EPROM, EEPROM, and flash memory devices; magnetic disks, such as internal hard disks or removable disks; magneto-optical disks; and CD-ROMs and DVD-ROMs. Processors and memory can be supplemented or incorporated therein by dedicated logic circuitry.

[0055] To provide interaction with the user, examples of the subjects described in this specification can be implemented on a computer having a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. 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; and input from the user can be received in any form, including sound, speech, or tactile input. Furthermore, the computer can interact with the user by sending documents to and receiving documents from the device used by the user; for example, by sending a webpage to a web browser on the user's client device in response to a request received from a web browser.

[0056] The embodiments of the subject matter described in this specification can be implemented in a computing system that includes back-end components (e.g., as a data server), or middleware components (e.g., an application server), or front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with the embodiments of the subject matter described in this specification), or any combination of one or more such back-end components, middleware components, or front-end components. The components of the system can be interconnected via digital data communication (e.g., a communication network) of any form or medium. Examples of communication networks include local area networks (LANs) and wide area networks (WANs), such as the Internet.

[0057] A computing system can include clients and servers. Clients and servers are typically geographically separated and usually interact through communication networks. The client-server relationship arises from computer programs running on their respective computers that have a client-server relationship with each other.

[0058] This article references conditional parameter curves, but it should be understood that such references refer to the evolution of degenerate parameters and may include linear sections and / or arcuate sections, such as sections that can be described by polynomials, exponentials, powers and other functions, and may vary smoothly or gradually or may contain sharp angular transitions between regions.

[0059] Although this specification contains many specific implementation details, these details should not be construed as limiting any invention or the scope of protection that may be claimed, but rather as descriptions of features that may be specific to particular embodiments of a particular invention. Certain features in this specification described in the context of individual embodiments can also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment can also be implemented individually or in any suitable sub-combination in multiple embodiments. Furthermore, although the foregoing features may be described as functioning in certain combinations and even initially claimed in this manner, in some cases, one or more features in the claimed combination can be removed from that combination, and the claimed combination can be a sub-combination or a variation of a sub-combination.

[0060] Typically, vibrational spectral analysis may include or consist of IR spectral analysis, wherein the IR spectral analysis includes 400-4000 cm⁻¹. -1 The wavelengths between [a certain range]. IR spectral analysis may include or may consist of ATR-IR spectral analysis.

[0061] However, it should be understood that other vibrational spectroscopy techniques, such as Raman spectroscopy, can be used.

[0062] The methods disclosed herein typically employ infrared (IR) spectroscopy analysis. Various IR techniques known in the art can be used in the methods disclosed herein. One method utilizes Fourier transform IR (FTIR) spectroscopy analysis. In FTIR, IR spectra can range from 400 to 4000 wavenumbers (cm²). -1 IR spectra are typically collected in the region of 10 cm⁻¹. -1 Or lower, 5cm -1 Or lower, or about 4cm -1 The resolution. FTIR spectral analysis can employ at least 10 scans, at least 15 scans, or at least 30 scans. FTIR spectral analysis can employ up to 100 scans, up to 50 scans, or up to 40 scans. For example, 16 scans can be used. Scans can be added together. As those skilled in the art will understand, the number of scans can be selected to optimize data content and data acquisition time.

[0063] Prior to spectral analysis, a background spectrum can be obtained. This background spectrum can provide correction for the background environment. For example, a background spectrum of air can be obtained to provide atmospheric correction. Optionally or additionally, a background spectrum of a solution, such as an aqueous solution like phosphate-buffered saline (PBS), can be obtained. In one embodiment, a background spectrum is obtained to provide correction for the background environment. In one embodiment, the spectral analysis may further include normalization (e.g., standard normal variables), noise reduction (e.g., based on principal component analysis (PCA), and derivatization (e.g., first or second) as preprocessing steps and / or Fourier transform IR spectral analysis.

[0064] The claimed method uses vibrational spectroscopy analysis, typically IR spectroscopy, on the basis of attenuated total reflectance (ATR)-IR spectroscopy, from a blood sample taken from a subject. In some embodiments, the spectral analysis used to obtain the spectral characteristics specific to the blood sample may be ATR-FTIR.

[0065] In a typical ATR-IR spectral analysis, a blood sample from a subject is loaded onto an internal reflection element (IRE). IR light travels through the IRE and is reflected at least once from the inner surface of the IRE in contact with the sample (e.g., by total internal reflection). This reflection forms an "evanescent wave," the extent to which it penetrates the blood sample depending on the wavelength of the light, the angle of incidence and refractive index of the IRE, and the blood sample itself. The penetration depth and path of the reflected light can be altered by changing the angle of incidence and / or wavelength of the incident light. The beam can be received by an IR detector as it exits the internal reflection element. The IRE is typically an optical material with a higher refractive index than the blood sample to achieve the evanescent wave effect.

[0066] The unique IR spectral features of blood samples (which can be referred to as the characteristic or fingerprint region) can typically range from 400 to 4000 cm⁻¹. -1 The relevant IR spectra between some or all.

[0067] The spectral characteristics specific to a blood sample can be compared with a database of representative characteristics previously obtained from samples from previous subjects who had cancer or adenomas, or from healthy subjects, in order to detect whether a subject has cancer or adenomas by comparing the corresponding characteristics. This comparison can be performed using pattern recognition software and / or machine learning analysis known in the art and / or as described herein.

[0068] The spectral characteristics specific to a blood sample can be compared with a database of representative characteristics from samples previously obtained from subjects with adenomas or healthy subjects, in order to detect whether a subject has an adenoma by comparing the corresponding characteristics. This comparison can be performed using pattern recognition software and / or machine learning analysis known in the art and / or as described herein.

[0069] In one implementation, the analysis of representative features from previous subjects includes applying a trained model to the spectral features. The trained model may include a trained machine learning model and optionally include a neural network, support vector machine (SVM), or random forest (RF) decision tree. The trained model may include or be used as a classifier by applying a probability threshold to the probability values ​​output by the trained model. The method may additionally include selecting and / or varying the probability threshold, thereby selecting and / or varying the specificity and / or sensitivity of the spectral feature analysis. Furthermore, the method may involve selecting the probability threshold based on the receiver operating characteristic (ROC) curve of the trained model, and / or selecting a trained model from a set of trained models based on the ROC of a set of trained models, for example, to obtain the desired specificity and / or sensitivity.

[0070] The term "sensitivity" is understood in this text as the proportion of diseased subjects correctly identified as "positive" by the test. Therefore, sensitivity can be defined as the percentage of true positives predicted by the test. A test with 100% sensitivity will correctly detect all patients with a given disease. In other words, high sensitivity means a low rate of false negatives.

[0071] The term "specificity" is understood in this text as the proportion of non-disease subjects correctly identified as "negative" by the test. Therefore, specificity can be defined as the percentage of true negatives predicted by the test. A test with 100% specificity will correctly detect all patients who do not have a given disease. In other words, high specificity means a low rate of false positives.

[0072] An advantageous feature of the method described herein is its ability to detect (colorectal) adenomas or (colorectal) adenomas and / or (colorectal) cancer in patients. Advantageously, the method is capable of detecting early-stage cancer and / or precancerous conditions, as early medical intervention can increase the likelihood of patient survival. For it to be possible as a first-line diagnostic test, high sensitivity is necessary to ensure that all patients with adenomas (and optionally cancer) are correctly identified. Existing precancerous diagnostic tests (e.g., FIT) are not sensitive enough to provide a reliable method. The spectral liquid biopsy method disclosed herein differs from other tests because the probability threshold of the machine learning analysis can be adjusted according to clinical requirements to maximize sensitivity or specificity. In one embodiment, a method for spectral liquid biopsy with machine learning analysis is provided, wherein the probability threshold of the analysis is adjusted to maximize the sensitivity for the detection of (colorectal) adenomas using a blood sample from a patient.

[0073] In some diagnostic analytics applications, such as when stratifying subjects with adenomas or cancer from healthy subjects in a triage setting, a classifier with high sensitivity and moderate to low specificity may be preferable to a model with low sensitivity and high specificity. In other diagnostic applications, such as for general adenoma / cancer detection with population-level screening tests, a classifier with high specificity and moderate to low sensitivity may be preferred. The desired performance level is typically chosen based on a trade-off between the number of false positives and false negatives that must be tolerated by the specific diagnostic application. This trade-off often depends on the medical consequences of the error. In one implementation, the method described herein includes adjusting the probability threshold of a machine learning algorithm to determine the sensitivity and / or specificity of the spectral analysis.

[0074] High sensitivity or specificity typically refers to a value greater than 70%. Moderate sensitivity or specificity is typically greater than 50% but less than 70%. In one embodiment, the analytical sensitivity is greater than 70% and the analytical specificity is greater than 40%. In another embodiment, the analytical sensitivity is greater than 40% and the analytical specificity is greater than 70%. In one embodiment, the analytical sensitivity is at least 80% and the analytical specificity is at least 40%. In one embodiment, the analytical sensitivity is at least 40% and the analytical specificity is at least 80%. In some embodiments, the sensitivity is at least 80% and the specificity is at least 40%, 45%, 50%, 55%, 60%, 65%, 70%, or 75%. In some embodiments, the specificity is at least 80% and the sensitivity is at least 40%, 45%, 50%, 55%, 60%, 65%, 70%, or 75%.

[0075] In one embodiment, the spectral analysis of this disclosure includes performing a highly sensitive spectral analysis to identify patients with adenoma, or with adenoma and / or cancer, and performing a highly specific spectral analysis to identify patients without adenoma, or without adenoma or cancer.

[0076] This disclosure also provides a method for detecting adenoma, or adenoma and / or cancer, using the spectral characteristics of a blood sample, wherein the characteristics of the blood sample include peak detection at specific molecular vibrational modes and / or at specific wavenumber regions. The detection of specific molecular vibrational modes and peaks at specific wavenumber regions of the spectral characteristics can also be used to distinguish between adenoma and cancer. The detection of specific molecular vibrational modes and peaks at specific wavenumber regions of the spectral characteristics can also be used to distinguish between different types of adenoma, such as advanced adenoma and non-advanced adenoma. Peak detection at specific wavenumber regions is associated with molecular vibrational modes, which can be used to identify the type of disease of interest in a subject, such as adenoma or cancer regardless of stage.

[0077] Analysis of representative features from previous subjects may include applying trained models to spectral features.

[0078] The trained model may include a trained machine learning model, optionally including a neural network, support vector machine (SVM), or random forest (RF) decision tree. Alternatively or additionally, any other suitable model or model feature may be used, such as ORF-PLS, ORF-SVM and / or bagging models, shrinkage discriminant analysis, or distance-weighted discrimination (DWD) linear models.

[0079] By applying a probability threshold to the probability values ​​output by the trained model, the trained model can be included or used as a classifier.

[0080] The method may further include selecting and / or changing the probability threshold, thereby selecting and / or changing the specificity and / or sensitivity of the spectral feature analysis.

[0081] The method may further include selecting a probability threshold based on the receiver operating characteristic (ROC) curve of the trained model, and / or selecting a trained model from the set of trained models based on the ROC of the set of trained models, for example, to obtain the desired specificity and / or sensitivity.

[0082] One method for detecting the presence of adenoma, or adenoma and / or cancer, in a subject using the methods described in this disclosure, and optionally, determining whether a subject has adenoma or cancer if the subject does, is by analyzing molecular vibrational patterns obtained using ATR-IR analysis or similar methods on a blood sample of the subject.

[0083] In one implementation, the method for detecting the presence of adenoma, or adenoma and / or cancer, in a subject includes analyzing molecular vibrational modes selected from: NH (in-plane) bending / deformation, CN stretching, CH stretching / deformation, CH2 stretching, CO stretching, CC stretching, C-OH deformation, CH2 wagging, C=O stretching, and asymmetric... Stretch and / or symmetry Stretch.

[0084] In one embodiment, the method for detecting whether a subject has an adenoma, or an adenoma and / or cancer, is based on the analysis of one or more vibrational modes of spectral characteristics, said vibrational modes being selected from: CO stretching, CC stretching, CH deformation, NH bending, CN stretching, and / or C=O stretching. In another embodiment, the method for detecting whether a subject has an adenoma, or an adenoma and / or cancer, is based on the analysis of vibrational modes of spectral characteristics, said vibrational modes including CO stretching, CC stretching, CH deformation, NH bending, CN stretching, and C=O stretching.

[0085] In one embodiment, the method for detecting colorectal adenoma, or colorectal adenoma and / or colorectal cancer in a subject, is based on the analysis of one or more vibrational modes of spectral characteristics, said vibrational modes being selected from: C=O stretching, CN stretching, NH bending, CO stretching, CC stretching, and / or CH deformation. In another embodiment, the method for detecting colorectal adenoma, or colorectal adenoma and / or colorectal cancer in a subject, is based on the analysis of vibrational modes of spectral characteristics, said vibrational modes including: C=O stretching, CN stretching, NH bending, CO stretching, CC stretching, and CH deformation.

[0086] Alternative spectral features that can be used to detect the presence of adenoma, or adenoma and / or cancer, in a subject are identified by peaks in different wavenumber regions of the spectral features of a sample from the subject.

[0087] In one implementation, the method for detecting whether a subject has an adenoma, or an adenoma and / or cancer, and optionally, if the subject has an adenoma and / or cancer, whether the subject has an adenoma or cancer, is based on a measurement of 400-4000 cm⁻¹. -1The identification of peaks in the spectral characteristics of a sample at one or more wavenumber regions within one or more portions thereof. In an alternative embodiment, the method for detecting whether a subject has an adenoma, or an adenoma and / or cancer, and optionally, if the subject has an adenoma and / or cancer, whether the subject has an adenoma or cancer, is based on peaks in the spectral characteristics of a sample at 1000-3700 cm⁻¹. -1 Identification of peaks in the spectral characteristics of samples located in one or more wavenumber regions within one or more of its parts.

[0088] In one implementation, the method for detecting whether a subject has an adenoma, or an adenoma and / or cancer, is based on one or more wavenumbers: 1163 cm⁻¹. -1 1578cm -1 and / or 1682cm -1 Peak identification at 1163 cm⁻¹. In another embodiment, the method for detecting whether a subject has an adenoma, or an adenoma and / or cancer, is based on the peak identification at 1163 cm⁻¹. -1 1578cm -1 and 1682cm -1 Peak identification at the location.

[0089] In one implementation, the method for detecting whether a subject has colorectal adenoma, or colorectal adenoma and / or colorectal cancer, is based on one or more wavenumbers: 1682 cm⁻¹. -1 1575cm -1 and / or 1165cm -1 Peak identification at 1682 cm⁻¹. In another embodiment, the method for detecting whether a subject has colorectal adenoma, or colorectal adenoma and / or colorectal cancer, is based on peak identification at 1682 cm⁻¹. -1 1575cm -1 and 1165cm -1 Peak identification at the location.

[0090] In one implementation, any of the aforementioned wavenumber regions can vary by ±100 cm. -1 ±50cm -1 ±40cm -1 ±30cm -1 ±20cm -1 and / or ±10cm -1 .

[0091] Based on a predictive model developed using a database of precorrelation analyses “trained” (e.g., through pattern recognition and / or machine learning algorithms), spectral features can be correlated with favorable or unfavorable diagnoses and / or prognoses. To train the model, we provide known cancer samples and known non-cancer samples from appropriate patient populations to identify discriminative features. Correlating the analysis results with favorable or unfavorable diagnoses and / or prognoses can be performed manually (e.g., by a clinician or other suitable analyst) or automatically (e.g., through computational methods). Correlation can be established qualitatively (e.g., through comparison of graphical trajectories or features) or quantitatively (e.g., by referencing predetermined thresholds or statistical limits). The correlation of analysis results can be performed using a predictive model, optionally as defined herein, developed by “training” a database of precorrelation measurements and / or analyses.

[0092] In one implementation, a computer program product may be provided containing executable, computer-readable instructions to perform the methods described herein. A trained machine learning model may be configured to receive input and provide output, said input comprising operations performed on a blood sample from a subject, including those ranging from 400 to 4000 cm... -1 The spectral characteristics of IR spectral analysis at wavelengths between 400-4000 cm⁻¹, the output of which indicates whether the subject has an adenoma, or an adenoma and / or cancer. In one embodiment, a method for training a machine learning model may include: receiving multiple datasets, the multiple datasets representing wavelengths between 400-4000 cm⁻¹ from blood samples from the subject. -1 The process involves: obtaining spectral characteristics of multiple subjects through IR spectral analysis at wavelengths between; receiving data indicating whether at least some of the subjects have adenoma, or adenoma and / or cancer; and training a model based on the patients' spectral characteristics to determine the probability that a patient has adenoma, or adenoma and / or cancer, among some of the subjects. Model training may also include adjusting at least one parameter of the model to optimize the area under the ROC curve and / or provide the desired sensitivity and / or specificity for determining whether a patient has adenoma, or adenoma and / or cancer.

[0093] The detection method disclosed herein provides a way to analyze macromolecules in small volumes of patient serum using IR spectroscopy analysis and machine learning algorithms. Since the analysis requires only a small volume of blood, it may be feasible to integrate this rapid liquid biopsy spectroscopy test with other existing blood-based tests without disrupting clinical practice (e.g., taking small aliquots of blood from routine blood tests). In one embodiment, the aliquot of blood sample used for IR analysis can typically be less than 1 mL. In alternative embodiments, the aliquot of blood sample used for IR analysis can be less than 100 μL, 80 μL, 60 μL, 40 μL, or 10 μL. In this way, a large portion (typically more than 80%, 90%, 95%, or 99%) of the original sample can still be used for subsequent analysis.

[0094] Therefore, by expediting further evaluation of higher-risk patients while excluding diagnoses in other patients (e.g., colonoscopy), the method of this disclosure can allow for efficient patient triage. Furthermore, when provided in conjunction with other blood-based diagnostic assays or routine blood tests, it can identify precancerous conditions, such as adenomas and optionally subtypes of adenomas, in asymptomatic patients or patients with nonspecific symptoms. Optionally, the method can further be effective regardless of the age and / or sex of the subject.

[0095] In one implementation, the method disclosed herein can be performed as a standalone test or optionally as an additional test in combination with other blood-based tests.

[0096] In one embodiment, a single blood sample can be combined with one or more other blood-based tests using the method disclosed herein, wherein a first aliquot is used for IR spectroscopy analysis as described herein, and a second aliquot, comprising the remainder or a portion of the blood sample, is used for one or more other blood-based tests. In one embodiment, the remainder of the blood sample can be used for standard routine blood tests or other liquid biopsy-based diagnostic tests. In alternative embodiments, other analyses may include subsequent testing, such as liquid biopsy sequencing assays (e.g., NGS-based ctDNA assays, methylation assays, biomarker assays). The single blood sample can be aliquoted into individual containers or receivers, or the single blood sample can be provided in a single container or receiver.

[0097] In another aspect, a computer program product is provided that includes executable computer-readable instructions to perform the methods or procedures claimed herein.

[0098] In another aspect, which can be provided independently, a trained model is provided, configured to receive input and provide output, said input comprising a 400-4000 cm... -1 The spectral characteristics of ATR-IR spectral analysis between wavelengths, the output of which indicates whether the subject has an adenoma, or an adenoma and / or cancer.

[0099] In another aspect, which can be provided independently, a method for training a model is provided, the method comprising: receiving multiple datasets, said multiple datasets representing samples from blood samples taken from subjects, ranging from 400 to 4000 cm. -1 The spectral characteristics of multiple subjects were obtained by ATR-IR spectral analysis of wavelengths between; data indicating whether at least some of the subjects had adenoma, or adenoma and / or cancer were received; and a model was trained based on the spectral characteristics of the patients to determine the probability that a patient had adenoma, or adenoma and / or cancer, some of whom had adenoma, or adenoma and / or cancer.

[0100] Training the model may include adjusting at least one parameter of the model to optimize the area under the ROC curve and / or provide the desired sensitivity and / or specificity for determining whether a patient has an adenoma, or an adenoma and / or cancer.

[0101] A feature in any aspect may be provided as a feature in any one or more other aspects. For example, any feature of a method, computer program product, model, or apparatus may be provided as any one or more other features of a method, computer program product, model, or apparatus. Attached Figure Description

[0102] This disclosure will now be further described by way of example and with reference to the accompanying drawings, in which:

[0103] Figure 1 : A schematic breakdown diagram of the complete patient cohort.

[0104] Figure 2 The results show the mean receiver operating characteristic (ROC) curves for (a-pink) colorectal cancer (CRC) versus non-cancerous (NC), (b-purple) adenoma (A) versus NC, and (c-blue) CRC+A versus NC, illustrating the trade-off between sensitivity (Sens) and specificity (Spec). AUC represents the area under the curve.

[0105] Figure 3The detection rate for each colorectal cancer (CRC) stage, with a specificity fixed at 90%, and the sensitivity for overall CRC and advanced adenoma (AA) from the CRC+A vNC model. AA sensitivity does not consider "non-advanced" adenomas (n=7); the overall adenoma sensitivity is 58%. Detailed Implementation

[0106] In this disclosure, numerous terms are referenced, and unless the context otherwise requires, these terms have the meanings provided below.

[0107] The term “comprising” or variations thereof should be understood herein to imply inclusion of the said element, integer or step, or group of elements, integers or steps, but does not exclude any other element, integer or step, or group of elements, integers or steps.

[0108] The term “consisting of” or variations thereof shall be understood to imply inclusion of the said element, integer or step, or group of elements, integers or steps, and exclude any other element, integer or step, or group of elements, integers or steps.

[0109] The term "about" in this document is used to refer to values ​​within ±5% of the specified value when defining numbers or values. For example, if a temperature is specified as about 5°C to about 13°C, then temperatures from 4.75°C to 13.65°C are included.

[0110] method

[0111] Patient sample cohort selection

[0112] Patient samples included in this study were obtained from the Tayside Biorepository (Dundie, UK). Sample collection was ethically approved (#22 / ES / 0041-TR628). Blood samples were collected from patients scheduled for colonoscopy prior to surgical resection and any anticancer therapy. The complete cohort consisted of 296 patients: 100 with CRC, 99 with adenomas (A), and 97 with non-cancerous (NC) diagnoses as colonoscopy screening controls. Figure 1 A schematic breakdown diagram of the complete patient cohort is shown in the figure.

[0113] All cancer samples were collected from patients with a histopathologically confirmed CRC diagnosis. Adenoma samples were classified as advanced adenoma (AA) if they were classified as (i) carcinoma in situ or highly dysplastic or villous growth pattern (any size) or (ii) adenoma / serrated lesion with a size >10 mm. Adenoma samples with a size <10 mm were classified as “non-advanced”.

[0114] The NC group consists of other non-malignant conditions such as small polyps, diverticulosis, and inflammation.

[0115] Blood samples were obtained via venipuncture using red-topped BD Vacutainer™ serum collection tubes and then anonymized. Serum was extracted by centrifugation and stored at -80°C. Non-identifiable clinical and demographic data were obtained in accordance with biobank data control procedures.

[0116] Patient sample analysis

[0117] Remove serum aliquots from frozen storage (-80°C) and thaw at room temperature (18-25°C) for up to 30 minutes, inverting three times to ensure mixing before use. Prepare each patient sample for analysis by pipetting 3 μL of serum onto each of the three wells of a Dxcover® Sample Slide (Dxcover Ltd, UK). Place the prepared slides in a 35°C incubator (Thermo Scientific™ Heratherm™, USA) for at least 10 minutes to produce a dried serum membrane. Then place each dried sample slide into a Dxcover® autosampler (Dxcover Ltd., UK) connected to a PerkinElmer® Spectrum Two™ FTIR spectrometer (PerkinElmer® Inc., USA) and Dxcover® Platform Software (Dxcover Ltd, UK) for automated spectral data acquisition. Collect three spectra for each well, resulting in nine replicates per patient.

[0118] Data Analysis

[0119] Machine learning models were developed to construct diagnostic algorithms based on known patient populations and to predict diseases in unknown samples from the test set. Three different classifications were examined in this study:

[0120] (a) CRC v NC;

[0121] (b) A v NC; and

[0122] (c) CRC+A v NC.

[0123] Nested cross-validation (CV) strategies were used to develop the model to reduce sampling bias. In this approach, patients were randomly split into training and test sets in a 70:30 ratio, repeated 51 times. Model hyperparameters were tuned on the training set (70%), which were used to predict spectra in the test set (30%). Since each patient sample provided nine spectra, the final diagnosis was considered as the consensus prediction (maximum vote) from all nine spectra. Based on the diagnostic algorithm results, the patient sample was reported as positive or negative. For a given resampling, spectra from the individual patient were not allowed to exist in either the training or test sets. The classification metrics obtained from all 51 external CV iterations were summarized, and the average has been reported. For each patient, predictions from all test sets in which the patient resided were collected, and the majority vote was taken as the overall test set prediction for that patient. Thus, the overall detection rate (sensitivity) was calculated as the ratio of correct predictions to the total number of predictions, which was reported for cancer and adenoma, and when split by stage, for CRC patients with a specificity fixed at 90%.

[0124] result

[0125] Figure 2 The figure shows the average receiver operating characteristic (ROC) curve for each category.

[0126] First, the CRC v NC model reported an area under the curve (AUC) of 0.93, indicating excellent discriminative power. The A v NC model had an AUC of 0.85, which is still a promising result considering that adenomas are precancerous lesions and are known to be difficult to detect with currently available screening methods. As expected, the CRC+A v NC ROC curve lies between the other two curves because we grouped cancer and adenoma together. It is conceivable that this could arguably represent the most useful classifier, as an effective screening test should identify both CRC and patients with precancerous lesions for rapid referral to colonoscopy. Therefore, the AUC of 0.88 highlights the potential of this liquid biopsy as a CRC screening tool.

[0127] However, it should be understood that other classifications can also be useful. For example, in other implementations, only certain types of adenomas (e.g., AA) can be grouped with CRC, such as in an AA+CRC vNC analysis represented by the associated AA+CRC vNC curves, while non-advanced adenomas (e.g., size <10 mm) can be grouped with the NC group. Ultimately, since the spectral characteristics of the subgroups are subgroup-specific and can be used by machine learning models and associated diagnostic algorithms, any selection of adenoma subgroups can be considered based on the final diagnostic outcome desired by the user or professional.

[0128] exist Figure 2 In the results shown, the CRC v NC model had a CRC sensitivity of 80% when the specificity was fixed at 90%. This model accurately predicted 83% of stage I cancers, 73% of stage II cancers, 76% of stage III cancers, and 100% of stage IV cancers. For the Av NC model, the sensitivity was 58% for all adenomas.

[0129] Figure 3 The sensitivity of the CRC+A v NC model with 90% specificity is shown, exhibiting results very similar to the CRC v NC model. However, in this case, the sensitivity for CRC is 78%, and 87% of stage IV cancers are detected. When staging is combined, 77% of early (I / II) and 79% of late (III / IV) tumors are successfully identified. Furthermore, 58% of all adenomas are detected using this model, with 7 classified as “non-late” (e.g., size <10 mm). Notably, 54 out of 92 (59%) AA patients are correctly predicted.

[0130] discuss

[0131] The findings from this study demonstrate the significant potential of the method disclosed herein to be used as a screening tool for CRC (including precancerous conditions such as adenomas). The classification metric can be fine-tuned according to the diagnostic pathway and healthcare system requirements. For example, sensitivity (or specificity) can be enhanced while ensuring that specificity (or sensitivity) remains approximate. Many liquid biopsy techniques currently under development for CRC report their maximum sensitivity when specificity is fixed at 90%. This is likely due to the minimum performance level set by the Centers for Medicare & Medicaid Services (CMS) for CRC testing (74% sensitivity / 90% specificity). For the CRC+A v NC model (which is likely the most suitable dataset for the screening setting), the overall CRC sensitivity is 78% (at 90% specificity), exceeding the CMS target. When split by stage, this method successfully detected 83% of stage I tumors and 73% of stage II tumors, demonstrating significant potential for early CRC detection. Additionally, 76% and 87% of stages III and IV were identified, respectively. Furthermore, 59% of AA patients were correctly predicted using the CRC+A v NC model. This is considered the highest sensitivity for AA reported to date from blood-based CRC testing.

[0132] It will be understood that this embodiment is provided merely as an example, and various modifications can be made to this embodiment without departing from the scope of the invention.

Claims

1. A method for determining or detecting whether a subject suspected of having a type of cancer has an adenoma or cancer regardless of stage, the method comprising: Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples. These spectral features are then analyzed against representative features from previous subjects who have and do not have cancer and who have and do not have adenomas, in order to detect whether a subject has cancer or adenomas based on the spectral features obtained from the subject.

2. A method for determining or detecting whether a subject suspected of having a type of cancer has an adenoma, the method comprising: Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples, wherein the spectral features of the blood samples are analyzed against representative features from previous subjects with and without adenomas in order to detect whether a subject has an adenoma based on the spectral features obtained from the subject.

3. A method for determining or detecting whether a subject suspected of having a type of cancer has advanced adenoma or cancer regardless of stage, said method comprising: Vibrational spectral analysis is performed on blood samples from subjects to generate spectral features specific to the blood samples. These spectral features are then analyzed against representative features from previous subjects who have and do not have cancer, and who have and do not have advanced adenomas, in order to detect whether a subject has cancer or advanced adenomas based on the spectral features obtained from the subject.

4. The method according to any one of claims 1 to 3, wherein the vibrational spectral analysis includes wavelengths in the range of 400-4000 cm⁻¹. -1 IR spectral analysis between them.

5. The method of claim 4, wherein the IR spectral analysis includes ATR-IR spectral analysis.

6. The method according to any one of the preceding claims, wherein the type of cancer is colorectal cancer (CRC) and / or wherein the type of adenoma is colorectal adenoma.

7. The method according to any one of the preceding claims, wherein the analysis of representative features from previous subjects comprises applying a trained model to spectral features.

8. The method of claim 7, wherein the trained model comprises a trained machine learning model, optionally comprising a neural network, a support vector machine (SVM), or a random forest (RF) decision tree.

9. The method of claim 7 or claim 8, wherein a probability threshold is applied to the probability value output by a trained model, the trained model comprising or serving as a classifier.

10. The method of claim 9, further comprising: Selecting and / or changing the probability threshold allows for the selection and / or alteration of the specificity and / or sensitivity of the spectral feature analysis.

11. The method according to claim 9 or claim 10, further comprising: Selecting probability thresholds based on the receiver operating characteristic (ROC) curve of a trained model, and / or selecting trained models from a set of trained models based on the ROC of a set of trained models, for example, to obtain the desired specificity and / or sensitivity.

12. The method according to any one of the preceding claims, wherein the spectral analysis comprises the detection of one or more vibrational modes and / or wavenumber regions.

13. The method of claim 11 or claim 12, wherein the method is based on spectral features containing information on one or more molecular vibrational modes.

14. The method according to any one of the preceding claims, wherein the method is based on a temperature of 400-4000 cm. -1 Identification of peaks in the spectral characteristics of samples located in one or more wavenumber regions within a given area.

15. A computer program product comprising executable computer-readable instructions to perform the method according to any one of claims 1 to 14.

16. A trained model configured to receive input and provide output, the input comprising spectral features of a vibrational spectral analysis performed on a blood sample from a subject, the output indicating whether the subject has an adenoma or cancer.

17. A trained model configured to receive input and provide output, the input comprising spectral features of a vibrational spectral analysis performed on a blood sample from a subject, the output indicating whether the subject has an adenoma.

18. A trained model configured to receive input and provide output, the input comprising spectral features of a vibrational spectral analysis performed on a blood sample from a subject, the output indicating whether the subject has an advanced adenoma or cancer.

19. A method for training a model, comprising: Receive multiple datasets, which represent the spectral characteristics of multiple subjects obtained from vibrational spectral analysis of blood samples from the subjects; Receive cancer data and adenoma data indicating whether at least some of the subjects have cancer or adenoma; and A model is trained based on the patient's spectral characteristics to determine the probability that the patient has cancer or adenoma. Some participants had cancer, and some had adenomas.

20. A method for training a model, comprising: Receive multiple datasets, which represent the spectral characteristics of multiple subjects obtained from vibrational spectral analysis of blood samples from the subjects; Receive adenoma data indicating whether at least some of the subjects have adenomas; and A model is trained based on the patient's spectral characteristics to determine the probability that the patient has an adenoma. Some of the participants had adenomas.

21. A method for training a model, comprising: Receive multiple datasets, which represent the spectral characteristics of multiple subjects obtained from vibrational spectral analysis of blood samples from the subjects; Receive cancer data and advanced adenoma data indicating whether at least some of the subjects have cancer or advanced adenoma; and A model is trained based on the patient's spectral characteristics to determine the probability that the patient has cancer or advanced adenoma. Some participants had cancer, and some had advanced adenomas.

22. The method according to any one of claims 19 to 21, wherein training the model includes adjusting at least one parameter of the model to optimize the area under the ROC curve and / or provide the desired sensitivity and / or specificity for determining whether a patient has an adenoma or cancer.

Images