Methods of detecting cancer
View Patent ↗The disclosure is related to methods of cancer early detection. In some embodiments, a panel of selected protein tumor markers are used for effective and affordable multi-cancer early detection.
1 . A method for early detection of the presence of cancer in a subject, the method comprising:
(a) collecting a blood sample of the subject;
(b) quantifying the level of a panel of biomarkers from the blood sample of the subject using one or more immunoassays, wherein the panel of biomarkers comprises AFP, CA125, CA15-3, CA19-9, CEA, and CYFRA 21-1;
(c) training a machine learning system to generate a classifier using a plurality of parameters for inputs on a cohort population comprising a group of cancer patients and a group of non-cancer individuals using a machine learning algorithm selected from the group consisting of Generalized Linear Model (GLM), Gradient Boosting Machine (GBM), Random Forest (RF), and Support Vector Machine (SVM), wherein the plurality of parameters comprises the level of the panel of biomarkers and at least one clinical parameter;
(d) determining a cancer predicting score of the subject by applying the classifier obtained in step (c) to the quantified level of the panel of biomarkers obtained in step (b) together with the at least one clinical parameter of the subject;
(e) performing an Outlier Analysis by:
(i) determining a cut-off value for each biomarker in the panel of biomarkers in step (b), wherein the cut-off value for a biomarker represents an abnormal high level of the biomarker in non-cancer individuals,
(ii) comparing the quantified level of each biomarker in step (b) with the corresponding cut-off value, wherein:
if any of the quantified level of the panel of biomarkers in step (b) is higher than the corresponding cut-off value, assigning a cancer predicting score to the subject that indicates the subject has cancer; and
if none of the quantified level of the panel of biomarkers in step (b) is higher than the corresponding cut-off value, assigning the cancer predicting score obtained in step (d) to the subject, wherein the cancer predicting score obtained in step (d) is an indicator of whether the subject has cancer
(f) determining that the subject has cancer; and
(g) treating the subject with a cancer therapy, wherein the subject has lung cancer, breast cancer, colorectal cancer, stomach cancer, liver cancer, cervical cancer, esophageal cancer, pancreatic cancer, ovarian cancer, or lymphoma; wherein the cancer therapy comprises surgery or chemotherapy.
2 . The method of claim 1 , wherein the plurality of parameters further comprises X-ray imaging, mammography, computerized tomography (CT), or Magnetic Resonance Imaging (MRI).
3 . The method of claim 1 , wherein the panel of biomarkers are selected from at least seven different biomarkers selected from AFP, CA125, CA15-3, CA19-9, CEA, CYFRA 21-1, ProGRP, SCC, and PSA.
4 . The method of claim 1 , wherein the panel of biomarkers comprises or consists of AFP, CA125, CA15-3, CA19-9, CA72-4, CEA, and CYFRA 21-1.
5 . The method of claim 1 , wherein the subject is a male and the panel of biomarkers comprises or consists of AFP, CA125, CA15-3, CA19-9, CEA, CYFRA 21-1, ProGRP, SCC, and PSA; or wherein the subject is a female and the panel of biomarkers comprises or consists of AFP, CA125, CA15-3, CA19-9, CEA, CYFRA 21-1, ProGRP, and SCC.
6 . The method of claim 1 , wherein the machine learning system is trained using GLM.
7 . The method of claim 1 , wherein the method comprises training the machine learning system using at least two machine learning algorithms selected from the group consisting of GLM, GBM, RF, and SVM.
8 . The method of claim 7 , wherein the method comprises applying GLM to the results from the at least two machine learning algorithms.
9 . The method of any one of claim 1 , wherein the quantified level of the panel of biomarkers is normalized by Modified Z-Score, wherein the Modified Z-Score is obtained by calculating the difference between the observed value and the median value, divided by the median absolute deviation (MAD).
10 . The method of claim 1 , wherein the method can aid early detection of the presence of at least two cancer types simultaneously.
11 . The method of claim 1 , wherein the panel of biomarkers consists of AFP, CA125, CA15-3, CA19-9, CEA, and CYFRA 21-1.
12 . The method of claim 1 , wherein the subject is a male and the panel of biomarkers comprises or consists of AFP, CA125, CA15-3, CA19-9, CEA, CYFRA 21-1, and PSA; or wherein the subject is a female and the panel of biomarkers comprises or consists of AFP, CA125, CA15-3, CA19-9, CEA, and CYFRA 21-1.
13 . The method of claim 1 , wherein the subject is a male and the panel of biomarkers comprises or consists of AFP, CA125, CA15-3, CA19-9, CEA, CYFRA 21-1, SCC, and PSA; or wherein the subject is a female and the panel of biomarkers comprises or consists of AFP, CA125, CA15-3, CA19-9, CEA, CYFRA 21-1, and SCC.
14 . The method of claim 1 , wherein the method further comprises determining the tissue of origin (TOO) in the subject, wherein the subject is indicated to have a cancer based on the machine learning-based supervised algorithm using the panel of biomarkers.
15 . The method of claim 1 , wherein the one or more immunoassays are conducted using an automated immunoassay analyzer.
16 . The method of claim 1 , wherein the one or more immunoassays comprise an enzyme-linked immunosorbent assay (ELISA), a chemiluminescent immunoassay (CLIA), a fluoroimmunoassay (HA), an electrochemiluminescence immunoassay, an enzyme immunoassay (EIA), a radioimmunoassay (RIA), a counting immunoassay (CIA), a filter media enzyme immunoassay (META), a fluorescence-linked immunosorbent assay (FLISA), an agglutination immunoassay, a multiplex fluorescent immunoassay, or an immunohistochemistry assay.
17 . The method of claim 1 , wherein the machine learning system is trained on at least 1000 subjects.