Jun Deng

Ph.D.

Yale University

Professor

PLCO (Learn more about this study)

PLCO-392

Aug 10, 2018

Stratifying Cancer Risks for Individuals Based on Deep Learning of PLCO Data

Cancer is a worldwide public health issue with an estimated 21.7 million new cases and 13 million cancer deaths globally by 2030. Despite a tremendous amount of money and resources have been spent on cancer care over the past 50 years, there are estimated 1,735,350 new cancer cases and 609,640 cancer deaths in the United States alone in 2018. Besides the disparities in resource, environment, diet, and lifestyle, one of the major reasons for high cancer mortality rates is the failure to diagnose cancers at early stages, missing perhaps the best window of opportunity for intervention and cure. Recently artificial intelligence has shown great promise in leveraging big health data to diagnose and stage diseases, reduce cost, improve healthcare and patient outcome. While access to big data stored in silo-like electronic medical record systems has been very restrictive, the large amount of health data collected in some nation-wide multi-centered studies, such as the National Health Interview Survey (NHIS) and the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO), is readily available. Recently, we have demonstrated that deep neural networks can be used for cancer prediction with acceptable sensitivity and specificity based only on personal health data from NHIS, without family history, cancer staging or follow-up information. The advantages of this approach lie in its predicting power, cost-effectiveness, non-invasiveness, real-time turnaround, and easy-to-obtain health data at individual level. Built upon our preliminary study, we hypothesize that the large amount of data collected in the PLCO trial, including not only personal health data but also the staging information of 18 types of cancer (i.e., Prostate, Lung, Colorectal, Ovarian, Biliary, Bladder, Breast, Endometrial, Glioma, Head & Neck, Hematopoietic, Liver, Male Breast, Melanoma, Pancreas, Renal, Thyroid, and Upper GI), long-term follow-up, family history, socio-behavior, lifestyle, dietary data, serum vitamin D, and cancer-specific screening results such as colonoscopy, PSA, and ovarian pathology, can be used to train and validate a deep learning algorithm to stratify cancer risks for individuals. Hence, the goal of this project is to develop a risk predictor and classifier based on deep learning of PLCO data for better risk stratification and more precise intervention, hence improving outcomes for high-risk people while minimizing overtreatment of low-risk disease. If implemented successfully, we envision a risk predictor and classifier tool embedded in the clinical workflow for effective risk stratification and prevention for millions of people worldwide, hence reducing cancer mortality in the long run. With the daunting cost of healthcare and nontrivial rate of cancer mortality, it is critical to stratify an individual’s cancer risks prior to its onset to maximize outcomes for people with aggressive disease and minimize unnecessary treatment of indolent disease.