Technological advancements in colonoscopic and polypectomy procedures have drastically reduced both the incidence and overall mortality due to colorectal cancer (CRC). However, even with the availability of multiple screening tools, only 40% of CRC are diagnosed at early stage, in part due to lack of compliance with follow-up colonoscopy procedure, limited access to colonoscopy. These factors contribute to inappropriate surveillance follow-up as well as development of `interval cancer' within 5-years of a completely negative colonoscopy. Histological characteristic of polypectized polyps (high-risk vs. low-risk) is among major criterion for follow-up surveillance colonoscopy. Given that current surveillance recommendations depend on the histologic type of the polyp, the underdiagnoses of these premalignant precursors often leads to inappropriate follow-up care and therapy, thus contributing to CRC. Our initial screening has identified that a combination of markers i.e. MUC17, MUC5AC and CA19-9 that accurately differentiate benign hyperplastic polyp (HP) from pre-malignant sessile serrated adenoma/polyp (SSA/P) and tubular adenoma (TA). Based on preliminary studies and the identified gaps in diagnosis, we hypothesize that this newly identified marker panel of MUC17, MUC5AC and CA19.9 can accurately classify the major colorectal polyp subtypes, and in conjunction with machine learning tools can provide an economical product for improved patient stratification for better surveillance and prevention of CRC. To meet these milestones, two specific aims are proposed:
aim 1 is designed to evaluate the potential of MUC17, MUC5AC and CA19.9 for effective stratification of benign from malignant polyps. The major milestones for this aim is the development of a MUC17, MUC5AC, and CA19.9 immunostaining kit and to evaluate the potential of the combination for differentiating HP, SSA/P and TA in highly suspicious cases with documented inter-observer variability amongst pathologists. Further, aim 2 focuses on the development of polyp differentiation deep learning computational program based upon histology, tissue markers (MUC17, MUC5AC and CA19.9), polyp size, number, and location (the most critical parameters for deciding the interval of surveillance colonoscopy) for accurate surveillance of CRC by colonoscopy. The major milestone is to develop and evaluate the colon cancer polyp stratification algorithm in an independent patient set. Overall, the present phase I SBIR application seeks to advance the CRC field by simplifying and improving CRC precursor classification, thus improving the ease of polyp classification which can facilitate better recommendations for follow-up surveillance colonoscopy. This should in turn further reduce the healthcare burden by improving patient adherence for CRC management, to reduce deaths from CRC. Altogether, results from Phase 1 will lead to validation in a multi-center trial and validation in a clinical setting (CLIA lab) during phase II to form the basis to seek FDA approval of this test for use in hospital testing to provide accurate classification of premalignant polyps.

Public Health Relevance

With only 40% of patient diagnosed at an early stage, colorectal cancer (CRC) is the second leading cause of cancer related deaths. The present Phase 1 SBIR study aims to harness the diagnostic potential of a combination of biomarkers, clinical characteristic of polyps and CRC as well as machine learning approaches for accurate distinction of malignant polyp/tissues from benign for improving CRC detection and surveillance.

National Institute of Health (NIH)
National Cancer Institute (NCI)
Small Business Innovation Research Grants (SBIR) - Phase I (R43)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Rahbar, Amir M
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Sanguine Diagnostics and Therapeutics
United States
Zip Code