Characterizing polyp progression and its relationship to colorectal cancer in a longitudinal study
Principal Investigator
Name
Scott Lipnick
Degrees
Ph.D.
Institution
Flagship Labs LLC
Position Title
Senior Principal, Preemptive Health and Medicine
Email
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCO-1485
Initial CDAS Request Approval
Feb 26, 2024
Title
Characterizing polyp progression and its relationship to colorectal cancer in a longitudinal study
Summary
Colon polyps are clinically considered as precursors of colorectal cancer, but there is also research indicating that not all polyps have malignant tendencies. To understand what triggers polyp progression is important to provide patients with more personalized care and efficacious treatment. From the existing literature, we know that depending on the histology and morphology of the polyps, some of them can regress / resolve over time while some of them are growing faster than others. We have developed an algorithm for estimating the severity of the polyps using different sources of clinical data. We are interested in validating our methods in a real representative population under CRC screening, and further exploring the association between polyp progression / growth and CRC incidence. PLCO screening trial is assessing the benefits of screening tools used to diagnose prostate, lung, colon, rectum, or ovary cancer. It's a longitudinal study including roughly 155,000 participants in the trial and following up for 10 years. It will provide valuable background population distribution as a representative U.S. population aged 55 years+ who are supposed to start their CRC screening. Many of these participants had endoscopies in records and colon polyps information during these procedures, giving us a chance to profile the characteristics / progression of different polyps and also to explore how this information is associated with future CRC incidence. We are also interested in evaluating the performance of our method in the PLCO dataset to fill in the gap of external validation of the algorithm and to better understand the essential steps needed to calibrate our method.
Aims
1. To profile polyps growth across different histological types of polyps, and compare the patient phenotypes between these polyps
2. To generate classic model performance metrics, such as AUROC, precision, recall, and F1 score to evaluate our algorithm
3. To study the association between different polyps and future colorectal cancer incidence
Collaborators
Scott Lipnick, Etiome, Inc.
William Yuan, Etiome, Inc.
Xiaoyue Zhang, Etiome, Inc.
Shiwei Xu, Etiome, Inc.