Radiomics-Based Companion Diagnostic for Predicting and Assessing Response to Chemoradiation in Rectal Cancers
Abstract
Rationale and Clinical Impact: This project focuses on the Topic Area of colorectal cancer, which is the third most frequently occurring cancer in the general population as well as the military. Nearly 30% of these tumors occur near or around the rectum, for which the definitive treatment is surgery. To ensure optimal outcomes, it is critical to enable early identification of rectal cancer patients who should receive chemoradiation therapy prior to surgery as well as those patients who should immediately undergo surgery with subsequent chemoradiation therapy. Developing markers to address this unmet need could enable personalization of therapy for over 43,000 rectal cancer patients every year, helping to improve their response, survival, and morbidity rates. This involves addressing two major challenges. First, over 15,000 patients with rectal tumors do not see any benefit from pre-surgical therapy (i.e., no tumor shrinkage) every year, which significantly impacts their overall survival and morbidity. If one could identify patients who will not respond to neoadjuvant therapy at diagnosis, they could be sent straight to surgery and later targeted with adjuvant therapy instead. These patients could avoid complications associated with pre-operative therapy such as bowel obstruction and radiation injury. This subgroup of patients is also more likely to suffer local recurrence of rectal cancer, which has a very poor survival rate (7-14 months). Second, up to 25% of rectal cancer patients will exhibit complete response (i.e., no residual tumor) after pre-operative chemoradiation and excellent survival. If identified accurately and triaged onto a less aggressive treatment pathway, these patients could avoid an unnecessary surgery and associated co-morbidities. These include fecal incontinence, urinary and sexual dysfunction, rectal leakage, or pelvic infection. Scientific Objective & Ultimate Application: There is a dire lack of predictive markers, blood-based tests, or companion diagnostic assays that can be used to identify which rectal cancer patients will benefit from which therapeutic pathway. MR imaging is part of the standard-of-care protocol in rectal cancer but suffers from limited assessment and inter-observer variability. If we had better tools to interrogate rectal magnetic resonance imaging (MRIs), we could noninvasively determine the response of rectal cancers to neoadjuvant therapy and thus, personalize management. Building on a Congressionally Directed Medical Research Programs (CDMRP) Career Development Award, the Principal Investigator (PI) has developed new computer-extracted image measurements to characterize domain-driven properties of tumors (as well as surrounding tissue regions) via routine MRIs. Leveraging these developments, the objective of this proposal is to develop a computerized imaging-based companion diagnostic that can be used to determine which rectal cancer patients most benefit from neoadjuvant therapy and which do not. Our computer-extracted features can correctly identify 80% of patients who will most benefit from standard-of-care pre-operative therapy with less than 10% being mis-called, representing a significant improvement over typical performance requirements for companion diagnostic assays. Upon successful completion, our short-term outcome will be computerized imaging tools that have been retrospectively validated across multiple sites for identifying which rectal cancer patients will benefit from preoperative chemoradiation and who can get surgery with post-operative therapy instead. In the long term, these tools will assist in the adoption of the next wave of treatment paradigms in rectal cancers, which focus on minimally invasive therapies (e.g., wait-and-watch protocols for complete responder patients, experimental therapies for non-responders) with validation on clinical trial datasets (e.g., NCT01515787, NCT02008656). Approach: We will leverage an active collabor
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Dec 05, 2021
- Source ID
- W81XWH2110345
Entities
People
- Satish E Viswanath
Organizations
- Case Western Reserve University
- United States Army