We have recently developed the PREMM1,2,6 model, a clinical prediction rule designed to be used by healthcare providers to estimate the probability that an individual carries a mutation in the MLH1, MSH2 and MSH6 mismatch repair (MMR) genes (Balmana et al. JAMA 2006, Kastrinos et. al Gastroenterology 2011). PREMM1,2,6 was developed and subsequently validated in thousands of patients who were undergoing genetic evaluation for Lynch Syndrome, the most common form of hereditary colon cancer. Since the development of the model, two additional genes, PMS2 and EPCAM, have been implicated in the condition. In addition to the PREMM1,2,6 model, we have developed a Polyposis model (Grover et al., JAMA 2012) which predicts the likelihood of an individual carrying a germline mutation in the APC and MYH genes, related to Familial Adenomatous Polyposis and MYH-Associated Polyposis. This model was developed from over 9000 individuals who had undergone genetic testing for these two genes but requires independent, external validation that necessitates a large, multicenter collaborative effort. Finally, although the PREMM1,2,6 model, (available on the Dana-Farber Cancer Institute website at www.dfci.org/premm) is widely used by providers around the world who are familiar with the syndrome, our work and that of others have consistently shown that many patients who are at risk for a familial cancer syndrome are not identified or referred for genetic testing. Systematic approaches of risk assessment are necessary to identify and correctly manage patients with inherited forms of cancer, an issue that will be increasingly important as the number of cancer susceptibility genes to be considered expands and complexity of test interpretation increases. In consideration of these issues, and as an expansion of our prior work, the aims of this R01 renewal application are (1) To expand and validate PREMM1,2,6 to include PMS2 and EPCAM gene mutation prediction using an unpublished cohort of 12,000 patients for model development, and data from an international consortium of family registries for model validation~ (2) To validate the performance of the clinical prediction model for the inherited polyposis syndromes, Familial Adenomatous and MYH-associated Polyposis, in subjects enrolled through an international consortium of collaborators~ and (3) (i) To adapt the Lynch Syndrome and Polyposis models into a single risk assessment tool that can be completed by patients electronically on a mobile device and subsequently used by physicians to generate individualized prediction risk scores of the patient's likelihood of carrying MMR, APC or MYH gene mutations for shared decision making, and (ii) to validate the patient application by comparing its predictions to those made when the model estimates were derived by genetic counselors and healthcare providers.
The results of this effort will achieve several goals of PA-10-025. It will support work in development, application, and evaluation of new and existing cancer risk and prognostic prediction models for use by researchers, clinicians, and the general public. It also addresses several major challenges in model development described in the Program Announcement: integrating diverse types of data, ensuring adequate validation of prediction models, and evaluating their utility in research and clinical settings. The ultimate goal of the application is to develop an extensively validated tool that is easy to use by patients and physicians, and improves identification of patients who are at increased risk for having an inherited colorectal cancer syndrome.
Showing the most recent 10 out of 14 publications
