Presentation Detail
Presenter
Josephine Tsang, Research Scientist at ImpriMed, Inc.
Research Abstract
Introduction
The ImpriMed precision oncology platform provides AI-based clinical decision support for anticancer drug selection. The AI currently learns predictive and prognostic features from several different sources including live-cell drug sensitivity testing and molecular profiling.Here we report ongoing research designed to add tumor mutation profiling datato the platform and to discover new predictive and prognostic biomarkers that will improve health outcomes for canine lymphoma patients.
Methods
Previously we performed paired/normal NGS sequencing on 47 canines selected from our biobank of >1,800 tumor samples using targeted enrichment of >300 cancer genes. We expanded the study population by adding 200 additional patient pairs enriched for relapsed patients and patients with T-cell immunophenotype.
Results
On average, we detected 344variants/patient with moderate to high impact on gene function (34 somatic, 310 germline). We detected FBXW mutations known to be mutated inB-cell lymphoma in 16% of B-cell patients, and SATB1 mutations known to be mutated in T-cell lymphoma in 12% of T-cell patients. We then validated the most highly predictive biomarker from our first study, Predictor of SlowRelapse-1 (PSR1), and identified 4 new candidate predictive biomarkers present in 29% of lymphoma patients. In addition, we identified TP53 as a potential mediator of chemotherapy resistance in relapsed patients.
Conclusion
By detecting known mutations, successfully validating PSR1, a novel predictive biomarker, and identifying new putative predictive biomarkers, we have completed critical steps towards integrating the power of tumor mutation profiling into our AI-based precision oncology platform.
