Presentation Detail
Presenter
John Callegari, PhD, Lead Data Scientist, ImpriMed, Inc.
Research Abstract
Introduction
Treatment of canine B-cell lymphoma after relapse from prior therapy is challenging due to evolution of drug resistance. Historical treatment metrics for these patients include a complete remission rate of 27% (CRR), overall response rate of 48% (ORR), complete response duration of 106 days (CRD), and survival after relapse of 110 days (SAR). ImpriMed's Personalized Prediction Profile (PPP) gives a real-time update on a patient's tumor drug sensitivity profile that may be taken into consideration by oncologists when choosing a rescue therapy after relapse occurs.
Methods
We assessed the impact of ImpriMed's AI predictions in a prospective study by analyzing clinical outcomes of 60 relapsed B-cell lymphoma patients after their oncologist received an AI prediction report. Patients were divided into two groups of equal size based on the degree to which drugs administered matched the prediction report provided by ImpriMed. Objective clinical outcomes were measured for both groups.
Results
The high-matching group experienced better outcomes than the low-matching group for every metric analyzed (CRR: 53 high, 4.3 low, P<0.001; CRD: 200 high, 48 low, P=0.1; ORR: 69 high, 41 low, P=0.06;SAR: 270 high, 83 low, P <0.001). Furthermore, the high-matching group experienced better clinical outcomes than historical controls for every metric analyzed (CRR: 53 study, 27 control; CRD: 200 study, 106 control; ORR: 69 study,48 control; SAR: 270 study, 110 control).
Conclusion
These results strongly support the efficacy of combining clinical knowledge with AI decision support to optimize rescue therapy outcomes for relapsed B-cell lymphoma patients.
