Use of an in-silico clinical trial intelligence solution to predict outcomes of tinengotinib, a potent multi-kinase small molecule FGFR inhibitor in patients with cholangiocarcinoma, based on the molecular matching score.

Abstract

e15190 Background: Cancer genomic diversity, with a median of ~5 pathogenic molecular alterations per tumor, presents a challenge in predicting clinical outcomes of investigational therapies. Precision medicine implies the ability of a drug/drug combination to address the unique molecular profile of an individual patient’s cancer. Methods: We describe an algorithmic solution developed by CureMatch, Inc. that evaluates drugs based on their ability to address a patient’s composite molecular profile from somatic next-generation sequencing (NGS) testing and other omics reports (Matching Score). The scoring system uses expert-curated knowledge base content within a digital reasoning framework, which assesses how well a given therapy option addresses a tumor’s biomarkers, with higher scores implying better molecular matching. Using TransThera’s phase 2 clinical trial results (NCT04919642) of tinengotinib, a potent small molecule multi-kinase inhibitor targeting FGFR, the CureMatch molecular Matching Scores were produced for all 37 patients with advanced cholangiocarcinoma and evaluable outcomes (includes cohorts with (i) FGFR2 fusion and primary progression on previous FGFR inhibitor (FGFRi) (ii) FGFR2 fusion(s) with progression after prior response to FGFRi; (iii) other FGFR alterations; and (iv) FGFR wild type. Evaluable patients had baseline tissue and/or liquid biopsy NGS (mostly performed by Foundation Medicine). Patients were considered to have better outcomes if they achieved stable disease (SD) > 5 months/complete or partial remission (CR/PR); vs. worse outcomes if best response was SD < 5 months/progressive disease (PD); patients with ongoing SD < 5 months were not evaluable. Results: Overall, 25 of 37 evaluable patients (68%) with cholangiocarcinoma who received tinengotinib achieved SD > 5 months/CR/PR; 29 patients (78% of 37) had > 1 FGFR alteration. Matching Score > 22% was found in 22 of 25 patients (88%) who achieved SD > 5 months/CR/PR vs. 4 of 12 patients (33%) with SD < 5 months/PD (p = 0.0004). Notably the Matching Score correctly predicted outcomes in 31 of 37 patients (84%), including in 6 of 8 patients (75%) without FGFR alterations and in 25 of 29 patients (86%) with FGFR alterations. Conclusions: The CureMatch Matching Score significantly correlated with outcomes for tinengotinib, a potent FGFR inhibitor given to patients with cholangiocarcinoma, with activity in both FGFR-altered and FGFR wild-type patients. Additional studies of this algorithmic Matching Score’s ability to predict outcome and therefore optimize treatment for a variety of pharmaceutical agents are warranted. Outcomes CureMatch Scores >22% CureMatch Scores <22% CureMatch Scores <22% P-Value SD>5 months/CR/PR N=23 N=2 Median = 40 (range, 0-99) 0.0004 SD<5 months/PD N=4 N=8 Median = 20 (range, 0-89)