O-001 - Systematic liquid biopsy identifies novel and heterogeneous mechanisms of acquired resistance in gastrointestinal (GI) cancer patients

Abstract

Introduction: Understanding mechanisms of acquired resistance to targeted therapy can guide strategies to improve clinical outcome. Circulating tumor DNA (ctDNA) provides a non-invasive means to identify concurrent heterogeneous resistance mechanisms emerging during therapy. Methods: Since January 2015, 35 patients (pts) with molecularly-defined GI cancers (24 colorectal (CRC), 8 biliary, 3 gastroesophageal (GE)) achieving response or prolonged stable disease on targeted therapies had plasma collected at disease progression for next-generation sequencing of ctDNA. Molecular alterations identified were compared to ctDNA and/or tissue obtained pre-treatment to identify mechanisms of acquired resistance. When possible, post-progression tumor biopsies were also analyzed. Serial ctDNA specimens were evaluated to determine if the change in ctDNA levels can predict response to targeted therapy. Results: In 35 pts, at least one molecular mechanism of resistance was identified in progression ctDNA in 28 pts (80%) with 15 (43%) exhibiting >1 resistance alteration (range 2-12, median 3). Overall, 72 total resistance alterations were identified. 14 pts had matched progression tumor biopsies, and resistance alterations were identified in 9 (61%), all of which were detected in matched ctDNA. In 64% of these pts, additional resistance mechanisms not detected in the matched tumor biopsy were identified in ctDNA. In 7 pts with multiple progression tumor biopsies available, distinct metastases showed different resistance alterations, all of which were detectable in ctDNA, but in 6 (86%) of these pts, ctDNA detected additional resistance alterations not found despite multiple tumor biopsies, reflecting extensive heterogeneity. Several critical and/or novel resistance alterations were identified across tumor types treated with diverse therapies. In 14 of 17 RAS wildtype CRC pts receiving EGFR antibodies, 28 total (13 distinct) resistance alterations affecting KRAS, EGFR extracellular domain, MEK1, MET, and ERBB2 were detected. In 3 of 5 BRAF mutant CRC pts receiving BRAF inhibitor combinations, 10 total (9 distinct) resistance alterations affecting MAPK genes (KRAS, NRAS, BRAF, MEK1, MEK2) were detected. In a MET-amplified GE pt receiving a MET inhibitor, 2 novel secondary mutations in the MET kinase domain were identified. In 5 FGFR2 fusion-positive biliary pts receiving an FGFR inhibitor, all patients developed novel secondary FGFR2 kinase mutations (13 total, 8 distinct), with 2 pts harboring 5 concurrent mutations, and one harboring 2. In serial ctDNA specimens, the decrease in ctDNA levels by 4 weeks correlated with response and therapeutic outcome. Conclusion: Systematic ctDNA analysis at disease progression can effectively identify novel and heterogeneous mechanisms of acquired resistance in pts receiving targeted therapies for a range of molecularly-defined GI cancers. Compared to parallel tumor biopsies, ctDNA more effectively captured the heterogeneity of acquired resistance mechanisms, which may be important to guide subsequent therapy. Real-time monitoring of ctDNA levels may also represent a potential approach to predict response and resistance to therapy. Further studies to optimize integration of ctDNA in clinical decision-making are warranted.