Diverging acquired resistance evolutionary pathways in fusion-positive patients treated with tyrosine kinase inhibitors.

Abstract

Listen

Translate article

3015 Background: Tyrosine kinase fusions are recurrent, actionable oncogenic drivers across cancer types. Multiple tyrosine kinase inhibitors (TKIs) are approved in a disease-specific or tumor-agnostic fashion for these cancers. While TKI resistance includes the acquisition of on-target and off-target alterations, a large subset of resistance remains unexplained and may not simply be mediated by singular emergent alterations. As such, we sought to define how genomic complexity evolves under the selective pressure of TKI therapy. Methods: ALK, RET, NTRK, ROS1 or FGFR fusion-positive patient (pt) tumor samples that underwent clinical FDA-authorized tumor-normal panel sequencing (MSK-IMPACT) between April 2015 and January 2023 were selected. Genomic data assessed included chromosomal instability, inferred by the fraction of genome altered (FGA), fraction of loss-of-heterozygosity (F-LOH) and tumor mutational burden (TMB). Samples with low tumor purity ( < 0.2) or indeterminate copy number profiles were excluded. Treatment data were obtained from natural language processing of electronic medical records. Results: Of 68,920 tumors with sequencing, 49,910 samples met quality requirements. Of these, 1,283 samples (2.6%) from 1,111 pts harbored the queried fusions; among these 64 pts had at least 1 TKI-naïve baseline and 1 or more post-TKI progression of disease (PD) samples, yielding a total of 147 samples usable for evaluating the intra-patient evolution of genomic complexity. The most common cancer types were non-small-cell lung cancer (NSCLC, n = 42), biliary tract cancers (n = 4), and thyroid cancer (n = 4). Paired analysis of TKI-naïve and first post-TKI PD samples in all 64 pts showed increases in FGA (median = 1.1; 0.4~6.6), F-LOH (median = 1.1; 0.2~11.5), and TMB (median = 1.3; 0.2~9.9) upon TKI-resistance (p < 0.01). Pts with primary baseline tumors (n = 22) and pts with metastatic tumors (n-42) were next evaluated independently. Increases in FGA, F-LOH and TMB upon TKI PD were statistically significant in both groups (p < 0.05). Accounting for tumor type, all 3 metrics showed increase in NSCLCs (p < 0.05), whereas only F-LOH was increased in non-lung tumors upon TKI-PD (p < 0.05). Post-PD FGA increase was seen in 22 pts, these didn’t harbor significant increase in TMB (p = 0.07; 57% increasing, 43% not increasing). Similarly, among the 38 pts with post-PD TMB increase, no significant increase in FGA was seen (p = 0.21; 56% increasing, 44% not increasing). F-LOH was increased in both pt groups (p < 0.05). Conclusions: Fusion-positive cancers acquire additional genomic alterations post-TKI, potentially via at least two main diverging evolutionary pathways: the acquisition of additional mutations or an increase in chromosomal instability. Improved understanding of pathways by which fusion-positive cancers acquire genomic complexity may yield novel treatment strategies.