Shifting landscape of resistance to next-generation ALK inhibitors with evolving treatment paradigm in ALK+ lung cancer.

Abstract

8607 Background: Next-generation (gen) ALK tyrosine kinase inhibitors (TKIs) are standard first-line (1L) therapy for patients (pts) with ALK -rearranged (ALK+) metastatic non-small cell lung cancer (mNSCLC), having supplanted crizotinib (criz). While past studies uncovered mechanisms of resistance to next-gen ALK TKIs, vast majority of analyzed biopsies (bx) were obtained from pts treated with next-gen TKIs after 1L criz, reflecting the outdated treatment paradigm. Limited knowledge exists on the mechanisms of resistance to second- (2G) and third-gen (3G) ALK TKIs received without 1L criz exposure. Methods: This retrospective study included pts with ALK+ mNSCLC who received 2G ALK TKIs (alectinib, brigatinib, ceritinib, ensartinib) or 3G TKI lorlatinib (lorl) and had post-progression tissue (TBx) or liquid bx (LBx) assessed by next-generation sequencing (NGS). Frequency (freq) of ALK mutations (mut) (on-target) or MET amplification (amp) and histologic transformation (off-target) was compared in pts who had vs had not received prior 1L criz. Results: We identified 270 pts (median age, 52; 61.1% women) who received 2G TKI (1L criz, n=116; no 1L criz, n=106) and/or 3G TKI (1L criz, n=69; no 1L criz, n=59) and underwent TKI-resistant bx (116 pts with ≥2 bx). In total, 436 post-next-gen TKI bx (280 post-2G TKI, 156 post-lorl) underwent NGS. Post-2G TKI bx detected lower freq of ALK mut in pts without vs with prior criz exposure (TBx: 36.8% vs 64.3%, p<0.001; LBx: 44.4% vs 71.7%, p=0.006). Of pts with post-lorl TBx, 43.8% had ≥1 ALK resistance mut detected, of which 23.6% had ≥2 co-occurring ALK mut. Post-lorl Tbx detected lower freq of ALK mut (29.7% vs 53.8%, p=0.024) and lower freq of ≥2 co-occurring ALK mut in pts without vs with prior criz (10.8% vs 32.7%, p=0.036), with consistent findings by LBx. In terms of off-target resistance. MET amp was detected by post-2G TKI TBx at higher freq in pts without vs with prior criz (17.2% vs 2.5%, p=0.002), but with no significant difference post-lorl without vs with prior criz (13.9% vs 5.8%, p=0.26). Of note, the two post-1L lorl TBx both had MET amp or polysomy, without ALK mut. Histologic transformation occurred at similar freq in pts without vs with 1L criz after 2G TKIs (4.3% vs 1.2%, p=0.33) and after lorl (2.7% vs 3.8%, p=0.99). Among pts with post-1L 2G TKI bx, on-target resistance ( ALK mut) was more common with EML4::ALK variant 3a/b vs variant 1 (TBx: 56.3% vs 20.0%, p=0.038; LBx: 75.0% vs 13.3%, p=0.003). Conclusions: In this largest analysis of post-2G/3G ALK TKI TBx/LBx to date, on-target resistance was less freq after 2G/3G TKIs in pts treated with the current paradigm (upfront 2G/3G ALK TKIs) than the past approach (2G/3G TKI after 1L criz). These findings crystallize a shifting resistance landscape and indicate an increasing role for off-target resistance with upfront 2G/3G TKIs, highlighting a need to uncover and therapeutically address off-target resistance.