1375P - Sequential ctDNA profiling in patients with advanced non-small cell lung cancer: An interim analysis of the COPE randomized study

Background

The full impact of serial ctDNA profiling on predicting outcome and guiding therapeutic decisions in patients (pts) with advanced non-small cell lung cancer (NSCLC) has not been explored in a randomized setting.

Methods

COPE, a multicentric phase 2 study, enrolls advanced NSCLC pts at diagnosis. Pts enter 2 arms: Arm A gets both tissue and ctDNA profiling initially and at each imaging, while Arm B only gets tissue profiling initially. Analyses use FDA-approved FoundationOne®Liquid CDx and FoundationOne®CDx assays, reviewed weekly by a national tumor board (MTB). Primary endpoint: 18-month survival. Secondary endpoints include comparing tissue/ctDNA results, proportion of actionable alterations, treatment outcomes, and identifying resistance mechanisms via ctDNA. This interim report focuses on secondary objectives.

Results

150 pts were enrolled in the study. 23 (20.5%) had oncogene-addicted disease. In Arm A (n=112), the median number of liquid biopsies (LB) per pts was 8, (range,1-19). Initial LB (LB1) pts revealed ESCAT 1 actionable molecular alterations in 24 pts (23%). Baseline high tumor fraction (>10%) correlated with worse response rates (RR) (45 vs 66%, p=0.046), progression-free (PFS, p=0.032) and overall survival (OS, p=0.17). Early decrease in average variant allele frequency (≥90% vs <90%,3 weeks after treatment onset) was associated with better RR (79 vs 49%, p=0.005), PFS (mPFS: 10.4 vs 5.6 months, p=0.026) and OS (mOS: NR vs 15.2 months, p=0.002). Among pts with non-oncogene addicted disease and treated with 1st line immune checkpoint therapy based regimen, the most frequent emerging alterations occurring at secondary resistance were KRAS, MAP2K1, NBN, RB1, SMARCA4, STAG2 and STK11 mutations. MTB recommendations were influenced by the results of LB1 in 26.6% of the pts (n=17/64), while findings from subsequent LB informed MTB decisions in 46.9% (n=30/64) of pts, 28.1% (n=18/64) with no previous LB1 MTB decisions.

Conclusions

This study highlights serial ctDNA profiling's benefits for managing advanced NSCLC, providing real-time genetic data. It offers predictive insights as potential imaging substitutes, enables dynamic treatment adjustments, and identifies resistance to therapies.

Clinical trial identification

NCT04258137.

Editorial acknowledgement

Legal entity responsible for the study

A. Italiano.

Funding

Roche.

Disclosure

A. Italiano: Financial Interests, Personal, Advisory Board: Bayer, Roche, Philips, Chugai, GSK; Financial Interests, Institutional, Coordinating PI: Bayer, AstraZeneca, Roche, MSD, Ipsen, Merck. T. Grellety: Financial Interests, Personal, Advisory Board: AstraZeneca, Pfizer, Novartis, Lilly, Roche. M. Cabart: Financial Interests, Institutional, Local PI, Amplitude Clinical Trial: Janssen; Other, Travel fees, accommodations and registration for ASCO 2024: Daiichi Sankyo; Other, Travel fees, accommodations and registration for ESMO 2022: Janssen. I. Soubeyran: Financial Interests, Institutional, Invited Speaker, congress; meeting: AstraZeneca; Financial Interests, Institutional, Invited Speaker: Roche, Amgen; Financial Interests, Institutional, Funding, research funding: AstraZeneca; Other, Expert meeting: Sophia Genetics. S. Cousin: Financial Interests, Personal, Advisory Board: AbbVie, Amgen, Lilly, Roche; Non-Financial Interests, Principal Investigator: AstraZeneca, Daiichi Sankyo, Gilead, Takeda, Sanofi, MSD, GSK, BMS. All other authors have declared no conflicts of interest.