Pre-Treatment Levels of ctDNA for Long-Term Survival Prediction in Stage IIIA NSCLC Treated With Neoadjuvant Chemo-Immunotherapy

ABSTRACT

Introduction: There are currently no predictive biomarkers for long-term survival after neoadjuvant chemoimmunotherapy. However, the identification of non-small lung cancer (NSCLC) patients who obtain long-term benefit from chemoimmunotherapy is essential to optimize therapies. Methods: Using samples from NADIM clinical trial (NCT03081689), in which resectable stage IIIA NSCLC patients were treated with neoadjuvant chemo-immunotherapy with nivolumab, we have evaluated the capacity of ctDNA levels before treatment initiation to predict overall survival (OS) and progression-free survival (PFS) by calculating Harrell’s C-statistic and we compare its predictive value with classical survival surrogates as the pathological response and clinical response assessed according to RECIST criteria v.1.1. The ctDNA was analyzed by NGS, using the Oncomine Pan-Cancer Cell-Free Assay™ (Thermo Fisher Scientific®). To explore the prognostic value of the amount of ctDNA at baseline, for each positive plasma sample, we calculated the sum of the mutant allele frequency (MAF) for all detected mutations. Patients who died from COVID19 were excluded from this analysis. Results: In our study, clinical responses based on RECIST criteria were not predictive for OS or PFS. On the contrary, in the multivariate analysis, patients with low ctDNA levels (<1% MAF), in the baseline sample, had significantly improved PFS and OS than patients in whom the opposite situation occurred (adjusted HR 0.22;95%CI 0.06-0.75;P=0.016 and adjusted HR 0.04;95%CI 0.00-0.45;P=0.008 for PFS and OS, respectively). The adjusted C-statistic (c) to predict PFS for ctDNA was 0.68 (95%CI 0.51-0.84), which was superior to that of RECIST criteria (c=0.61;95%CI 0.45-0.78) and similar to that of pathological response (c=0.68;95%CI 0.52-0.84). Similarly, baseline ctDNA levels predicted OS (c=0.85;95%CI 0.72-0.99) better than RECIST criteria (c=0.68;95%CI 0.44-0.93). Conclusion: Pre-treatment ctDNA levels predicted more accurately long-term survival than radiological assessments in NADIM study and might be useful for the design of new clinical trials.