Improved Prediction of Chemoradiation and Immune Checkpoint Blockade Efficacy with Dynamic bTMB Combined with ctDNA in Unresectable Locally Advanced NSCLC

Abstract

Consolidation durvalumab after definitive chemoradiation therapy (CRT) has become the new standard of care for patients with unresectable locally advanced non-small-cell lung cancer (LA-NSCLC). However, only a fraction of patients can benefit from it, and current decision-making procedures have limited accuracy. Blood-based tumor mutational burden (bTMB) is a promising biomarker, but whether bTMB alone or combined with circulating tumor DNA (ctDNA) can predict the efficacy of definitive CRT and immune checkpoint inhibitors (ICIs) in patients with LA-NSCLC remains unclear. This cohort study enrolled patients diagnosed with unresectable LA-NSCLC from 2018 to 2022. Patients were assigned to the cohort A receiving definitive CRT alone (intensity modulated radiation therapy or volumetric modulated arc therapy with the prescribed dose of 60 Gy, concurrently or sequentially with two or more cycles of platinum-based doublet chemotherapy), or cohort B undergoing definitive CRT and immunotherapy. Peripheral blood specimens were collected before and after CRT and subjected to next-generation sequencing panel to analyze ctDNA and bTMB. The dynamic change in bTMB (∆bTMB) was calculated as the bTMB level after CRT minus the baseline bTMB level. Potential correlations were identified by Spearman's correlation tests and expressed as R coefficients. Time-dependent receiver operating characteristic curves and areas under the curve (AUCs) were employed to evaluate the predictive power of different models on survival. A total of 73 LA-NSCLC patients were included, with 70 (95.9%) at stage III and 3 at stage II (4.1%). Thirty-six patients (49.3%) assigned to the cohort A were treated with CRT alone and 37 (50.7%) in the cohort B receiving CRT and ICIs. Patients with CRT + ICIs in the cohort B showed significantly improved overall survival (OS; P < 0.01) and progression-free survival (PFS; P = 0.02) than those with CRT alone. Baseline bTMB at cutoff values of 4 to 22 did not predict outcomes (all P > 0.10), while patients with increased bTMB (∆bTMB > 0) after CRT had significantly worse OS and PFS (both P < 0.01) than those with decreased or stable bTMB (∆bTMB ≤ 0). ∆bTMB was independent of the ctDNA level after CRT (P = 0.76, R = -0.04). Patients were further divided into 3 groups based on ∆bTMB and post-CRT ctDNA (∆bTMB > 0/detectable ctDNA vs. ∆bTMB ≤ 0/detectable ctDNA vs. ∆bTMB ≤ 0/undetectable ctDNA), and significant survival differences were observed in the pairwise comparisons of 3 groups (all P < 0.05). The model of ∆bTMB combined with post-CRT ctDNA status exhibited the optimal predictive power on both OS (AUC = 0.80) and PFS (AUC = 0.75) and outperformed each factor, which had been respectively validated in the cohort A and B as well. Dynamic bTMB (∆bTMB) combined with post-CRT ctDNA status is a novel and effective biomarker model of predicting survival outcomes in LA-NSCLC patients treated with CRT ± ICIs, and outperforms each individual feature.