Baseline Tumor Proliferation and Ki-67 Are Associated With Pathological Response to Neoadjuvant Chemoimmunotherapy in...

Heterogeneity in the pathological response to neoadjuvant chemoimmunotherapy underscores the need for predictive biomarkers in resectable non-small cell lung cancer (NSCLC). This study identified baseline molecular features-particularly tumor proliferation signatures-associated with pathological response. Two retrospective cohorts (test, n = 81; validation, n = 107) of patients with NSCLC who received neoadjuvant chemoimmunotherapy were analyzed. Bulk RNA sequencing was performed on baseline biopsies from the test cohort, with immunohistochemistry (IHC) for PD-L1 and Ki-67 in both cohorts. Outcomes included major pathological response (MPR), pathological complete response (pCR), and event-free survival. In the test cohort, transcriptomic analysis showed that responders had significant upregulation of proliferation-related genes (eg, TP63, SOX2, NTRK2, and HMGA2) and activation of proliferation signatures (eg, chromosomal instability signature and core embryonic stem cell-like module), without activation of canonical immune-inflamed pathways. PD-L1 expression had limited predictive value (area under the curves [AUCs], 0.56-0.59 for MPR and 0.52-0.54 for pCR). In contrast, proliferation markers demonstrated higher performance: MKI67 messenger RNA predicted both MPR and pCR with an AUC of 0.71, and the Ki-67 IHC index yielded AUCs of 0.71 for MPR and 0.64 for pCR. The predictive value of the Ki-67 IHC index was validated in the independent cohort, with AUCs of 0.74 for MPR and 0.70 for pCR, and this association was generally consistent across squamous cell carcinoma and adenocarcinoma. A Ki-67 index ≥50% was significantly correlated with improved event-free survival in both cohorts (both Ps < .05). These findings indicate baseline tumor proliferation, particularly MKI67/Ki-67, as predictors of pathological response in NSCLC patients receiving neoadjuvant chemoimmunotherapy, supporting its potential clinical utility for patient selection.