Abstract
PURPOSE: Major pathologic response (MPR), defined as ≤ 10% of residual viable tumor (VT), is a prognostic factor in non-small cell lung cancer (NSCLC) after neoadjuvant therapy. This study evaluated interobserver reproducibility in assessing MPR, compared area-weighted and unweighted VT (%) calculation, and determined optimal VT (%) cutoffs across histologic subtypes for survival prediction. MATERIALS AND METHODS: This retrospective study included 108 patients with NSCLC who underwent surgical resection after neoadjuvant chemotherapy or chemoradiation at Seoul National University Bundang Hospital between 2009-2018. Three observers with varying expertise independently assessed tumor bed and VT (%) based on digital whole-slide images. RESULTS: Reproducibility in tumor bed delineation was reduced in squamous cell carcinoma (SqCC) with smaller tumor bed, although overall concordance was high (Dice coefficient, 0.96; intersection-over-union score, 0.92). Excellent agreement was achieved for VT (%) (intraclass correlation coefficient=0.959) and MPR using 10% cutoff (Fleiss' kappa=0.911). Shifting between area-weighted and unweighted VT (%) showed only one case differing in MPR status out of 81 cases. The optimal cutoff was 10% for both adenocarcinoma (ADC) and SqCC. MPR+ was observed in 18 patients (17%), with SqCC showing higher MPR+ rates (p=0.044), lower VT (%) (p < 0.001), and better event-free survival (p=0.015) than ADC. MPR+ significantly improved overall survival (p=0.023), event-free survival (p=0.001), and lung cancer-specific survival (p=0.012). CONCLUSION: While MPR assessment demonstrated robust reproducibility with minimal impact from the tumor bed, attention is warranted when evaluating smaller tumor beds in SqCC. A 10% cutoff reliably predicted survival across histologic subtypes with higher interobserver reproducibility.