Prognostic Value of Histology-specific Ki-67 Labeling Index Cutoffs in Stage I-III Non-small Cell Lung Cancer

Background/Aim: Although the Ki-67 labeling index (LI) is widely used as a marker of tumor proliferative activity, its prognostic utility in non-small cell lung cancer (NSCLC) remains uncertain because of heterogeneous study populations and the lack of standardized, histology-specific cutoff values. Patients and Methods: We retrospectively analyzed 1,245 patients with completely resected pathologic stage I-III NSCLC. Time-dependent receiver operating characteristic (ROC) analysis for postoperative recurrence within 5 years was used to determine Ki-67 LI cutoff values. Associations between Ki-67 LI status and recurrence-free survival (RFS) and overall survival (OS) were assessed using Kaplan-Meier analysis and multivariable Cox proportional hazards models. Results: Ki-67 LI was significantly higher in squamous cell carcinoma than in adenocarcinoma (median, 40.0% vs. 15.0%; p<0.001). Time-dependent ROC analysis yielded a cutoff value of 21.0% for the overall NSCLC cohort and histology-specific cutoff values of 13.0% for adenocarcinoma and 40.5% for squamous cell carcinoma. The corresponding 5-year area under the curves (AUCs) were 0.656 for the overall NSCLC cohort, 0.691 for adenocarcinoma, and 0.607 for squamous cell carcinoma. After multivariable adjustment, high Ki-67 LI was independently associated with poorer RFS in the overall NSCLC cohort [hazard ratio (HR)=1.40; 95% confidence interval (CI)=1.11-1.77; p=0.005], adenocarcinoma cohort (HR=1.57; 95%CI=1.15-2.16; p=0.005), and squamous cell carcinoma cohort (HR=1.82; 95%CI=1.22-2.73; p=0.004). High Ki-67 LI was also independently associated with poorer OS in the overall NSCLC and adenocarcinoma cohorts, but not in the squamous cell carcinoma cohort. Conclusion: High Ki-67 LI was independently associated with an increased risk of postoperative recurrence in both adenocarcinoma and squamous cell carcinoma. Histology-specific cutoffs derived by time-dependent ROC analysis may support Ki-67-based postoperative risk stratification in resected NSCLC.