A Predictive Model for 90-Day Mortality After Lobectomy in Patients with Non-Small Cell Lung Cancer: A Multicenter Retrospective Study

Background: Over the past decade, multiple prognostic models have been developed to estimate short-term (including in-hospital) mortality following lobectomy. However, 90-day postoperative mortality is not as widely used as a standard measure for evaluating outcomes after lobectomy. The objective of this study was to evaluate 90-day mortality following lobectomy with mediastinal lymphadenectomy for non-small cell lung cancer (NSCLC), and to identify factors independently associated with this outcome. Methods and Results: This retrospective multicenter study included data from 700 patients who underwent anatomical lobectomy with systematic mediastinal lymphadenectomy for histologically confirmed NSCLC. A clinically homogeneous cohort was formed to validate the assessment of intra- and postoperative risk factors associated with 90-day mortality. Among 700 patients included in the study, the 90-day postoperative mortality rate was 3.7%. Multivariate logistic regression identified the following independent predictors of 90-day mortality: prolonged air leak (AOR=2.505; 95% CI: 1.115–5.629, P=0.026), intraoperative blood loss (AOR=1.003; 95% CI: 1.001–1.005, P=0.004), and forced expiratory volume in one second (AOR=0.965; 95% CI: 0.945–0.985, P=0.001). To assess the relationship between the probability of 90-day postoperative mortality and the value of the logistic function P, a ROC analysis was performed. The area under the ROC curve (AUC) was 0.719 (95% CI: 0.606–0.832; P<0.001). The sensitivity and specificity of the resulting predictive model were 65.4% and 77.1%, respectively. Conclusion: The proposed model incorporating three key clinical variables may serve as a practical tool for postoperative risk stratification and guiding follow-up strategies in patients undergoing lobectomy for NSCLC.