The STS 2024 Risk Models for Lung Cancer Resection: Continued Refinement and Improved Outcomes.

ABSTRACT

BACKGROUND: The Society of Thoracic Surgeons General Thoracic Surgery Database (STS GTSD) has been used to develop risk models for patients undergoing pulmonary resection for cancer. Leveraging a contemporary and more inclusive cohort, we sought to refine these models. METHODS: The study population consisted of adult patients in the STS GTSD who underwent pulmonary resection for cancer between 2015 and 2022. Unlike previous models, non-elective operations were included. Separate risk models were derived for operative mortality, major morbidity, and composite morbidity or mortality. Logistic regression with backward selection was used with predictors retained in models if p<0.10. All derived models were validated using 9-fold cross validation. Model discrimination and calibration were assessed for the overall cohort and for surgical procedure, demographic, and risk factor subgroups. RESULTS: Data from 140,927 patients at 337 participating centers were included in the study. Overall operative mortality rate was 1.1%, major morbidity 7.3%, and composite morbidity or mortality 7.6%. Novel predictors of short-term outcomes included interstitial lung disease, DLCO, and payor status. Overall discrimination was superior to previous STS pulmonary resection models for operative mortality [C-statistic = 0.80] and for composite morbidity or mortality [C-statistic = 0.70]. Model discrimination was comparable and model calibration was excellent across all procedure- and demographic-specific sub-cohorts. CONCLUSIONS: Among STS GTSD participants, major morbidity and operative mortality rates remain low following pulmonary resection. The newly derived pulmonary resection risk models demonstrate superior performance compared to previous models, with broader real-life applicability and clinical face validity.