The Society of Thoracic Surgeons 2024 Risk Models for Lung Cancer Resection: Continued Refinement and Improved...

Area of Science:

• Thoracic surgery
• Cardiothoracic surgery
• Surgical oncology

Background:

• The Society of Thoracic Surgeons (STS) General Thoracic Surgery Database (GTSD) has been instrumental in developing risk models for pulmonary resection in cancer patients.
• This study aimed to refine existing risk models by utilizing a more contemporary and inclusive patient cohort.

Purpose of the Study:

• To develop and validate updated risk prediction models for operative mortality, major morbidity, and composite morbidity or mortality following pulmonary resection for cancer.
• To improve the accuracy and applicability of risk models by incorporating a broader range of patient data and including nonelective operations.

Main Methods:

• Utilized data from adult patients undergoing pulmonary resection for cancer between 2015 and 2022 from the STS GTSD.
• Included nonelective operations in the analysis, a departure from previous models.
• Developed separate logistic regression models for operative mortality, major morbidity, and composite morbidity or mortality, validated using 9-fold cross-validation.

Main Results:

• Included 140,927 patients from 337 centers; operative mortality was 1.1%, major morbidity 7.3%, and composite morbidity or mortality 7.6%.
• Identified novel predictors such as interstitial lung disease, diffusing capacity of lung for carbon monoxide, and payer status.
• Achieved superior discrimination compared to previous STS models for operative mortality (C-statistic = 0.80) and composite morbidity or mortality (C-statistic = 0.70).

Conclusions:

• Pulmonary resection for cancer continues to have low rates of major morbidity and operative mortality within the STS GTSD cohort.
• The newly derived risk models demonstrate enhanced performance, broader real-world applicability, and clinical validity compared to prior models.