Can Preoperative PET/CT SUVmax Predict Aggressive Pathological Features in Non-Small Cell Lung Cancer? A Retrospective Cohort Study of STAS, LVI, and VPI

Abstract Background Spread through air spaces (STAS), lymphovascular invasion (LVI), and visceral pleural invasion (VPI) are adverse prognostic features in non-small cell lung cancer (NSCLC) that can only be assessed postoperatively. This study evaluated whether preoperative primary tumor SUVmax on 18F-FDG PET/CT independently predicts these features. Methods In this retrospective cohort of 478 patients undergoing surgical resection for NSCLC, associations between primary tumor SUVmax and STAS, LVI, and VPI were assessed using univariate and multivariable logistic regression, adjusted for tumor size, histologic subtype, and other clinically relevant covariates. Discriminative performance was evaluated using receiver operating characteristic (ROC) curve analysis. Results STAS, LVI, and VPI were present in 269 (56.3%), 237 (49.6%), and 202 (42.3%) patients, respectively. SUVmax was independently associated with all three outcomes: STAS (adjusted OR = 1.051, 95% CI 1.020–1.083, p = 0.001), LVI (adjusted OR = 1.083, 95% CI 1.048–1.119, p < 0.001), and VPI (adjusted OR = 1.044, 95% CI 1.015–1.073, p = 0.002). Discriminative performance was acceptable for LVI (AUC = 0.701) but weak for STAS (AUC = 0.578) and VPI (AUC = 0.584). Sensitivity analyses excluding neoadjuvant-treated patients and interaction testing confirmed these associations were independent of neoadjuvant treatment status. Conclusion Preoperative SUVmax is independently associated with STAS, LVI, and VPI in resected NSCLC, most strongly for LVI. Although its standalone discriminative value is limited, SUVmax—already available at diagnosis—may contribute to multivariable preoperative risk stratification alongside histologic subtype and tumor size.