Quantitative FDG PET/CT may help risk-stratify early-stage non-small cell lung cancer patients at risk for recurrence following anatomic resection

• Published in: Journal of thoracic disease 2019-04, Vol.11 (4), p.1106-1116
• Main Authors: Harmon, Stephanie, Seder, Christopher W, Chen, Song, Traynor, Anne, Jeraj, Robert, Blasberg, Justin D
• Format: Article
• Language: English
• Subjects: Original
• ISSN: 2072-1439; 2077-6624
• DOI: 10.21037/jtd.2019.04.46

Preoperative identification of non-small cell lung cancer (NSCLC) patients at risk for disease recurrence has proven unreliable. The extraction of quantitative metrics from imaging based on tumor intensity and texture may enhanced disease characterization. This study evaluated tumor-specific F-fluorodeoxyglucose ( F-FDG) positron emission tomography/computerized tomography (PET/CT) uptake patterns and their association with disease recurrence in early-stage NSCLC. Sixty-four stage I/II NSCLC patients who underwent anatomic resection between 2001 and 2014 were examined. Pathologically or radiographic confirmed disease recurrence within 5 years of resection comprised the study group. Quantitative imaging metrics were extracted within the primary tumor volume. Squamous cell carcinoma (SCC) (N=27) and adenocarcinoma (AC) (N=41) patients were compared using a Wilcoxon signed-rank test. Associations between imaging and clinical variables with 5-year disease-free survival (DFS) and overall survival (OS) were evaluated by Cox proportional-hazards regression. Clinical and pathologic characteristics were similar between recurrence (N=34) and patients achieving 5-year DFS (N=30). Standardized uptake value (SUV) and SUV varied significantly by histology, with SCC demonstrating higher uptake intensity and heterogeneity patterns. Entropy-grey-level co-occurrence matrix (GLCM) was a significant univariate predictor of DFS (HR =0.72, P=0.04) and OS (HR =0.65, P=0.007) independent of histology. Texture features showed higher predictive ability for DFS in SCC than AC. Pathologic node status and staging classification were the strongest clinical predictors of DFS, independent of histology. Several imaging metrics correlate with increased risk for disease recurrence in early-stage NSCLC. The predictive ability of imaging was strongest when patients are stratified by histology. The incorporation of F-FDG PET/CT texture features with preoperative risk factors and tumor characteristics may improve identification of high-risk patients.