Invasive Mediastinal Staging for Lung Cancer by The Society of Thoracic Surgeons Database Participants.

BACKGROUND: Prior studies suggest underutilization of invasive mediastinal staging for lung cancer. We hypothesized that The Society of Thoracic Surgeons General Thoracic Surgery Database (STS-GTSD) participants would have higher rates of invasive staging compared with previous reports. METHODS: We conducted a retrospective cohort study (2012 to 2016) of lung cancer patients staged by computed tomography and positron-emission tomography and first treated with an anatomic resection. We defined invasive staging by the use of mediastinoscopy, endosonography, or thoracoscopy. Standardized incidence ratios were used to compare participant-level rates of invasive staging, and Poisson regression was used to identify factors associated with invasive staging. RESULTS: Among 29,015 patients across 256 participating STS-GTSD sites, 34% (95% confidence interval: 33% to 34%) underwent invasive staging. The overall rate of invasive staging did not change between 2012 and 2016 (p trend = 0.16). Increasing clinical stage and features suggestive of a central tumor were associated with invasive staging (p < 0.001). Rates of invasive staging among patients with clinical stage IB or greater or features suggestive of a central tumor were 43% (95% confidence interval: 42% to 44%) and 52% (95% confidence interval: 50% to 54%), respectively. There was a more than 40-fold variation in rates of invasive staging across 251 centers contributing at least 10 cases (standardized incidence ratio: lowest = 0.08; highest = 3.26); 66 sites (26%) performed invasive mediastinal staging less often than average and 77 sites (31%) performed invasive staging more often than average. CONCLUSIONS: The STS-GTSD participants performed invasive mediastinal staging more frequently than prior reports, and yet only in a minority of patients. Rates of invasive mediastinal staging vary widely across STS-GTSD participants.

Duke Scholars

Author Andrzej Stanislaw Kosinski Biostatistics & Bioinformatics, Division of Biostatistics