Background Clinical staging of non-small cell lung cancer (NSCLC) determines the initial treatment offered to a patient. The similarity between clinical and pathologic staging in some studies is as low as 50%, and others publish results as high as 91%. The Dutch Lung Surgery Audit is a clinical database that registers the clinical and pathologic TNM of almost all NSCLC patients who undergo operations in the Netherlands. The objective of this study was to determine the accuracy of clinical staging of NSCLC. Methods Prospective data were derived from the Dutch Lung Surgery Audit in 2013 and 2014. Patients were included if they had undergone a surgical resection for stage IA to IIIB NSCLC without neoadjuvant treatment and had a positron emission tomography–computed tomography scan as part of the clinical workup. Clinical (c)TNM and pathologic (p)TNM were compared, and whether discrepancy was based on tumor or nodal staging was determined. Results From 2,834 patients identified, 2,336 (82.4%) fulfilled the inclusion criteria and had complete data. Of these 2,336, 1,276 (54.6%) were staged accurately, 707 (30.3%) were clinically understaged, and 353 (15.1%) were clinically overstaged. In the understaged group, 346 patients had a higher pN stage (14.8%), of which 148 patients had unforeseen N2 disease (6.3%). In the overstaged group, 133 patients had a cN that was higher than the pN (5.7%). Conclusions Accuracy of NSCLC staging in the Netherlands is low (54.6%), even in the era of positron emission tomography–computed tomography. Especially accurate nodal staging remains challenging. Future efforts should include the identification of specific pitfalls in NSCLC staging.@en