BackgroundPrimary lung cancer is the leading cause of cancer death in the United States. Most lung cancers are diagnosed in an outpatient setting, but a subset requires intraoperative diagnosis. Two intraoperative diagnostic methods are available, frozen section (FS) and fine needle aspiration (FNA) cytology. This study compares intraoperative FNA cytology and FS based diagnosis in thoracic malignancies within the same clinical practice.MethodsPathology reports from thoracic intraoperative FNA cytology or FS (January 2017–December 2019) were reviewed. Resection diagnosis was the gold standard. If unavailable, concurrent biopsy and final FNA cytology diagnosis were the gold standard.ResultsOf 300 FNA specimens (155 patients), 142 (47%) cases were benign, and 158 (53%) were malignant. Adenocarcinoma was the most common malignant diagnosis (40%), followed by squamous cell carcinoma (26%), neuroendocrine tumors (18%), and other (16%). Intraoperative FNA yielded 88% sensitivity, 99% specificity, and 92% accuracy (p < .001).Of 298 FS specimens (252 patients), 215 (72%) cases were malignant and 83 (28%) were benign. Adenocarcinomas was the most common malignant diagnosis (48%), followed by squamous cell carcinoma (25%), metastatic carcinomas (13%), and other (14%). FS yielded 97% sensitivity, 99% specificity, and 97% accuracy (p < .001).ConclusionOur findings confirm FS is the gold standard for intraoperative diagnosis. FNA cytology may be useful as a non‐invasive, inexpensive initial diagnostic tool intraoperatively, given the similar specificity (99% FNA, 99% FS) and accuracy (92% FNA, 97% FS). Negative FNA could be followed by the costlier and invasive FS. We encourage surgeons to utilize intraoperative FNA first.