Background: Bladder cancer is the fourth most common malignancy in men and a considerable disease burden globally. Multiple studies have focused on the accuracy of optical coherence tomography for bladder cancer diagnosis; however, the findings are inconsistent. Here, we assessed the accuracy of optical coherence tomography for bladder cancer diagnosis. Methods: Embase, PubMed, Medline, Web of Science, and the Cochrane Library database were searched for relevant studies from the earliest date available through March 11, 2019. Studies evaluating the accuracy of optical coherence tomography bladder cancer diagnosis were included. Pooled sensitivity, specificity, and area under the curve values of weighted symmetric summary receiver operating curves, were calculated at the perlesion level. Results: Eleven studies, with a total of 1933 lesions, were included in the final analysis. The pooled results indicated that optical coherence tomography can differentiate bladder cancer from benign lesions: sensitivity, 94.9% (95% confidence interval: 92.7%-96.6%); specificity, 84.6% (95% confidence interval: 82.6%-86.4%); area under the curve, 0.97. Moreover, compared with optical coherence tomography alone, combined optical coherence tomography and fluorescence cystoscopy increased the diagnostic accuracy (sensitivity, 94.3% vs. 87.3%; specificity, 89.2% vs. 73.9%). Cross-polarization optical coherence tomography could also distinguish bladder cancer from normal tissue: sensitivity, 92.0% (95% confidence interval: 87.0%-95.6%); specificity, 84.4% (95% confidence interval: 81.7%-86.9%); area under the curve, 0.95. Conclusions: Optical coherence tomography can accurately differentiate malignant from benign bladder lesions, particularly when combined with fluorescence cystoscopy.