Comparison of fluorescence in situ hybridization (FISH) and urine cytology in diagnosing urothelial carcinoma: a single-center retrospective cohort study

Background and purpose: Urothelial carcinoma (UC) is a prevalent malignant tumor of the urinary system, and early diagnosis is crucial for improving patient prognosis. This study evaluated the diagnostic efficacy of fluorescence in situ hybridization (FISH), urine cytology and their combination for UC, as well as for its different subtypes. Methods: This study included patients who underwent transurethral resection of bladder tumor (TURBT) from January 2022 to December 2023 and approved by Ethics Commetce of Fudan Univesity Shanghai Cancer Center, No.: 050432-4-2307E) that met the inclusion and exclusion criteria. We collected TURBT pathological results and pre-procedure FISH and cytology results. Diagnostic accuracy, sensitivity and specificity of FISH, cytology and their combination were analyzed and compared for urothelial carcinoma. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist and Standards for Reporting of Diagnostic Accuracy (STARD) were followed for this study. Results: A total of 283 patients were enrolled in this study, 136 were diagnosed with UC, and 147 were not. Of the 136 UC cases, 79 (58.09%) were invasive and 57 (41.91%) were non-invasive. In terms of malignancy grade, 112 (82.35%) were high-grade UC and 24 (17.65%) were low-grade UC. Using histopathology as the gold standard, the accuracy of FISH, cytology and their combination in diagnosing UC was 79.51%, 72.08% and 77.39%, respectively; sensitivity was 72.06%, 58.82% and 78.68%, respectively; specificity was 86.39%, 84.35% and 76.19%, respectively. The area under the curve (AUC) for FISH and the combination was similar but higher than that for cytology (0.792 vs 0.716, P=0.006; 0.774 vs 0.716, P=0.004); the Net Reclassification Improvement (NRI) for FISH compared to cytology was 15.28% (P=0.006). In the 79 cases of invasive UC, FISH had higher accuracy than cytology (86.28% vs 78.32%, P=0.011). The sensitivity of FISH and the combination was higher than that of cytology (86.08% vs 67.09%, P=0.004; 91.14% vs 67.09%, P<0.001), and the AUC values were also higher (0.808 vs 0.713, P=0.004; 0.784 vs 0.713, P=0.007). The NRI for FISH compared to cytology was 21.03% (P=0.003). In the 57 cases of non-invasive UC, the AUC values for all three methods were low (AUC<0.700). Among the 112 cases of high-grade UC, FISH had higher accuracy (84.94% vs 76.45%,P=0.005), and the combination had higher sensitivity (89.29% vs 66.07%, P<0.001) compared to cytology. The AUC values for FISH and the combination were also superior to that for cytology (0.847 vs 0.752, P=0.002; 0.827 vs 0.752, P=0.001). The NRI for FISH compared to cytology was 19.01% (P=0.003). In the 24 cases of low-grade UC, the AUC values for all three methods were low (AUC<0.600). Conclusion: For UC, particularly invasive and high-grade subtypes, FISH shows superior diagnostic efficacy compared to cytology. FISH alone offers accuracy and sensitivity comparable to the combination test, with higher specificity. In cases of non-invasive or low-grade UC, however, all three diagnostic methods demonstrate relatively low efficacy.