Objective: to determine the influence of the knowledge of the endoscopic findings and the influence of the patient's history on the cytologist's judgement, as urinary cytology is known to be subjective and has several limitations, in particular a high inter- and intra-observer variability.
Patients and methods: we analysed the cytological and histological findings of patients who underwent transurethral resection of a bladder tumour, and determined whether the cytologist was aware of the endoscopic findings or not. The sensitivity and specificity of cytology were calculated with or without this knowledge, and that of the patients' bladder cancer history.
Results: the findings of 1705 patients were reviewed; in 641 the histological examination confirmed a malignant tumour and 1046 were classified as benign. The sensitivity of cytology was 66.0% and the specificity was 78.4%. The cytologist was aware of the endoscopic finding and patient history in 742 cases, and unaware of the endoscopic findings in 963. The specificity was higher in the latter group (80.2% vs 73.0%; P= 0.006). The specificity in patients with the endoscopic findings described as 'negative', 'inflammation', 'scar tissue', 'flat lesion', 'suspicious for tumour', and 'exophytic tumour' was 89.8%, 89.9%, 85.0%, 77.1%, 63.2% and 48.6%, respectively (P < 0.001). In 898 patients the history was negative for bladder tumours. Among these patients the sensitivity and specificity of cytology was 67.3% and 79.7%; the sensitivity and specificity was 65.4% and 74.8% for the 807 patients with a positive history of bladder cancer (P= 0.054).
Conclusion: both being aware of the endoscopic findings and a positive patient history for bladder cancer lowers the specificity of cytology. Consequently, the cytologist should be unaware of the endoscopic findings.