Suspicious urine cytology (class III) in patients with bladder cancer: should it be considered as negative or positive?

OBJECTIVES: Urine cytology is the gold standard in the diagnosis and follow-up of bladder cancer. Cytology, however, exhibits variable sensitivity depending on tumour grade and interpretation of urine specimens is highly dependent on the skill of the examiner. Positive cytology, classes IV and V by Papanicolaou classification, is a strong predictor for coexisting or subsequent malignancy, while the role of suspicious cytology, class III, is controversial. The objective of the study was to evaluate the role of the suspicious finding in cytological analysis, and whether it should be considered as a negative or positive sign for coexisting malignancy.

MATERIAL AND METHODS: Six hundred and fifty-two consecutive patients with bladder cancer were studied in a prospective multicenter trial. One hundred and fifty-one of the patients were newly diagnosed, and the remaining 501 patients were under follow-up. A voided urine sample was obtained prior to TURB or prior to routine follow-up cystoscopy in those under the surveillance and split for culture and cytology. The cytopathological results were analyzed by a central review and only patients with samples available for review analysis were included. Sensitivity and specificity, as well as positive (PPV) and negative (NPV) predictive values of urine cytology were calculated by classifying the class III samples as negative or positive.

RESULTS: A total of 570 patients were evaluable. One hundred and twenty nine (22.6%) were newly diagnosed and 441 were under follow-up, of whom 117 (26.5%) had recurrence. Cytology was classified as suspicious in 33/129 (25.6%) patients with primary tumour, and in 41/441 (9.3%) of those under the follow-up, of whom 20 (48.8%) had recurrence. Sensitivity increased from to 31.0% to 56.6% in primary tumours (p < 0.001) and from 17.8% to 34.7% in recurrent tumours (p < 0.001) if class III was determined as positive, whereas the specificity decreased from 96.6% to 90.1% (p < 0.001). Accordingly, the NPV increased from 76.3% to 79.1% and the PPV decreased from 65.6% to 56.2%.

CONCLUSIONS: The poor sensitivity of voided urine cytology improved significantly when suspicious samples were determined as positive while the specificity remained high, a clear advantage compared with most of the new tumour marker tests. In addition, nearly half of the follow-up patients with suspicious class III cytology had recurrence implying that this patient category is at substantial risk for co-existing malignancy. Therefore, it is recommended that suspicious class III cytology together with class IV and V specimens should be considered positive.