Does urinary cytology have a role in haematuria investigations?

To determine the diagnostic accuracy of urinary cytology to diagnose bladder cancer and upper tract urothelial cancer (UTUC) as well as the outcome of patients with a positive urine cytology and normal haematuria investigations in patients in a multicentre prospective observational study of patients investigated for haematuria.

The DETECT I study (clinicaltrials.gov NCT02676180) recruited patients presenting with haematuria following referral to secondary case at 40 hospitals. All patients had a cystoscopy and upper tract imaging (renal bladder ultrasound [RBUS] and/ or CT urogram [CTU]). Patients, where urine cytology were performed, were sub-analysed. The reference standard for the diagnosis of bladder cancer and UTUC was histological confirmation of cancer. A positive urine cytology was defined as a urine cytology suspicious for neoplastic cells or atypical cells.

Of the 3 556 patients recruited, urine cytology was performed in 567 (15.9%) patients from nine hospitals. Median time between positive urine cytology and endoscopic tumour resection was 27 (IQR: 21.3-33.8) days. Bladder cancer was diagnosed in 39 (6.9%) patients and UTUC in 8 (1.4%) patients. The accuracy of urinary cytology for the diagnosis of bladder cancer and UTUC was: sensitivity 43.5%, specificity 95.7%, positive predictive value (PPV) 47.6% and negative predictive value (NPV) 94.9%. A total of 21 bladder cancers and 5 UTUC were missed. Bladder cancers missed according to grade and stage were as follows: 4 (19%) were ≥ pT2, 2 (9.5%) were G3 pT1, 10 (47.6%) were G3/2 pTa and 5 (23.8%) were G1 pTa. High-risk cancer was confirmed in 8 (38%) patients. There was a marginal improvement in sensitivity (57.7%) for high-risk cancers. When urine cytology was combined with imaging, the diagnostic performance improved with CTU (sensitivity 90.2%, specificity 94.9%) superior to RBUS (sensitivity 66.7%, specificity 96.7%). False positive cytology results were confirmed in 22 patients, of which 12 (54.5%) had further invasive tests and 5 (22.7%) had a repeat cytology. No cancer was identified in these patients during follow-up.

Urine cytology will miss a significant number of muscle-invasive bladder cancer and high-risk disease. Our results suggest that urine cytology should not be routinely performed as part of haematuria investigations. The role of urine cytology in select cases should be considered in the context of the impact of a false positive result leading to further potentially invasive tests conducted under general anaesthesia.