Studies that compared FISH with urine cytology for detection of bladder cancer that used cystoscopy or pathological examination as the reference standard were eligible for inclusion. Patients who underwent urine cytology served as the control group. Included studies had to report sufficient data to populate 2x2 contingency tables of test performance (number of true positive, false negative, false positive and true negative test results) for more than 80% of participants.
Studies of patients with concomitant pelvic carcinoma, ureteral carcinoma or other urothelial tumours where separate data could not be extracted for bladder cancer were excluded.
Included studies were published between 2000 and 2008. Where reported, study participants were aged between 28 and 94 years. Criteria for a positive bladder cancer test on FISH varied between studies, but in most cases included five or more cells with gains of two or more chromosomes. One study examined chromosomes 9 and 17 and the rest examined chromosomes 3, 7, 9P21 and 17.
Two reviewers independently assessed studies for inclusion. Any disagreements were resolved by discussion with a third reviewer.