Molecular markers reliably predicting failure or success of Bacillus Calmette-Guerin (BCG) in the treatment of nonmuscle-invasive urothelial bladder cancer (NMIBC) are lacking. The aim of our study was to evaluate the value of cytology and chromosomal aberrations detected by fluorescence in situ hybridization (FISH) in predicting failure to BCG therapy. Sixty-eight patients with NMIBC were prospectively recruited. Bladder washings collected before and after BCG instillation were analyzed by conventional cytology and by multitarget FISH assay (UroVysion 1 , Abbott/ Vysis, Des Plaines, IL) for aberrations of chromosomes 3, 7, 17 and 9p21. Persistent and recurrent bladder cancers were defined as positive events during follow-up. Twenty-six of 68 (38%) NMIBC failed to BCG. Both positive post-BCG cytology and positive post-BCG FISH were significantly associated with failure of BCG (hazard ratio (HR)5 5.1 and HR5 5.6, respectively; p < 0.001 each) when compared to those with negative results. In the subgroup of nondefinitive cytology (all except those with unequivocally positive cytology), FISH was superior to cytology as a marker of relapse (HR5 6.2 and 1.4, respectively). Cytology and FISH in post-BCG bladder washings are highly interrelated and a positive result predicts failure to BCG therapy in patients with NMIBC equally well. FISH is most useful in the diagnostically less certain cytology categories but does not provide additional information in clearly malignant cytology. '
UICCKey words: nonmuscle-invasive urothelial bladder cancer; Bacillus Calmette-Guerin; cytology; fluorescence in situ hybridization High-and intermediate risk nonmuscle-invasive urothelial bladder cancers (NMIBC) frequently recur and, more importantly, can progress to potentially life-threatening muscle-invasive carcinomas. Intravesical instillation of Bacillus Calmette-Guerin (BCG) after transurethral resection of bladder cancer (TURB) is the most effective treatment. However, 50% of NMIBC do not respond to BCG treatment. [1][2][3] The progression rate has been estimated to be 9.8%, and 5.6% of patients with NMIBC die from bladder cancer. 4 Although an early cystectomy provides an excellent chance of cure, 50% of patients are overtreated with this aggressive approach. 5,6 Decision on therapy strategy of bladder cancer is based on clinicopathological tumor characteristics (number of tumors, tumor size, prior recurrence rate, tumor category, presence of pTis and tumor grade). Several clinicopathological tumor characteristics have recently been suggested as predictors for recurrence and progression after BCG therapy. 7 However, clinicopathological or molecular markers that can reliably predict failure or success of BCG therapy are still lacking. 8 Such markers would be of great value to individually tailor the follow-up schedule and stratify patients at high risk of tumor recurrence or progression.Although conventional cytology has a very high diagnostic accuracy for high-grade urothelial carcinomas, false-positive results have been repor...