Prognostic value of MIB-1 antibody labeling index to predict response to bacillus Calmette-Guerin therapy in a high-risk selected population of patients with stage T1 grade G3 bladder cancer

Purpose: Proliferation rate is an important determinant of bladder tumor pr ogression. However, this factor has not yet been correlated to bacillus Cal mette-Guerin (BCG) therapy response in a selected high-risk population of p atients with stage T1 grade G3 bladder cancer. To assess the predictive val ue of the proliferation rate, an immunoreactivity test with monoclonal anti bodies MIB-1 was carried out. The aim of this study was to evaluate the pro gnostic value of an MIB-1 labeling index by selecting a group of responsive patients prior to intravesical therapy. Materials and Methods: After complete transurethral resection, 35 patients with T1G3 bladder carcinoma received 6 weekly installations of BCG (intrave sical Pasteur strain: 75 mg in 50 ml course of BCG therapy). After treatmen t a cystoscopy and randomized biopsies of the bladder mucosa were carried o ut and all recurrences were systematically resected. All tissue samples wer e fixed in Bouin's solution, embedded in paraffin and stained with hematoxy lin-eosin-safran. Pathologists had sufficient material to perform immunomar king in 25 patients using peroxidase-anti peroxidase (PAP) technique, with antiprotein monoclonal antibody MIB-1 (Immunotech, Marseilles, France) to s tudy MIB-1 expression before BCG therapy. Consensus was obtained from three independent pathologists for all sections. The results were expressed in a percentage of marked nuclei. Ten percent increment thresholds were establi shed from 10 to 60%. Contingency tables were established, chi 2 (p1) and Fi sher exact test(p2) were performed for each threshold of 10%. Results: Median follow-up was 57.3 months (range 25-144). Of the 25 patient s, 8 (32%) did not respond to BCG therapy, 17 (68%) responded positively. W ith a 20% threshold, there was a statistical difference (p1 = 0.03, p2 = 0. 04) between responder (R) and nonresponder (NR) patients. All the 7 patient s with less than 20% of nuclear activity positively responded to BCG. At th is threshold level, sensitivity was high but specificity low (positive pred ictive value = 0.44). If we consider other reactivity thresholds there were no statistical differences between R and NR patients (10%) threshold p1 = 0.13, p2 = 0.19; 30% p1 and p2 = 0.20; 40% p1 = 0.82, p2 = 0.61; 50% p1 = 0 .57, p2 = 0.55). Conclusion: Our study indicates that the proliferation rate, assessed by MI B-1 immunoreactivity in Bouin's solution-fixed primary tissue, could be a u seful predictive marker of outcome in T1G3 bladder carcinoma. With a 20% re activity cut-off, a negative MIB-1 immunomarking appears to predict a posit ive response to BCG instillations. However, on the other hand, MIB-1 is of limited clinical use because the low specificity of th is test cannot predi ct failure a nd then select candidates for cystectomies. Copyright (C) 2000 S. Karger AG, Basel.