EAU 2011 - Prediction of locally advanced bladder cancer before cystectomy and outcome assessment of patients with bladder cancer after cystectomy using artificial neural networks

VIENNA, AUSTRIA (UroToday.com) - Data of 2,477 patients from 8 German centers who had undergone radical cystectomy for bladder cancer were analyzed for oncological outcome by artificial neural network (ANN).

Freedom from tumor recurrence, cancer-specific survival and overall survival was 65%, 69% and 53%, respectively. The best ANN had 75%, 77% and 69% for the prediction of the given outcome parameters. In comparison, logistic regression models had an accuracy of 70%, 71% and 64%, respectively. The authors concluded that ANN performed better in predicting tumor recurrence, cancer-specific survival and overall survival following radical cystectomy for bladder cancer compared to logistic regressions models.

While the accuracy of ANN is higher compared to logistic regression it is unclear whether these results can influence follow-up diagnostic and therapeutic investigations.

Presented by Alexander Buchner, MD, et al. at the 26th Annual European Association of Urology (EAU) Congress - March 18 - 21, 2011 - Austria Centre Vienna, Vienna, Austria

Reported for UroToday by Christian Doehn, MD, PhD, Department of Urology, University of Lübeck Medical School, Lübeck Germany.

The opinions expressed in this article are those of the UroToday.com Contributing Medical Editor and do not necessarily reflect the viewpoints of the European Association of Urology (EAU)

View EAU 2011 Annual Meeting Coverage

Login