This Article Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Catto, J. W. F. Articles by Hamdy, F. C. Search for Related Content PubMed PubMed Citation Articles by Catto, J. W. F. Articles by Hamdy, F. C.

Artificial Intelligence in Predicting Bladder Cancer Outcome

A Comparison of Neuro-Fuzzy Modeling and Artificial Neural Networks¹

The Academic Urology Unit [J. W. F. C., F. C. H.] and the Academic Pathology Unit [J. L. B.], Department of Automatic Control and Systems Engineering [D. A. L., M. F. A., M. C.], University of Sheffield, and the Department of Pathology, Royal Hallamshire Hospital [K. M. F.], Sheffield S10 2JF, United Kingdom

Purpose: New techniques for the prediction of tumor behavior are needed, because statistical analysis has a poor accuracy and is not applicable to the individual. Artificial intelligence (AI) may provide these suitable methods. Whereas artificial neural networks (ANN), the best-studied form of AI, have been used successfully, its’ hidden networks remain an obstacle to its acceptance. Neuro-fuzzy modeling (NFM), another AI method, has a transparent functional layer and is without many of the drawbacks of ANN. We have compared the predictive accuracies of NFM, ANN, and traditional statistical methods, for the behavior of bladder cancer.

Experimental Design: Experimental molecular biomarkers, including p53 and the mismatch repair proteins, and conventional clinicopathological data were studied in a cohort of 109 patients with bladder cancer. For all three of the methods, models were produced to predict the presence and timing of a tumor relapse.

Results: Both methods of AI predicted relapse with an accuracy ranging from 88% to 95%. This was superior to statistical methods (71–77%; P < 0.0006). NFM appeared better than ANN at predicting the timing of relapse (P = 0.073).

Conclusions: The use of AI can accurately predict cancer behavior. NFM has a similar or superior predictive accuracy to ANN. However, unlike the impenetrable "black-box" of a neural network, the rules of NFM are transparent, enabling validation from clinical knowledge and the manipulation of input variables to allow exploratory predictions. This technique could be used widely in a variety of areas of medicine.

This article has been cited by other articles:

				J. W.F. Catto, M. F. Abbod, D. A. Linkens, S. Larre, D. J. Rosario, and F. C. Hamdy Neurofuzzy Modeling to Determine Recurrence Risk Following Radical Cystectomy for Nonmetastatic Urothelial Carcinoma of the Bladder Clin. Cancer Res., May 1, 2009; 15(9): 3150 - 3155. [Abstract] [Full Text] [PDF]

				S. F. Shariat, P. I. Karakiewicz, G. Godoy, and S. P. Lerner Review: Use of nomograms for predictions of outcome in patients with advanced bladder cancer Therapeutic Advances in Urology, April 1, 2009; 1(1): 13 - 26. [Abstract] [PDF]

				M. Koc, H. Dizen, S. Ocak, and M. Tez Scoring Systems: Are They Ideal? Ann. Surg. Oncol., November 1, 2008; 15(11): 3316 - 3316. [Full Text] [PDF]

				J. Cho, W. Han, and D.-Y. Noh What Is the Best? Ann. Surg. Oncol., November 1, 2008; 15(11): 3317 - 3317. [Full Text] [PDF]

				D. R. Yates, I. Rehman, M. F. Abbod, M. Meuth, S. S. Cross, D. A. Linkens, F. C. Hamdy, and J. W.F. Catto Promoter Hypermethylation Identifies Progression Risk in Bladder Cancer Clin. Cancer Res., April 1, 2007; 13(7): 2046 - 2053. [Abstract] [Full Text] [PDF]