Saratov JOURNAL of Medical and Scientific Research

Prognosis of prostate gland morphology study using artificial neural network

Year: 2014, volume 10 Issue: №2 Pages: 328-332
Heading: Proceedings of all-Russia week of science with international participants Article type: Original article
Authors: Popkov V.M., Shatylko T.V., Fomkin R.N.
Organization: Saratov State Medical University

The research goal is to optimize the management of patients with serum PSA level falling in the range of 4-10 ng/ ml by designing and educating of an artificial neural network, which may be used to predict prostate gland morphology basing on clinical, laboratory and imaging data. Material and methods. Data of 254 patients, who were admitted to the oncological Department of S. R. Mirotvortsev Clinical hospital for transrectal prostate biopsy, was collected to construct several artificial neural networks with different architecture. External validation was performed on 27 patients, who had prostate biopsy in January-February 2014. Results. One-layer network, consisting of 11 input, 9 hidden and 3 output neurons, was determined to be the most successful: in 92.6% cases it was correct in predicting prostate cancer or its absence. Input factors were evaluated according to their relative importance, from more important to less important: prostate volume, serum PSA, patient's age, prostate consistency, PSA velocity, prostate symmetry, previous negative biopsy, free serum PSA, intake of 5-alpha-reductase inhibitors. Conclusion. Artificial neural networks may be used to predict morphological findings in prostate biopsy. High PSA density and firm prostate consistency should cause suspicion of prostate cancer.

1. Murphy DG, Ahlering T, Catalona WJ, et al. The Melbourne Consensus Statement on the early detection of prostate cancer. BJU lnt2014Feb; 113(2): 186-8
2. Ecke TH, Bartel P, Hallmann S, et al.Outcome prediction for prostate cancer detection rate with artificial neural network (ANN) in daily routine. Urol Oncol 2012 Mar-Apr; 30 (2): 139-44
3. Ecke TH, Hallmann S, Koch S, et al. External validation of an artificial neural network and two nomograms for prostate cancer detection. ISRN Urol 2012; 2012: 643181
4. Meijer RP, Gemen EF, van Onna IE, et al. The value of an artificial neural network in the decision-making for prostate biopsies. World J Urol 2009 Oct; 27 (5): 593-8.

2014_02_328-332.pdf1005.39 KB

No votes yet