Classifying Disorders of Prostate using Master-Slave Configuration of Backpropagation Neural Networks

Anilkumar Kothalil Gopalakrishnan


Backpropagation neural network, benign prostatic hyperplasia, prostate cancer


This paper classifies disorders of prostate into Benign Prostatic Hyperplasia (BPH), prostate cancer (PC) and other inflammations using master-slave configuration of two Back propagation Neural Networks (BPNNs). Symptoms from elderly men including urinary dribbling, urinary hesitancy, feeling of non-empty bladder, burning urination, hematuria, testosterone level and others are given as the attributes of the master BPNN. The slave BPNN has the following attributes; output of the master BPNN (symptom level), free prostate specific antigen (PSA) level, insulin-like growth factor-1 (IGF-1), CAG (DNA glutamate) repeat, heredity, PCA3 urine level, JM-27 blood level and others. The slave BPNN classifies prostate disorders into PC, BPH and other inflammations. Data from Gleason score is used to detect PC stages. The simulation shows that the proposed procedure could be effective in classifying disorders of the prostate gland in men.

Important Links:

Go Back