A classification model for the prostate cancer based on deep learning

Regarded as one of the common cancers, the prostate cancer is a main reason harming the health of senile men, especially in Europe and the United States. In China, with increasing of living standards and aging of populations, the incidence of prostate cancer has an upward and younger tendency. Early detection and early treatment are helpful to patients. In the imaging diagnosis methods of prostate, magnetic resonance imaging (MRI) has been recognized as the most effective way. Whereas MRI image has several specialized configurations with a lot of medical information, and diagnosis results have a strong relationship with doctor's professional skill and experience, which makes the diagnosis for prostate cancer more difficult. Based on deep learning and the Convolutional Neural Networks (CNN), an image classification model which can provide some diagnosis classification reference was proposed in this paper. The data sets used in this paper consisted of 10056 diffusion weighted magnetic resonance imaging (DWI) images. Three quarters of the images were used for training and the rest images for testing. Experiments show that the accuracy rate of training set is 80.1539%, and the accuracy rate of testing set is 78.1538%. The curves of testing accuracy rate and loss show that this model has been trained steadily. The accuracy rate for single images are above 64.91%, and some may reach 99.99%. With certain clinical application value, this deep learning method can be widely applied to the grading and staging of prostate cancer and other cancer tasks.