Abstract
Prostate cancer is a commonly diagnosed cancerous disease among menworld-wide. Even with modern technology such as multi-parametric magneticresonance tomography and guided biopsies, the process for diagnosing prostatecancer remains time consuming and requires highly trained professionals. Inthis paper, different convolutional neural networks (CNN) are evaluated ontheir abilities to reliably classify whether an MRI sequence contains malignantlesions. Implementations of a ResNet, a ConvNet and a ConvNeXt for 3D imagedata are trained and evaluated. The models are trained using different dataaugmentation techniques, learning rates, and optimizers. The data is taken froma private dataset, provided by Cantonal Hospital Aarau. The best result wasachieved by a ResNet3D, yielding an average precision score of 0.4583 and AUCROC score of 0.6214.