Prostate cancer is one of the most common types of cancer among men worldwide, and early detection is crucial for effective treatment. The involvement of artificial intelligence in the detection process represents a significant advancement in medical diagnostics. A standard method to train a neural network for this task requires time-intensive, detailed pixel-level annotations by skilled pathologists, which are time-consuming and costly. In contrast, the global information, whether the whole histopathological image (slide) contains cancer, is readily available from the clinical practice. The thesis aims to train and evaluate a convolutional neural network for predicting prostate cancer using the slide-level annotations of clinical data provided by Masaryk Memorial Cancer Institute. The student will familiarise themself and present the basics of digital pathology. They will select suitable architecture and perform and asses a comparative test in the cloud infrastructure of the RationAI group. The student will optionally contribute the code to HistoPipe, the RationAI open-source library for digital histopathology.