Systematic Review: Machine Learning and Deep Learning based Prostate Cancer Prediction

Abstract

The purpose of this study focuses on using various computer models and imaging systems to diagnose prostate cancer, a major cause of cancer-related deaths globally. These models analyze patient samples using tools and advanced algorithms like DL (Deep Learning) to spot tumors and predict critical symptoms using methods such as AH or UNet. AI models like RNN and CNN help predict and detect prostate cancer, reducing risks. By employing SOM (Self-Organizing Maps) based on deep learning, they enhanced accuracy in disease detection, extracting parameters from CT scans for better treatment. Using DL and ML for prediction and classification, they observed improvements in computer-aided diagnosis. Techniques like KNN or SVM, along with multitask learning, helped in therapy reporting and optimizing prostate cancer assessments. AI-based clinical tools and technologies improved patient outcomes, utilizing biopsies and MRI scans for disease detection. The study explores various AI models such as Machine Learning and Deep Learning (like RNN, CNN, KNN, SVM, random forest, logistics regression) for detecting, predicting, diagnosing, and classifying prostate cancer. These models have used publicly available datasets from different websites, demonstrating their high performance in improving the treatment of prostate cancer.