Differentiation of Benign Vs. Malignant Colorectal Polyps Using Machine Learning Techniques: A Systematic Review and Meta-Analysis

Colorectal polyps are critical markers for colorectal cancer (CRC), and distinguishing between benign and malignant polyps is essential for early diagnosis and treatment. This systematic review and meta-analysis assess the performance of machine learning techniques in accurately differentiating benign from malignant colorectal polyps. The review examines recent research that integrates advanced technologies, including computer-aided diagnosis (CADx) systems and deep learning models, particularly convolutional neural networks (CNNs). These machine learning techniques show great promise in analyzing histopathological images, identifying subtle patterns and characteristics that may be overlooked by traditional methods. Additionally, the integration of optical coherence tomography (OCT) with CADx systems enhances tissue structure interpretation, leading to more precise diagnoses. By synthesizing data from multiple studies, this review highlights the significant advancements in CRC diagnosis and the potential for improved patient outcomes through the application of machine learning techniques.