Analysing DCE-MRI scans using hybrid techniques for early detection of prostate cancer based on fusion features of handcrafted and deep learning features

Prostate cancer is among the most diagnosed malignancies in men worldwide and a leading cause of cancer-related mortality. Early and accurate diagnosis is critical to improve patient outcomes and reduce the risks of overtreatment or missed detection. Conventional diagnostic approaches, including prostate-specific antigen (PSA) testing, digital rectal examination (DRE), and histopathological analysis, often suffer from limited sensitivity and specificity, leading to false positive or delayed diagnosis. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) has recently emerged as an effective modality for prostate cancer detection, providing complementary anatomical and functional information. This study proposes a novel hybrid diagnostic framework that integrates Generalized Quantum Gamma Polynomial (GQGP) features, kinetic signal intensity features, and deep learning-based representations. GQGP features capture subtle intensity variations and quantum-inspired statistical characteristics, while kinetic features quantify contrast-enhancement dynamics to discriminate malignant from benign tissues. These handcrafted descriptors are fused with high-level features extracted using convolutional neural networks (CNNs) to construct a comprehensive feature representation. Experimental evaluation on publicly available prostate imaging datasets demonstrates that the proposed fusion framework significantly outperforms single-feature and traditional methods, achieving a classification accuracy of 97.32%. The results highlight the effectiveness of combining mathematical modeling, radiomics, and artificial intelligence for improved prostate cancer diagnosis.