Portal de Periódicos da CAPES

Pathology and clinical practice

2023; Volume: 7; Issue: S1 Linguagem: Inglês

10.53730/ijhs.v7ns1.15110

2550-6978

Sulaiman Sleem Alatawi, Ali Moharag Hadadi, Munirah Mohammed Almulhim, Maryam Mousa Ahmed Almousa, Alkhathami Alkhathami, Ahmed K Adel, ‏Bakr Mansour Alqahtani, Almuhaysh Almuhaysh, Ahmed K Maryam, Jawaher Sadun Alsadun, Mazen Ibrahim Mohammed Otaif, Lujain Yousef Almulhim, Abdullah Mohammed Alanazi,

Cancer Genomics and Diagnostics

Background: The advent of molecular biomarkers has revolutionized cancer diagnosis and treatment, enhancing the precision of therapeutic strategies for solid tumors. However, the complexity of clinical decision-making has escalated with the increasing number of prognostic and predictive biomarkers. The integration of deep learning (DL) in histology image analysis promises to streamline these processes. Aim: This review aims to evaluate the latest diagnostic techniques and tools in cancer diagnosis, focusing on the role of molecular biomarkers and deep learning in enhancing clinical outcomes. Methods: A comprehensive review of recent studies and clinical trials was conducted, examining the impact of molecular biomarkers on cancer treatment and the application of DL in histology image analysis. The review covered fundamental DL applications in tumor identification, grading, subtyping, and advanced applications in predicting genetic mutations, treatment responses, and survival outcomes. Results: DL-based methods have shown high accuracy in automating histopathology workflows, matching or surpassing human performance in tumor detection and classification. Advanced DL applications offer new insights by predicting genetic alterations and clinical outcomes directly from histology images, which could significantly impact clinical decision-making.