AJSP2404 - CME/CMLE - CME/CMLE - Artificial intelligence–based algorithms for the diagnosis of prostate cancer-Artificial intelligence–based algorithms for the diagnosis of prostate cancer

Artificial intelligence–based algorithms for the diagnosis of prostate cancer

Pdf Summary

The article discusses the application of artificial intelligence (AI) in aiding the diagnosis and grading of prostate cancer using digital pathology methods. Prostate cancer is a prevalent form of cancer that impacts pathology laboratories' workflows significantly. The use of AI algorithms, particularly convolutional neural networks (CNNs), has shown promising results in detecting and grading prostate cancer based on histologic features. Studies have indicated that AI can enhance the efficiency and accuracy of cancer detection, resulting in reduced turnaround times and improved prognostic predictions.<br /><br />Various AI-based algorithms have been developed and tested, showing good to excellent results in cancer detection, grading, and identification of prognostic parameters like extraprostatic extension and biochemical recurrence. The use of whole-slide imaging (WSI) combined with AI tools has revolutionized prostate pathology by providing pathologists with advanced tools for improved diagnosis. AI algorithms have also shown potential in identifying specific histological features such as cribriform patterns and perineural invasion, which are crucial in determining prognosis.<br /><br />The review highlights the importance of further technological advancements to broaden the scope of prostate cancer variants recognized by AI algorithms and reduce the reliance on annotated datasets for training. Despite some limitations, such as challenges in classifying rare variants and the need for external validation, the integration of AI in prostate pathology holds great promise for enhancing diagnostic accuracy and prognostic potential. Further developments in this field are necessary to optimize AI tools for routine clinical practice.

Keywords

Artificial intelligence

Prostate cancer

Digital pathology

Convolutional neural networks

Histologic features

Cancer detection

Grading

Whole-slide imaging

Prognostic predictions

Pathology laboratories