Deep Learning Approaches in Histopathology

Area of Science:

• Computational pathology
• Medical artificial intelligence (AI)
• Machine learning (ML) and deep learning (DL) in oncology

Background:

• AI, ML, and DL are increasingly impacting daily life and scientific research.
• These technologies are being developed for various tumor pathology tasks, including detection, classification, grading, and genomic analysis.
• Pathologists seek AI to enhance diagnostic precision, reduce workload, and minimize time constraints.

Purpose of the Study:

• To review the implementation of ML and DL methods in healthcare settings for tumor morphology analysis.
• To identify current obstacles and future opportunities for AI application in pathology.

Main Methods:

• Literature review of AI, ML, and DL applications in tumor pathology.
• Survey of existing algorithms and computational technologies for pathological tasks.
• Analysis of challenges related to algorithm validation, training, and acceptance.

Main Results:

• AI algorithms are designed for diverse tasks in tumor pathology, from detection to genomic mutation analysis.
• Potential benefits include improved diagnostic accuracy and efficiency for pathologists.
• Significant obstacles remain, including algorithm validation, computational infrastructure, and user training/acceptance.

Conclusions:

• ML and DL offer significant potential to revolutionize tumor morphology analysis and improve patient care.
• Overcoming implementation barriers is crucial for successful integration of AI into routine pathology workflows.
• Further research and development are needed to fully realize the opportunities of AI in diagnostic pathology.