Computer-Aided Prostate Cancer Diagnosis From Digitized Histopathology: A Review on Texture-Based Systems

Area of Science:

• Digital pathology
• Medical image analysis
• Computational oncology

Background:

• Prostate cancer diagnosis relies on subjective microscopic evaluation of biopsy samples.
• Pathologist expertise and interpretation variability lead to inconsistent diagnostic outcomes.
• Computer-aided diagnosis (CAD) systems aim to enhance accuracy and reproducibility in prostate cancer detection and grading.

Purpose of the Study:

• To review existing literature on CAD systems for prostate cancer (PCa) detection and grading.
• To emphasize the role of texture analysis in describing prostatic tissue characteristics.
• To provide an overview of image analysis techniques relevant to PCa CAD development.

Main Methods:

• Review of selected literature focusing on CAD systems for PCa.
• Emphasis on texture analysis for feature extraction and classification.
• Discussion of image preprocessing, feature extraction, classification, and validation techniques.

Main Results:

• Numerous CAD systems have been proposed for quantitative image analysis and classification of prostatic carcinoma.
• Texture analysis methods are crucial for detailed tissue description in PCa histopathology.
• The complexity of prostatic tissue and large data volumes present challenges in CAD development.

Conclusions:

• CAD systems, particularly those utilizing texture analysis, hold significant potential to improve the accuracy and reproducibility of prostate cancer diagnosis.
• Further research is needed to overcome challenges and develop more robust texture-based CAD systems.
• Advancements in CAD can reduce pathologist reading time and improve diagnostic consistency.