General Pathologists Achieve Near-Specialist Diagnostic Performance Using Deep Learning-Based Virtual Staining for...

Area of Science:

• Nephrology
• Digital Pathology
• Artificial Intelligence

Background:

• Hematoxylin and eosin (H&E) staining inadequately visualizes collagen fibers, hindering interstitial fibrosis assessment in donor kidneys.
• General pathologists often lack specialized renal training, leading to challenges in evaluating chronic kidney pathologies.

Purpose of the Study:

• To investigate the efficacy of artificial intelligence-based virtual staining in enhancing donor kidney evaluation, specifically for interstitial fibrosis and chronic pathologies.
• To determine if virtual staining can improve diagnostic accuracy and inter-observer agreement among pathologists with varying expertise levels.

Main Methods:

• Developed and validated a CycleGAN-based artificial intelligence model to transform H&E images into virtual Masson's trichrome (VMT) representations.
• Assessed interstitial fibrosis and chronic changes using the Remuzzi scoring system with and without VMT, comparing evaluations by a renal pathologist and general pathologists.
• Conducted prospective validation on frozen sections to confirm the model's performance.

Main Results:

• Virtual Masson's trichrome (VMT) effectively visualized interstitial collagen fibers, enabling more reliable fibrosis classification.
• Diagnostic accuracy significantly improved with VMT, markedly increasing inter-observer agreement for both general and renal pathologists.
• General pathologists achieved near-specialist performance in fibrosis assessment using VMT, bridging the expertise gap.

Conclusions:

• Deep learning-based virtual staining significantly enhances the precision of donor kidney evaluations by general pathologists, approaching specialist-level performance.
• This technology offers an efficient, cost-effective solution for assessing fibrosis and chronic pathologies in transplantation medicine.
• Virtual staining has the potential to eliminate diagnostic disparities between general and specialist pathologists.