top of page
  • Writer's picturePathy

Quantifying Fibrosis with AI: Towards Faster and More Objective Diagnosis

Updated: Apr 18, 2023

What is Fibrosis and why is it important to diagnose it accurately?

  • Fibrosis is a scarring process in tissue caused by the accumulation of collagen proteins and is a common feature of many chronic diseases.

  • Accurate and timely diagnosis of fibrosis is crucial for the effective treatment and management of these diseases.

What are the traditional methods for diagnosing fibrosis?

  • Traditionally, fibrosis is diagnosed based on the visual examination of tissue samples by pathologists, who grade the severity of fibrosis using subjective scales such as the Ishak or METAVIR scoring systems.

  • To have an accurate quantification of fibrosis, a special stain is used: Trichrome staining which is a widely used technique in pathology that allows visualization of collagen fibers in blue, muscle fibers in red, and cytoplasm in yellow-green.

What are the challenges with traditional methods of diagnosing fibrosis?

  • The traditional approach is time-consuming, prone to inter-observer variability, and can miss subtle changes in fibrosis progression.

  • There is a growing demand for more quantitative and standardized methods for fibrosis diagnosis, especially in the context of clinical trials and drug development.

What is the solution being developed for diagnosing fibrosis?

  • Our AI-based model can automatically quantify fibrosis in trichrome-stained histopathology slides of the Kidney.

  • The model is trained on a dataset of kidney tissue slides annotated in collaboration with an 8-year fibrosis researcher and a pathologist.

  • The model aims to enable users to select a region of interest in the slide and get a quantification of the percentage of low, medium, and high fibrosis in that region.

What are the benefits of the proposed solution?

  • The model reduces the subjectivity and variability of fibrosis diagnosis and provides faster and more objective results for clinicians and researchers.

  • The model provides a quantitative score that can be easily converted into clinically validated scores, such as the Banff Lesion Score, a widely accepted scoring system for kidney fibrosis.

What is the vision for expanding the solution?

  • The vision is to extend the model to other organs, such as the lung and liver, and potentially use other staining techniques, such as H&E staining.

  • By building a platform with an extensive model library for pathology, the goal is to provide a comprehensive and scalable solution for automated diagnosis and quantification of fibrosis and other pathological features across different organs and staining techniques.

What is the potential impact of this solution?

  • The model has the potential to revolutionize the way fibrosis is diagnosed and managed by providing a faster, more accurate, and more objective method that can improve patient outcomes and accelerate drug development.

  • The platform, which we plan to grow even further, can be a game-changer in the field of pathology and has the potential to bring innovative solutions to research labs and hospitals looking to advance the field of pathology.

89 views0 comments


bottom of page