Reproducibility and explainability in digital pathology: The need to make black-box artificial intelligence systems more transparent DOI Creative Commons
Gavino Faa, Matteo Fraschini, Luigi Barberini

et al.

Deleted Journal, Journal Year: 2024, Volume and Issue: 13(4)

Published: Oct. 1, 2024

Artificial intelligence (AI), and more specifically Machine Learning (ML) Deep learning (DL), has permeated the digital pathology field in recent years, with many algorithms successfully applied as new advanced tools to analyze pathological tissues. The introduction of high-resolution scanners histopathology services represented a real revolution for pathologists, allowing analysis whole-slide images (WSI) on screen without microscope at hand. However, it means transition from absence specific training most pathologists involved clinical practice. WSI approach represents major transformation, even computational point view. multiple ML DL developed may enhance diagnostic process fields human pathology. AI-driven models allow achievement consistent results, providing valid support detecting, H&E-stained sections, biomarkers, including microsatellite instability, that are missed by expert pathologists.

Language: Английский

Colorectal cancer: Biology and pathology DOI
Gavino Faa, Andrea Pretta, Matteo Fraschini

et al.

Elsevier eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 3 - 15

Published: Jan. 1, 2025

Language: Английский

Citations

0
Artificial Intelligence Models for the Detection of Microsatellite Instability from Whole-Slide Imaging of Colorectal Cancer DOI Creative Commons
Gavino Faa,

Ferdinando Coghe,

Andrea Pretta

et al.

Diagnostics, Journal Year: 2024, Volume and Issue: 14(15), P. 1605 - 1605

Published: July 25, 2024

With the advent of whole-slide imaging (WSI), a technology that can digitally scan whole slides in high resolution, pathology is undergoing digital revolution. Detecting microsatellite instability (MSI) colorectal cancer crucial for proper treatment, as it identifies patients responsible immunotherapy. Even though universal testing MSI recommended, particularly affected by (CRC), many remain untested, and they reside mainly low-income countries. A critical need exists accessible, low-cost tools to perform pre-screening. Here, potential predictive role most relevant artificial intelligence-driven models predicting directly from histology alone discussed, focusing on CRC. The deep learning (DL) identifying status here analyzed studies reporting development algorithms trained this end. important performance deficiencies are discussed every AI method. proposed algorithm sharing among multiple research clinical centers, including federal (FL) swarm (SL), reported. According all reported here, valuable WSI use digitized H&E-stained sections allow extraction molecular information, such status, short time at low cost. possible advantages related introducing DL methods routine surgical underlined acceleration transformation departments services recommended.

Language: Английский

Citations

1
“Artificial histology” in colonic Neoplasia: A critical approach DOI
Gavino Faa, Matteo Fraschini, Luca Didaci

et al.

Digestive and Liver Disease, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 1, 2024

Language: Английский

Citations

1
An Unsupervised Learning Tool for Plaque Tissue Characterization in Histopathological Images DOI Creative Commons
Matteo Fraschini, Massimo Castagnola, Luigi Barberini

et al.

Sensors, Journal Year: 2024, Volume and Issue: 24(16), P. 5383 - 5383

Published: Aug. 20, 2024

Stroke is the second leading cause of death and a major disability around world, development atherosclerotic plaques in carotid arteries generally considered severe cerebrovascular events. In recent years, new reports have reinforced role an accurate histopathological analysis to perform stratification affected patients proceed correct prevention complications. This work proposes applying unsupervised learning approach analyze complex whole-slide images (WSIs) allow simple fast examination their most relevant features. All code developed for present freely available. The proposed method offers qualitative quantitative tools assist pathologists examining complexity more effectively. Nevertheless, future studies using supervised methods should provide evidence correspondence between clusters estimated textural-based regions manually annotated by expert pathologists.

Language: Английский

Citations

0
Reproducibility and explainability in digital pathology: The need to make black-box artificial intelligence systems more transparent DOI Creative Commons
Gavino Faa, Matteo Fraschini, Luigi Barberini

et al.

Deleted Journal, Journal Year: 2024, Volume and Issue: 13(4)

Published: Oct. 1, 2024

Artificial intelligence (AI), and more specifically Machine Learning (ML) Deep learning (DL), has permeated the digital pathology field in recent years, with many algorithms successfully applied as new advanced tools to analyze pathological tissues. The introduction of high-resolution scanners histopathology services represented a real revolution for pathologists, allowing analysis whole-slide images (WSI) on screen without microscope at hand. However, it means transition from absence specific training most pathologists involved clinical practice. WSI approach represents major transformation, even computational point view. multiple ML DL developed may enhance diagnostic process fields human pathology. AI-driven models allow achievement consistent results, providing valid support detecting, H&E-stained sections, biomarkers, including microsatellite instability, that are missed by expert pathologists.

Language: Английский

Citations

0