Review Article
Noninvasive Assessment of Liver Fibrosis
Laurent Castera, M.D., Ph.D., Mary E. Rinella, M.D., and Emmanuel A. Tsochatzis, M.D., Ph.D.
A figure showing the diagnosis, staging, and progression of liver fibrosis.
The NEJM identity sits at the bottom.
The prognosis of liver disease depends on the extent of fibrosis, which is usually staged with the use of liver biopsy. The limitations of biopsy have led to the development of noninvasive tests, which the authors review. Learn more: nej.md/4hwZMgj
#MedSky #GastroSky
29.10.2025 22:00
π 16
π 5
π¬ 0
π 0
HOTSPoT paves the way for scalable, explainable, and accessible AI in liver pathology.
Open science, clinical relevance, and technical excellence β all in one model.
Congrats to the entire team!
#DigitalPathology #OpenSource #AIinMedicine
21.07.2025 07:13
π 1
π 1
π¬ 0
π 0
Behind this innovation is a multidisciplinary, international team β but a special shoutout to Giorgio Cazzaniga, who led dataset creation, annotation, model design & implementation. A stellar example of clinician-engineering leadership. π
21.07.2025 07:13
π 0
π 0
π¬ 1
π 0
Clinical validation? Yes. In 35 liver biopsies, HOTSPoT's automated tract count matched human observers (ΞΊ = 0.90) and correlated strongly with fibrosis stage (r = 0.87). A true step forward in reproducible pathology. ππ©Ί
21.07.2025 07:13
π 0
π 0
π¬ 1
π 0
GitHub - Gizmopath/HOTSPoT: Hematoxylin & Eosin-based Open-access Tool for Segmentation of Portal Tracts
Hematoxylin & Eosin-based Open-access Tool for Segmentation of Portal Tracts - Gizmopath/HOTSPoT
HOTSPoT isn't just a model β it's a tool. It's released in TorchScript format, integrates with QuPath, and includes a custom WSInfer pipeline for WSI-level inference. βοΈ
Code & model π github.com/Gizmopath/HO...
21.07.2025 07:13
π 0
π 0
π¬ 1
π 0
Trained on 223 cases and tested across 5 international centers, HOTSPoT achieved Dice scores up to 0.92, showing minimal domain shift. That means: robust performance across scanners, stains, and disease contexts. ππ
21.07.2025 07:13
π 0
π 0
π¬ 1
π 0
Liver histology is essential in diagnosing autoimmune and inflammatory diseases β but manual annotation is time-consuming and variable. HOTSPoT tackles this with automated, high-performance segmentation of portal tracts in H&E-stained slides. π§ π§«
21.07.2025 07:13
π 0
π 0
π¬ 1
π 0
Automating liver biopsy segmentation with a robust, open-source tool for pathology research: the HOTSPoT model
npj Digital Medicine - Automating liver biopsy segmentation with a robust, open-source tool for pathology research: the HOTSPoT model
π₯ HOTSPoT sets a new benchmark in liver pathology: open-source, accurate, and externally validated tool for portal tract segmentation. A brilliant team effort led by Giorgio Cazzaniga β whose vision and drive made it possible π
rdcu.be/ew7qh #AI #Digitalpathology
21.07.2025 07:08
π 2
π 1
π¬ 1
π 0
Great visit to EKFZ Digital Health Institute! Gave a talk at Jakob Katherβs lab, had great 1:1s with team members, and visited our PhD student Elisa Merelli, who's spending part of her PhD hereπ‘π¬ #DigitalHealth #AI #Research
@jnkt.bsky.social @janclusmann.bsky.social
10.06.2025 15:32
π 7
π 1
π¬ 0
π 1
People Overtrust AI-Generated Medical Advice despite Low Accuracy
This article presents a comprehensive analysis of how artificial intelligence (AI)βgenerated medical responses are perceived and evaluated by nonexperts. We conducted a study in which a total of 30...
This article analyzes how nonexperts perceive AI-generated medical advice, finding that it is often rated as accurate as β or better than β doctor responses, raising concerns about overreliance and potential harm from incorrect guidance. Full article: nejm.ai/43xKDG0
#AI #MedSky #MLSky
22.05.2025 13:17
π 9
π 7
π¬ 0
π 1
A graph showing the time to first flare of giant-cell arteritis through week 52.
In the SELECT-GCA phase 3 trial involving patients with giant-cell arteritis, the oral Janus kinase inhibitor upadacitinib (15 mg) significantly improved remission of disease, with less glucocorticoid use. Full trial results: nej.md/3XBdkia
#MedSky
03.04.2025 15:02
π 12
π 4
π¬ 0
π 1
Happy to share our latest paper in @ai.nejm.org, led by Isabella Wiest: "Deidentifying Medical Documents with Local, Privacy-Preserving Large Language Models: The LLM-Anonymizer" ai.nejm.org/doi/full/10....
You can use our open source tool with local LLMs to robustly de-identify medical documents.
27.03.2025 18:32
π 9
π 5
π¬ 1
π 0