Post

New research · Radiology
Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association · 1d
AI / informaticsClinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association · 2026

Automated deep learning detection of hepatic steatosis on non-contrast CT scans and discrepancy with scan reports.

David Yardeni, Jianfei Liu, Pritam Mukherjee … Yaron Rotman
Read paper
RadiologyAI / informatics

Only a minority of incidental hepatic steatosis cases are reported on non-contrast CT.

Automated deep learning detection of hepatic steatosis on non-contrast CT scans and discrepancy with scan reports.

David Yardeni et al. · Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association · 2026
Background

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a major cause of liver disease that is growing in prevalence.

Purpose

We hypothesized that incidental finding of steatosis on NCCT is often not reported if not specifically sought in the imaging request.

Methods

A retrospective cross-sectional single-center analysis of abdominal NCCT scans performed between 2012-20 for any indication in adult subjects.

n = 2,710 adult
32.7%
Results
Radiologists reported hepatic steatosis in only 32.7% of scans where AI detected it.
n = 2,710 adult
More results

The mean liver attenuation derived from the deep-learning algorithm was 50.4 ± 11.8 HU.

Image-based steatosis was found in 480 (13.1%) scans, with a mean liver attenuation of 29.8 ± 14 HU.

More results

Predictors of unreported steatosis included higher average liver attenuation (even if <40 HU), high variability of fat distribution in the liver and low BMI.

“
Conclusion

We found that incidental hepatic steatosis in non-contrast CT is reported in a minority of scans. Incorporating artificial intelligence-based hepatic attenuation measurement in CT scan reading may increase reporting rates.

Read paper
0 comments

No comments yet. Be the first.

Related papers

LatestFoundational
Review
10-60 mGy
the range of radiation exposure to the baby from belly or pelvis CT scans
Cohort Study
16 min 08 s
PACScrawler was faster at getting medical images than manual methods
Cohort Study
6.7%
Among hereditary hemorrhagic telangiectasia patients re-imaged, 6.7% had clinically significant liver arteriovenous malformations.
Guideline
“In those instances where peer reviewed literature is lacking or equivocal, experts may be the primary evidentiary source available to formulate a recommendation.”
Cross-sectional
56.5%
More than half of computed tomography reports did not mention tumor deposits.
Study
5.24 million
the number of brain scans used to train the new artificial intelligence model
Cohort Study
17.70%
fewer than one in five senior authors in medical imaging research are women
AI / Informatics
OR = 6.15
Large language models were much more likely to answer easier questions correctly than harder ones