Automated patient-tailored screening of the liver for diffuse steatosis and iron overload using MRI.

Journal Article (Journal Article)

OBJECTIVE: The purpose of this article is to validate an automated screening method for evaluation of hepatic steatosis or siderosis. MATERIALS AND METHODS: This was a two-part study, with retrospective and prospective portions. First, 130 consecutive abdominal MRI examinations, including both the automated algorithm and reference standard fat and iron quantification, were retrospectively identified. The algorithm's performance was validated against the reference standard and was compared with the performance of three expert readers. Subsequently, 39 subjects undergoing liver MRI were prospectively identified and enrolled. These subjects were scanned with a protocol where quantification sequences were either performed or not performed on the basis of the recommendation of the algorithm. Total examination time in these subjects was compared with examination times in the 90 subjects from the retrospective cohort who had undergone a similar liver MRI protocol with complete quantification. RESULTS: The automated algorithm was accurate in determining the presence of deposition disease (93.1%), with no significant difference between its conclusions and those of any of the readers (p=0.48-1.0). Use of the algorithm resulted in a small but statistically significant time savings compared with performing quantification in all subjects (28 minutes 56 seconds vs 31 minutes 20 seconds; p<0.05). CONCLUSION: Automated screening for hepatic steatosis and siderosis can be performed in real time during abdominal MRI examinations, can save total scan time compared with always performing quantification, and could serve as a gatekeeper for dedicated quantification sequences.

Full Text

Duke Authors

Cited Authors

  • Bashir, MR; Zhong, X; Dale, BM; Gupta, RT; Boll, DT; Merkle, EM

Published Date

  • September 2013

Published In

Volume / Issue

  • 201 / 3

Start / End Page

  • 583 - 588

PubMed ID

  • 23971450

Electronic International Standard Serial Number (EISSN)

  • 1546-3141

Digital Object Identifier (DOI)

  • 10.2214/AJR.12.10051


  • eng

Conference Location

  • United States