Reliability, Validity, and Reader Acceptance of LI-RADS-An In-depth Analysis

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Hepatocellular carcinoma (HCC) diagnosis relies heavily on imaging.
• Standardized reporting systems like LI-RADS aim to improve diagnostic consistency.
• Nonstandardized scales are also used, but their agreement and accuracy vary.

Purpose of the Study:

• To compare the interreader agreement and diagnostic accuracy of LI-RADS with a nonstandardized 5-point Likert scale.
• To assess reader acceptance of LI-RADS for routine clinical use in HCC detection.
• To analyze the performance of specific LI-RADS features.

Main Methods:

• Retrospective analysis of 84 patients with HCC risk undergoing liver MRI.
• Four readers assessed 104 liver observations using LI-RADS and a Likert scale.
• Interreader agreement assessed via kappa statistics; diagnostic accuracy via ROC analysis.
• Reader acceptance evaluated through a questionnaire.

Main Results:

• LI-RADS showed similar overall interreader agreement (κ=0.44) compared to the Likert scale (κ=0.35), with a trend towards higher agreement for LI-RADS.
• Diagnostic accuracy was comparable between LI-RADS (AUC=0.87) and the Likert scale (AUC=0.86).
• Readers indicated a short version of LI-RADS would enhance clinical routine use.

Conclusions:

• LI-RADS offers comparable diagnostic accuracy and interreader agreement to nonstandardized scales for HCC.
• Further simplification and feature refinement of LI-RADS may be beneficial for clinical implementation.
• LI-RADS shows potential for standardized HCC reporting, but optimization is suggested.