Agreement Between Criteria Free Psychometric Scale (Likert) and Liver Imaging Reporting and Data System (Lirads) in Diagnosing HCC in Patients With Chronic Liver Disease.

Abstract

Title: Agreement between criteria free psychometric scale (LIKERT) and liver imaging reporting and data system (LI-RADS) in diagnosing HCC in patients with chronic liver disease. Aim: LI-RADS was introduced to improve the accuracy and consistency of liver lesion reporting while imaging a cirrhotic patient. The aim of our study is to determine the agreement between conventional criteria free LIKERT scale and LI-RADS criteria in diagnosing HCC and also to study the interobserver agreement between three radiologists in diagnosing HCC using LIKERT and LI-RADS criteria. Methods: Data set consists of CECT observations of liver cirrhosis patients who underwent triple phase CT in HCC protocol from the department of Radiodiagnosis of Amrita Institute of Medical Sciences, Kochi during the period 2015 to 2020. Interpretation of data sets were done by 3 radiologists trained in abdominal imaging with a minimum 4 -week gap between LIKERT and LI-RADS scoring. The agreement between LIKERT and LI-RADS for categorizing liver nodules by the senior radiologist is studied by using Cohens Kappa. Interobserver agreement between 3 radiologists for LI-RADS and for LIKERT is also studied separately using Fleiss Kappa statistics. Results: There is moderate agreement between LIKERT and LI-RADS [κ- 0.6 (0.516-0.751)] for characterizing liver lesions in CLD patients. The Fleiss Kappa coefficient shows almost perfect interreader agreement between three observers in characterizing liver lesions into HCC and non -HCC using LIKERT approach [κ- 0.81 (0.807-0.813)] and using LIRADS criteria [κ- 0.93 (0.921-0.927)] with more interreader agreement while reporting using LI-RADS criteria, especially in characterizing liver observations less than 3 cm . Conclusion: Characterizing liver observations in CLD patients using LIKERT and LI-RADS approaches had significant uniformity and consistency in characterizing liver observations. However, LI-RADS showed a higher interreader agreement compared to non-standardized reporting especially for nodules less than 3 cm.