Impact of a Structured Report Template on the Quality of CT and MRI Reports for Hepatocellular Carcinoma Diagnosis.
Academic Article
Overview
abstract
PURPOSE: To assess the impact of a Liver Imaging Reporting and Data System (LI-RADS) structured template on quality of reports for hepatocellular carcinoma (HCC). MATERIALS AND METHODS: A departmental structured LI-RADS template was adopted in April 2015. CT and MRI reports from September 2014 to February 2016 with probable or definite HCC were reviewed. Reporting of the following was recorded for each lesion and compared between template and free-text reports: (1) LI-RADS category, (2) Couinaud segment, and (3) unequivocal description of presence or absence of major LI-RADS HCC features: arterial phase hyperenhancement, "washout," diameter, threshold growth, and "capsule." RESULTS: There were 306 definite or probable HCCs, 125 (40.8%) reported with free text and 181 (59.2%) with the template. LI-RADS category was reported in 23 of 125 (18.4%) HCCs with free text and in 178 of 181 (98.3%) HCCs with the template (P < .001). Couinaud segment was reported in 102 of 125 (81.6%) HCCs with free text and in 181 of 181 (100%) HCCs with the template (P < .001). Diameter was reported in 118 of 125 (94.4%) HCCs with free text and in 181 of 181 (100%) HCCs with the template (P = .001). Threshold growth was reported in 36 of 125 (28.8%) HCCs with free text and in 169 of 181 (93.4%) HCCs with the template (P < .001). Arterial phase hyperenhancement was reported in 101 of 125 (80.8%) HCCs with free text and in 177 of 181 (97.8%) HCCs with the template (P < .001). Washout was reported in 93 of 125 (74.4%) HCCs with free text and in 178 of 181 (98.3%) HCCs with the template (P < .001). Capsule was reported in 24 of 125 (19.2%) HCCs with free text and in 176 of 181 (97.2%) HCCs with the template (P < .001). CONCLUSIONS: Use of structured LI-RADS template resulted in more comprehensive and consistent reporting of major HCC features and LI-RADS category compared with free-text reporting.
