Comprehensive analysis of various strategies to improve the diagnostic performance of LR-5 observations in the Liver Imaging Reporting and Data System (LI-RADS) version 2018

Abstract

The Liver Imaging Reporting Data System (LI-RADS) is widely adopted for the non-invasive diagnosis of hepatocellular carcinoma (HCC). In this study, the following objectives were investigated: (1) possible strategies to improve the diagnostic performance of LR-5 observations without reducing specificity for HCC were investigated, (2) whether threshold growth is suitable as a major criterion for HCC diagnosis and (3) whether hepatobiliary phase signal intensity can be used to difference HCC with atypical imaging features from non-HCC malignancies within LR-M observations. Herein, treatment naïve patients who underwent gadoxetate-disodium enhanced magnetic resonance imaging and surgical resection for focal hepatic observation were retrospectively analyzed. When evaluating various strategies to improve LR-5 diagnostic performance, hepatic observations were categorized according to the Liver Imaging Reporting Data System (LI-RADS) version 2018 and the final categories were readjusted by upgrading LR-4 to LR-5 using ancillary features, arterial phase hyperenhancement interpreted with subtraction images, indication of no washout when arterial phase hyperenhancement was absent, extension of washout to the transitional phase, and subthreshold growth as a major feature. Category-readjusted LR-5 after upgrading LR-4 to LR-5 using ancillary features favoring HCC in particular, subthreshold growth as a major feature, extending washout to transitional phase and arterial phase hyperenhancement interpreted using subtraction images showed significantly increased sensitivity (P<0.001) without decreased specificity (Ps>0.05). When investigating whether threshold growth should remain as a major feature in the diagnosis of HCC, frequency of the major and ancillary features outlined in the Liver Imaging Reporting Data System (LI-RADS) were evaluated in HCC and non-HCC malignancies. Ancillary feature showing significantly higher prevalence in HCC was used to replace threshold growth as a major feature or was added as an additional major feature and the diagnostic performance of readjusted LR-category was compared to that based on the Liver Imaging Reporting Data System (LI-RADS) version 2018. Unlike APHE, washout, or enhancing capsule which were more frequent in HCCs than non-HCC malignancies, threshold growth was more frequent in non-HCC malignancies than HCCs (P<0.001). The mean size of non-HCC malignancies showing threshold growth was smaller than those without threshold growth (22.2mm vs. 42.9mm, P=0.040) and similarly for HCCs but without significant difference (26.8mm vs. 33.1mm, P=0.184). Fat-in-nodule was more frequent in HCCs than non-HCC malignancies (P=0.027). When fat-in-nodule replaced threshold growth as major feature, LR-5 sensitivity and specificity increased nonsignificantly from 73.2% to 73.9% (P=0.289) and from 98.2% to 98.5% (P>0.999), respectively. When evaluating whether hepatobiliary phase signal intensity of the tumor can be used to difference HCC with atypical imaging features from non-HCC malignancies, the hepatobiliary phase signal intensity of LR-M observation was categorized into dark, low, and iso-to-high groups. Signal intensity of the tumor was classified as dark when more than 50% of tumor showed hypointensity compared to spleen, as low when more than 50% of tumor showed hyperintensity compared to spleen but hypointensity compared to liver parenchyme, and as iso-to-high if there was even a focal iso-intensity or hyperintensity compared to liver parenchyma. Analysis of clinicopathological factors and association between imaging and histology was performed. Out of 106 LR-M, 42 (42%) were showed dark, 61 (58%) showed low, and 3 (3%) showed iso-to-high SI in HBP. All 3 iso-to-high SI LR-M were HCCs (P=0.060) and their major histologic differentiation was Edmondson grade 1 (P=0.001). 43 out of 61 (71%) low SI LR-M were iCCA or cHCC-CCA (P=0.002). Inter-reader agreement of HBP SI classification was excellent, with a kappa coefficient of 0.872. In conclusion, sensitivity of LR-5 was improved without loss of specificity via category readjustment using AFs favoring HCC in particular, subthreshold growth as a major feature, extending washout to transitional phase and APHE interpreted with subtraction images. In addition, threshold growth was found nondiagnostic of HCC and was more common in non-HCC malignancies. Based on our results, comparable diagnostic performance of LR-5 can be obtained if threshold growth is replaced by a more HCC-specific ancillary feature. Lastly, LR-M with iso-to-high signal intensity in hepatobiliary phase was prone to being HCC while LR-M with low signal intensity in hepatobiliary phase was prone to being tumor with fibrous stroma such as iCCA and cHCC-CCA. Classification of LR-M based on hepatobiliary phase signal intensity may serve as a promising method of differentiating HCC with atypical imaging features from non-HCC malignancies.