Clinical predictors of progression of a hepatic lesion from Li-RADS (LR) 3 to LR5 among patients (pts) at risk of hepatocellular carcinoma (HCC).

Abstract

481 Background: We sought to identify predictors of progression of LR3 lesions (i.e. indeterminate for HCC) to LR5 lesions (i.e. definitely HCC) on follow-up imaging among cirrhotic pts. Methods: Imaging reports with LR assignments were identified among pts seen at the University of Washington, 2013-2017. Cirrhotic pts with a LR3 lesion and follow-up scan within 1 year (yr) of LR3 lesion date were included (n = 313). Clinical features were abstracted from chart review. Survival analyses employing interval censoring were performed. Variables as potentially predictive of LR3 progression were identified in univariate analyses, with backwards elimination done (p < 0.05) to obtain the final multivariate model. Results: 20.4% of LR3 lesions progressed to LR5 within 1 yr; 73% were still LR3, 8% progressed to LR4. The population was predominantly male (61%), Caucasian (71%), older than 55 (63%). The most common cirrhotic etiologies were HCV (46.7%), alcohol (32.6%), and NASH (12.8%), not mutually exclusive. AFP at the time of LR3 scan was low if available (39% with AFP <5, 16% 5-10, 28% unknown). 22.7% had impaired liver function (ALBI grade 3); 19.5% lacked data to calculate ALBI grade. CT scan was the most common exam (56%). Multiple LR3 lesions were seen on 51% of scans. Most LR3 lesions were right sided (75%), < 1 cm (51%); 7% of lesions were > 2cm. Men (HR 2.0, p = 0.02), earlier scan yr (HR 0.47 per yr, p < 0.0001), older age (HR 1.42 per 15 yr, p = 0.047), lesion size (HR 1.21 for 2cm+, global p = 0.02) appeared as independent predictors of LR3 to LR5 progression based on the final model. Of 16 variables examined, men were more likely to have chronic HCV, history of alcohol use and less likely to have autoimmune hepatitis. No other differences were seen. In an a priori analysis, risk of male sex (HR 1.99, p = 0.03) persisted despite control for HCV, alcohol, age, race, scan yr, lesion size, and number of lesions. Conclusions: Identification of clinical factors associated with LR3 progression may allow for risk modeling tools that may assist in determining imaging frequency and timing of intervention. The increased risk among men vs women is not explained by clinical or radiographic features listed above.