Search

Background Accurate prognostic stratification of hepatocellular carcinoma (HCC) is vital for clinical trial enrollment and treatment allocation. Multiple scoring systems have been created to predict patient survival, but no standardized scoring systems account for radiologic tumor features. We sought to create a generalizable scoring system for HCC which incorporates standardized radiologic tumor features and more accurately predicts overall survival (OS) than established systems. Methods Clinicopathologic parameters were collected from a prospectively collected cohort of patients with HCC treated at a single institution. Imaging studies were evaluated for tumor characteristics. Patients were randomly divided into a training set for identification of covariates that impacted OS and a validation set. Cox models were used to determine the association of various factors with OS and a scoring system was created. Results We identified 383 patients with HCC with imaging and survival outcomes, n  = 255 in the training set and 128 in the validation cohort. Factors associated with OS on multivariate analysis included: tumor margin appearance on CT or MRI (hazard ratio [HR] 1.37, 95% CI, 1.01-1.88) with infiltrative margins portending worse outcomes than encapsulated margins, massive tumor morphology (HR 1.64, 95% CI, 1.06-2.54); >2 lesions (HR 2.06, 95% CI, 1.46-2.88), Child-Turcotte-Pugh class C (HR 3.7, 95% CI, 2.23-6.16), and portal vein thrombus (HR 2.41, 95% CI, 1.71-3.39). A new scoring system was developed and more predictive of OS than other well-established systems. Conclusions Incorporation of standardized imaging characteristics to established clinical and lab predictors of outcome resulted in an improved predictive scoring system for patients with HCC. [ABSTRACT FROM AUTHOR]

Background & Aims: The liver Imaging Reporting and Data System (LI‐RADS) is a comprehensive system for standardizing liver imaging in patients at high risk for hepatocellular carcinoma (HCC). We performed a meta‐analysis to determine the diagnostic performance of the LR‐5 category for HCC and the pooled proportions of HCCs in each LI‐RADS category using CT/MRI LI‐RADS v2017. Methods: We searched multiple databases for original studies reporting on the diagnostic accuracy of CT/MRI LI‐RADS v2017. Random‐effects models were used to determine the summary estimates of the diagnostic performance of the LR‐5 category and the pooled proportions of HCCs for each LI‐RADS category. Risk of bias and concerns regarding applicability were evaluated with the Quality Assessment of Diagnostic Accuracy Studies‐2 tool. Results: Fourteen studies (3 prospective studies and 11 retrospective studies) were included in the final analysis, consisting of 2056 patients, 2589 observations, and 1693 HCCs. The pooled per‐observation sensitivity was 67% (95% confidence interval [CI], 62%‐72%) with specificity of 92% (95% CI, 88%‐95%) in the LR‐5 category of CT/MRI LI‐RADS v2017 for diagnosing HCC. The pooled proportions of HCCs were 0% (95% CI, 0%‐0%) for LR‐1, 4% (95% CI, 0%‐8%) for LR‐2, 34% (95% CI, 23%‐44%) for LR‐3, 67% (95% CI, 53%‐81%) for LR‐4, and 92% (95% CI, 87%‐96%) for LR‐5. The proportions of HCCs were significantly different among LI‐RADS categories 1‐5 (P =.034). Conclusions: The LR‐5 category of CT/MRI LI‐RADS v2017 shows moderate sensitivity and high specificity for diagnosing HCC. Higher LI‐RADS categories contained higher proportions of HCCs. [ABSTRACT FROM AUTHOR]