OBJECTIVES: Dynamic contrast-enhanced CT scan and MRI are essential for preoperative diagnosis of hepatocellular carcinoma (HCC), using the established Barcelona and AASLD criteria, which have been validated in only a few reports. The aim of this study is to retrospectively assess the diagnostic performance of these criteria in diagnosing or excluding HCC in at-risk patients with histopathology confirmation. METHODS: After institutional review board approval, a HIPAA compliant study was performed. The study cohort consisted of 156 de novo hepatic nodules imaged by either dynamic contrast-enhanced CT or MRI within 90 days of histopathology. Images were retrospectively reviewed by two abdominal radiologists blinded to clinical details, and all nodules were categorized as either meeting or not meeting AASLD and Barcelona criteria. By AASLD or Barcelona criteria, HCC was defined as any nodule greater than or equal to 1 or 2 cm, respectively, with hyperenhancement relative to background liver on arterial phase and hypoenhancement relative to background liver on portal venous or delayed phases. Significant differences in cohorts were analyzed using chi squared analysis (p < 0.05). RESULTS: On biopsy, 141/156 (90.38%) nodules were diagnosed as HCC. The respective sensitivity, specificity, accuracy and positive predictive value of AASLD and Barcelona were 78.7% and 63.1% (sensitivity), 73.3% and 86.7% (specificity), 78.2% and 65.4% (accuracy), and 82% and 63% (positive predictive value) (p < 0.001). CONCLUSIONS: Using established imaging criteria, up to 21.8% of presumed HCC nodules are inaccurately characterized and many small HCC nodules remain undiagnosed.
OBJECTIVES: Dynamic contrast-enhanced CT scan and MRI are essential for preoperative diagnosis of hepatocellular carcinoma (HCC), using the established Barcelona and AASLD criteria, which have been validated in only a few reports. The aim of this study is to retrospectively assess the diagnostic performance of these criteria in diagnosing or excluding HCC in at-risk patients with histopathology confirmation. METHODS: After institutional review board approval, a HIPAA compliant study was performed. The study cohort consisted of 156 de novo hepatic nodules imaged by either dynamic contrast-enhanced CT or MRI within 90 days of histopathology. Images were retrospectively reviewed by two abdominal radiologists blinded to clinical details, and all nodules were categorized as either meeting or not meeting AASLD and Barcelona criteria. By AASLD or Barcelona criteria, HCC was defined as any nodule greater than or equal to 1 or 2 cm, respectively, with hyperenhancement relative to background liver on arterial phase and hypoenhancement relative to background liver on portal venous or delayed phases. Significant differences in cohorts were analyzed using chi squared analysis (p < 0.05). RESULTS: On biopsy, 141/156 (90.38%) nodules were diagnosed as HCC. The respective sensitivity, specificity, accuracy and positive predictive value of AASLD and Barcelona were 78.7% and 63.1% (sensitivity), 73.3% and 86.7% (specificity), 78.2% and 65.4% (accuracy), and 82% and 63% (positive predictive value) (p < 0.001). CONCLUSIONS: Using established imaging criteria, up to 21.8% of presumed HCC nodules are inaccurately characterized and many small HCC nodules remain undiagnosed.