Qualitative and Quantitative Assessment of Enhancement and Washout Characteristics of HCC and non HCC Hypervascular Liver Tumours

Introduction: The image based diagnosis of Hepatocellular Carcinoma (HCC) relies on assessment of arterial enhancement and venous or delayed phase washout. This assessment is based on visual analysis; however, studies quantifying such enhancement and washout are lacking. Aim: To qualitative assessment of the imaging characteristics of HCC and non HCC hypervascular liver tumours followed by quantitative assessment of the enhancement and washout kinetics, and also to derive objective values that may help in differentiating HCC from non HCC on Computed Tomography (CT). Materials and Methods: The present study was carried out in PSG hospitals, Coimbatore, Tamil Nadu, India. Retrospective analysis of hypervascular liver lesions was performed over a study period of four years (January 2012 to February 2016). These lesions were divided into two groups namely; HCC and non HCC. All non HCCs were pathologically proven, while HCC cases included those diagnosed on histopathology or typical imaging findings based on American Association for the Study of Liver Diseases (AASLD) (2010) and Liver Imaging Reporting and Data Systems (LIRADS) (2014) diagnostic criteria along with an elevated Alpha Fetoprotein (AFP) of more than 400 IU/mL. A qualitative analysis of the enhancement and washout patterns was done followed by quantitative analysis using Regions of Interest (ROI) measurement of attenuation in Hounsfield units (HU). The quantitative parameters assessed were; Mean Attenuation (MA) of lesions in different phases, Attenuation Gain Percentage (AGP) on arterial phase, Arterial Phase Attenuation Change (AAC), Attenuation Loss Percentage in venous and delayed phases (ALP-V, ALP-D), Percentage Attenuation Ratio in venous and delayed phases (PAR-V, PAR-D) and Relative Washout Ratios from arterial to venous/delayed (RWR A-V, RWR A-D) and from venous to delayed phases (RWR V-D). Sensitivity and specificity were calculated for qualitative washout. For quantitative variables comparative analysis between HCCs and non HCCs was done using Student's t-test. ANOVA was used to compare different categories. ROC curves were drawn for values that were significantly different to establish a cutoff value. Results: A total of 120 lesions were evaluated consisting of HCCs (n=88) and non HCCs (n=32). Qualitatively majority (81) of HCCs demonstrated arterial enhancement and washout with eight cases of these demonstrating subtle washout requiring settlement by consensus. Washout were absent in seven cases. True mimics of HCC on qualitative analysis were hepatic adenomas, few cases of metastasis and hepatoblastoma were also there. The overall sensitivity of the washout appearance for predicting HCC was 92% while specificity was 78%. On quantitative analysis a PAR-V >= 106 has sensitivity of 70% and specificity of 75%, RWR-V-D >= 41.79 has sensitivity of 75% and specificity of 70% for HCC detection. Additionally significant difference was seen between RWR V-D of HCC when compared to cholangiocarcinomas (p-value 0.035) and metastasis (p-value 0.026) and also between PAR-V of HCC and metastasis (p-value 0.033). Conclusion: The qualitative analysis combined with quantitative assessment of washout using percentage attenuation ratio in venous phase and relative washout ratio from venous to delayed phase offers good sensitivity and specificity in image based diagnosis of HCC.