Introduction Hepatocellular carcinoma (HCC) behaves differently in cirrhotic and non-cirrhotic livers. Microvascular invasion (MVI) is a key determinant of outcome following curative treatment for HCC. Despite attempts to reach a consensus, reliably identifying microvascular invasion remains difficult. Interrogating factors influencing MVI in patients with or without cirrhosis may determine more reliably identifiable factors or surrogates for MVI and provide valuable insight into the underlying biology of HCC.
Methods Review of a prospective database of 229 consecutive patients undergoing hepatectomy (n=164) or transplantation (n=65) for HCC at St James's University Hospital, UK between 1998 and 2011. Diagnosis was based on published guidelines and incidental explant tumours excluded. 115, 50.2% occurred in cirrhotic compared to 114, 49.8% in non-cirrhotic livers. Clinico-pathological characteristics were correlated with survival and MVI. Survival was calculated using the Kaplan–Meier method with Log-rank and Cox stepwise regression for survival comparisons. Univariate χ2 and multivariate logistic regression were used to analyse relationships between clinico-pathological variables and MVI (p<0.05 was indicative of statistical significance).
Results In non-cirrhotic patients recurrence independently predicted overall survival (OS) (p=0.001) while multifocal tumours (p=0.042) and viral aetiology (p=0.029) independently predicted disease free survival (DFS). In cirrhotic patients recurrence (p<0.001), MVI (p<0.001) and tumour size >5 cm (p<0.005) predicted overall survival (OS) and disease free survival (DFS) in univariate analysis. Only recurrence (p=0.001) for OS and MVI (p=0.002) and tumour size >5 cm (p=0.027) for DFS retained independence on multivariate analysis. Univariate analysis of pre-operative variables revealed MVI was significantly associated with multifocal HCC and poor differentiation in non-cirrhotic patients (p=0.04 and p=0.019), and with viral aetiology in cirrhotic patients (p=0.047).
Conclusion In cirrhotic patients MVI was an independent predictor of DFS while recurrence strongly determined OS. Viral aetiology was the only significant pre-operative factor associated with MVI in the explant. In non-cirrhotic patients multifocality strongly predicted DFS and was associated with MVI. We hypothesise that multifocality in non-cirrhotic HCC may actually be representative of MVI rather than multifocal de novo tumour formation. Given the challenges of robustly identifying MVI in these patients, multifocality could be an extremely useful prognosticator and histopathological indicator of MVI, which we know carries series implications for our patients.
Competing interests None declared.