Background: Preoperative radiology has been widely used to detect and measure hepatocellular carcinoma (HCC). However, its accuracy and reliability are unclear. This study aimed to assess the ability of current radiology to measure tumour size in patients undergoing resection.
Methods: We evaluated 212 HCC patients undergoing curative resection. Tumour size measured in the pathological examination was correlated with that obtained in preoperative ultrasound (US) and contrast-enhanced dynamic computed tomography (CT). Accuracy and association with tumour recurrence were investigated.
Results: The mean size of the tumour was 4.5 +/- 2.6 cm and was accurate in both US and CT in only 6 (3%) patients. Cirrhosis (P = 0.015), absence of tumour stain (P = 0.002) and small (< or = 4 cm) tumour (P < 0.001) were the significant factors associated with size deviation using both US and CT. Ninety-four (44%) patients developed tumour recurrence within 17 +/- 11 months of resection. Recurrence rate was 52%, 52% and 67% in patients with underestimation in US (relative risk [RR]: 2.0, 95% confidence interval [CI]: 1.2-3.4, P = 0.01), CT (RR: 2.1, 95% CI: 1.1-4, P = 0.022) and both modalities (RR: 2.5, 95% CI: 1.4-4.2, P = 0.001), respectively, compared to 30% recurrence in patients with accurate estimation of tumour size.
Conclusion: The accuracy of radiology in measuring tumour size was poor, and may lead to inappropriate treatment. The finding that underestimation of tumour size was associated with a higher tumour recurrence rate is consistent with the hypothesis that HCC may recur from pre-existing tumour foci which could not be identified from the current imaging modalities.