Diagnostic Performance of Contrast-enhanced Imaging Techniques for Small Hepatocellular Carcinoma Measuring less than 2 cm

Abstract

BACKGROUND: Recent guidelines recommend ultrasound follow-up for nodules (< 1 cm) discovered via ultrasonography surveillance (US) in liver cirrhosis patients. Nodules larger than 1cm should be followed the recall policy with either dynamic contrast 4-phase CT scan or dynamic enhanced contrast MRI. There is no specified recall policy of nodule less than 1 cm found on US. We evaluated CT and MRI efficacy in hepatocellular carcinoma (HCC) diagnosis in nodules lesser than 1cm and 1 ∼2 cm.

METHODS: We retrospectively analyzed 130 ultrasonography-detected liver nodules <2 cm. The liver nodules were proved histologic confirmation either by percutaneous biopsy or surgical specimen analysis between 2007 and 2013. The patients with liver nodules underwent quadruple-phase CT at 16 MDCT and Gd-EOB-DTPA-enhanced 3T MRI. And image studies were examined by 2 radiologists experienced in imaging of the liver, who were unaware of the results of the biopsy. The detailed criteria for HCC were: a lesion that had an arterial hypervascularization with definite delayed washout and/or capsular rim enhancement (fibrous capsule) was HCC. We separated the patients with two diagnostic performance of the imaging work-up strategies. The two strategies were single imaging scan only, and two imaging scans. Sensitivity, positive predictive values and diagnostic accuracy were evaluated.

RESULTS: Thirty-six nodules (27.6%) were <1 cm, while 94 nodules were 1-2 cm in size. Histolog-ically, HCC was confirmed in 110 of 130 (84.6%) nodules; specifically 29 of 36 (80.5%) nodules <1 cm and 81 of 94 (86.2%) nodules 1–2 cm in size were HCCs. Imaging revealed that 24 of 29 HCCs <1 cm displayed the typical HCC vas-culature (82.7%). Imaging revealed false-negatives in 14% (5/36) of nodules <1 cm. In nodules <1 cm, the sensitivity, positive predictive value (PPV), and diagnostic accuracy were 80.0%, 66.6%, and 66.6%, respectively for single imaging and 83.3%, 95.2%, and 81.4%, for two images. Of 81 HCC nodules measuring 1–2 cm, 79 possessed the typical HCC vasculature (97.5%). One nodule was histologically diagnosed as HCC, despite a negative imaging result upon single image study. In these nodules, the sensitivity, PPV, and diagnostic accuracy were 96.0%, 92.3%, and 90.3%, respectively, for single imaging and 98.2%, 94.8%, and 93.6%, respectively, for two images.

CONCLUSIONS: In HCCs (< 1 cm), our study suggests that two imaging modalities provided limited diagnostic efficacy; invasive percutaneous biopsy may be needed. In HCCs (1-2 cm), single or two imaging modalities produced similar diagnostic efficacy. Further studies are needed, with long term follow-up in a larger cohort of (< 1 cm) HCC patients.