We read with interest the article by Khalili et al. 1 addressing the management of small liver nodules detected in patients with cirrhosis under surveillance with abdominal ultrasound (US) that gave indeterminate results by contrast imaging. To optimize American Association for the Study of the Liver Disease (AASLD) guidelines, 2 the authors suggest performing a fine-needle biopsy examination of nodules showing either arterial hypervascularity on computed tomography (CT) / magnetic resonance imaging (MRI) or accompanied by a synchronous hepatocellular carcinoma (HCC) only, since these were the only independent variables associated with malignancy in their retrospective study. According to this algorithm, approximately 20% of additional tumors will be identified, with a sensitivity of 62%, a specificity of 79%, and a 73% save of liver biopsies.When we applied this algorithm to our patients with a de novo liver nodule prospectively detected during surveillance 3 ( Fig. 1), the corresponding figures were 44% for sensitivity and 55% for specificity, with a positive predictive value of 44% and a negative predictive value of 55%, respectively. Overall, among 36 1-2 cm indeterminate nodules the modified algorithm would have diagnosed 7 (44%) of tumors only of the 16 identified by histology, including 15 HCC and 1 intrahepatic cholangiocarcinoma (ICC). At the same time, the diagnosis of HCC would have been significantly delayed in nine (56%) patients compared with none if treated according to AASLD guidelines. The fact that the majority (75%) of delayed diagnoses were in patients with a very early HCC, i.e., the ideal candidates for radical treatment with local ablation, 4 attenuates the appeal of the modified algorithm, which in addition would have also led to a misdiagnosis of ICC in one nodule devoid of contrast uptake during the arterial phase of CT/MRI. Due to the high incidence of HCC in patients with compensated cirrhosis and the low risk of liver biopsy complications, we strongly endorse unmodified AASLD guidelines for the management of patients with cirrhosis with 1-2 cm liver nodules with undefined radiological diagnosis. Fig. 1. The diagnostic yield of revised AASLD algorithm in our patients undergoing surveillance. Liver histology was the gold standard for the diagnosis of 1-2 cm HCC.
651Reply:We thank Drs. Iavarone and Sangiovanni for the interest in our study, 1 but caution the authors with regard to several points in their letter. First, their conclusions are based on a small sample size of 36 indeterminate nodules. While they calculate sensitivity and specificity of 44% and 55%, respectively, using our proposed criteria, the 95% confidence interval was not reported. We calculate their confidence interval to be 21%-69% for sensitivity and 32%-76% for specificity. Small sample sizes lead to real uncertainty.Second The substantial lower sensitivity reported by Iavarone and Sangiovanni may be a result of their small sample size, but should lead to reexamination of their methodology. It is unclear whether ...