We thank Shiha et al for their interest in our prediction model for hepatocellular carcinoma (HCC) after the sustained virologic response (SVR)1 and for the external validation of this model in Egyptian patients with hepatitis C virus (HCV)-related advanced fibrosis (F3) or cirrhosis (F4).2 Shiha et al evaluated their cohort using our prediction model and proposed that the model should be reclassified into three groups; low risk (score of 0-1), intermediate risk (score of 2-3) and high risk (score of 4-5). In addition, they noted that obesity levels varied among different countries, and the advantages of using the AFP level at SVR had not been fully emphasized. We would like to respond to these issues as follows.

First, Shiha et al reported that the c statics of our prediction model, when validated in their cohort, was 0.66 [c statics of their proposed General Evaluating Score (GES score) was approximately 0.8 in their cohort3]. Their cohort included only patients with advanced fibrosis (F3 or F4), whereas our cohort included patients both with and without (F2 or lower) advanced fibrosis. Therefore, we cannot directly compare the c statics of both these studies.

Second, one of the strengths of our prediction model was the objective indicators, which were commonly measured in the clinical management of the liver disease. In the GES score, the presence of cirrhosis was included in the component.3 Indeed, cirrhosis has been well established as a risk factor for HCC after SVR,4, 5 but a definitive non-invasive diagnosis of cirrhosis was difficult.

Third, in our prediction model, when cumulative HCC occurrence rates were divided by the cut-off value of 2, the annual HCC risk was approximately 2% in patients with scores ranging from 2 to 5 and less than 1% in those with scores of 0-1.1 In our cohort, more than half of the cases were patients with scores of 0-1, which was classified as the low-risk group that Shiha et al advocated; HCC surveillance could be reduced in such patients.

Fourth, as Shiha et al pointed out, obesity levels varied across different countries. The World Health Organization defines obesity as a body mass index (BMI) of 30 kg/m2 or more, but only a few patients in our cohort had a BMI of ≥30 kg/m2. It may be useful to change the cut-off value of BMI between Asians and Westerners.

Fifth, in our prediction model, we used the AFP level at SVR since the baseline AFP level was affected by liver inflammation. AFP levels at SVR may more accurately reflect the potential of liver carcinogenesis, and its inclusion is one of the strengths of our prediction model. In recent studies, post-treatment AFP levels were reported to be associated with HCC occurrence after SVR, and these studies supported our data.6-8

Finally, further validation in different cohorts may be needed.

The authors' declarations of personal and financial interests are unchanged from those in the original article.1