Comparison of usefulness of clinical diagnostic criteria for hepatocellular carcinoma in a hepatitis B endemic area

Abstract

Background/aims: We compared the accuracy and usefulness of clinical diagnostic criteria for hepatocellular carcinoma in a hepatitis B virus (HBV)-endemic area.

Methods: We reviewed the medical records of 355 patients who had undergone liver resection or biopsy at our institution between January 2008 and December 2009. These patients were reevaluated using four noninvasive diagnostic criteria for hepatocellular carcinoma proposed by the European Association for the Study of the Liver (EASL), the American Association for the Study of Liver Diseases (AASLD), the Korean Liver Cancer Study Group and the National Cancer Center (KLCSG/NCC), and National Comprehensive Cancer Network (NCCN) guidelines.

Results: The overall sensitivity was highest using the KLCSG/NCC criteria (79.8%), followed by the AASLD (51.5%), EASL (38.4%), and NCCN (10.1%; P<0.001) criteria, whereas the specificity (84.5-98.3%) and positive predictive value (96.2-98.3%) were similar for all of the criteria. The KLCSG/NCC criteria had an acceptable false-positive rate and the highest sensitivity among all of the patients, including those positive for HBsAg, those without liver cancer, and those with a tumor of at least 2 cm.

Conclusions: The KLCSG/NCC and AASLD criteria exhibited the highest sensitivity, and all four guidelines had a high specificity among all of the patients. Based on the sensitivity and false-positive rate, the KLCSG/NCC criteria was the most useful in the majority of patients. Inclusion of HBV infection in the clinical diagnostic criteria for hepatocellular carcinoma would be reasonable and may lead to an improvement in the sensitivity, with acceptable false-positive rates, in HBV-endemic areas.

Keywords: Clinical diagnostic criteria; Comparison; Hepatocellular carcinoma.

Figure 1

Comparison of sensitivity among clinical diagnostic criteria for hepatocellular carcinoma. The sensitivity was the highest using the KLCSG/NCC criteria (79.8%), followed by the AASLD (51.5%), EASL (38.4%), and NCCN (10.1%) criteria. The sensitivity differed significantly between the patient groups (P<0.001).

Figure 2

Comparison of specificity among the clinical diagnostic criteria for hepatocellular carcinoma. The specificity was 96.6% for EASL, 89.7% for AASLD, 98.3% for NCCN, and 84.5% for KLSCG/NCC criteria. The differences between the criteria were not statistically significant (P=0.05).

Figure 3

Comparison of positive predictive values (PPVs) among the clinical diagnostic criteria for hepatocellular carcinoma. The PPV was similar among the four noninvasive criteria (98.3%, 96.2%, 96.8%, and 96.3% for EASL, AASLD, NCCN, and KLSCG/NCC criteria, respectively).

Figure 4

Comparison of sensitivity among the clinical diagnostic criteria in HBsAg (+) and HBsAg (-) subgroups with or without liver cancer (LC). The sensitivities of the AASLD and KLCSG/NCC criteria were highest in the HBsAg (+) LC (+) subgroup (92.0% and 91.3%, respectively), followed by the EASL criteria (66.7%, P<0.001). The sensitivities of the EASL, AASLD, and KLCSG/NCC criteria in the HBsAg (-) LC (+) subgroup were 73.7%, 78.9%, and 78.9%, respectively. It is noteworthy that the KLCSG/NCC criteria had a sensitivity of 90.5% in the HBsAg (+) LC (-) subgroup.

Figure 5

Comparison of specificity among the clinical diagnostic criteria in the HBsAg (+) and HBsAg (-) subgroups with or without LC. In the LC (+) subgroup with or without HBV infection, the NCCN criteria had a specificity of 100%; the specificity of the EASL criteria was 85.7% for the HBsAg (+) group and 75.0% for the HBsAg (-) group.