A threshold analysis of alpha-fetoprotein in diagnosis and screening of hepatocellular carcinoma

Objective To investigate the optimal cut-off value of alpha-fetoprotein ( AFP) in the diagnosis and early screening of hepatocellular carcinoma ( HCC) . Methods The clinical data of 2212 HCC patients who were diagnosed and hospitalized in our hospital and 1998 non-HCC patients were collected, and the AFP level was summarized. The AFP level was divided into 10 ranges of 10-20 μg/L, 21-65μg/L, 66-110 μg/L, 111-155 μg/L, 156-200 μg/L, 201-250 μg/L, 251-300 μg/L, 301-350 μg/L, 351-400 μg/L, and >400 μg/L, and a comparative analysis was performed for the sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio of each cut-off value of AFP, ultrasound, and their combination in the diagnosis of HCC. The receiver operating characteristic ( ROC) curve was plotted to determine the optimal cut-off value. Results The cut-off valve of AFP of 200-250 μg/L had the largest sum of sensitivity and specificity ( 1. 370 1) and the largest area under the ROC curve ( 0. 896 4) . AFP > 20 μg/L combined with ultrasound had the highest sensitivity ( 95. 35%) in the diagnosis of HCC, with a diagnostic odds ratio of 26. 13. Conclusion The optimal cut-off value of AFP in the diagnosis of HCC is 200 μg/L. When AFP combined with ultrasound is used for the screening of people at a high risk of HCC, AFP > 20 μg/L is recommended as a positive index, and its combination with differential diagnosis and close follow-up and examinations can reduce the false negative rate of screening.