Published online Jul 28, 2021. doi: 10.3748/wjg.v27.i28.4687
Peer-review started: January 27, 2021
First decision: February 24, 2021
Revised: March 10, 2021
Accepted: July 13, 2021
Article in press: July 13, 2021
Published online: July 28, 2021
Diagnostic accuracy of various tumor markers and their combinations for hepatocellular carcinoma (HCC) was not fully investigated.
To evaluate the diagnostic accuracy of alpha-fetoprotein (AFP), the Lens culinaris agglutinin-reactive fraction of AFP (AFP-L3), and protein induced by vitamin K absence or antagonist-II (PIVKA-II) and their combination for HCC diagnosis.
Patients with newly detected liver mass or elevated serum AFP levels were considered eligible. Serum AFP level, AFP-L3 fraction, and PIVKA-II level were measured at the first visit.
In total, 622 patients were included; 355 patients (57.1%) had chronic liver disease, and 208 (33.4%) had liver cirrhosis. HCC was diagnosed in 160 patients (25.7%). The area under the receiver operating characteristics curves (AUROCs) of the serum AFP, AFP-L3 fraction, AFP-L3, and PIVKA-II levels for the diagnosis of HCC were 0.775, 0.792, 0.814, and 0.834, respectively. A novel diagnostic model was developed by classifying patients in a 1:1 ratio into training and validation sets. Using the binary regression analysis of the training cohort, the AFP, AFP-L3 fraction, and PIVKA-II (ALPs) score was calculated as follows: ALPs score = 3.8 × [serum AFP level (ng/mL) × AFP-L3 fraction (%) × 0.01] + 0.2 × PIVKA-II level (mAU/mL). The AUROC of the ALPs score for diagnosis of HCC was 0.878, significantly higher than that of serum AFP level (P < 0.001), AFP-L3 fraction (P < 0.001), PIVKA-II level (P = 0.036), and AFP-L3 level (P = 0.006). The optimal ALPs score cut-off was 5.3 (sensitivity, 85.0%, specificity 80.1%). The validation cohort showed similar results.
The ALPs score calculated using serum AFP level, AFP-L3 fraction, and PIVKA-II level showed improved accuracy in HCC diagnosis.
Core Tip: The value of Lens culinaris agglutinin-reactive fraction of alpha-fetoprotein (AFP-L3) for improving diagnostic accuracy of hepatocellular carcinoma has not been fully evaluated. We investigated performance of AFP-L3 in patients with newly detected liver mass or elevated serum AFP level. In total of 622 patients, we observed significantly higher diagnostic accuracy of AFP-L3 level than with AFP level (P < 0.001). In addition, we developed a novel diagnostic model, AFP, AFP-L3 fraction, and protein induced by vitamin K absence or antagonist-II (PIVKA-II) score, derived from AFP level, AFP-L3 fraction, and PIVKA-II level, and it showed significantly higher area under the receiver operating characteristics curves of 0.878 than those of AFP level, AFP-L3 fraction, and PIVKA-II level (all P < 0.05).