Dear Editor,

We read with great interest the paper by Tian et al.¹ titled “Clinical value of serum AFP and PIVKA-II for diagnosis, treatment and prognosis of hepatocellular carcinoma.” We congratulate Tian et al. for the excellent research, but some issues of the research warrant discussion.

In figure 2a, b in Ref. [1], serum AFP and PIVKA-II levels were compared pairwise among the five groups by rank sum test according to the introduction of statistical methods. Multiple testing in the five groups increased the chance of making type I errors; however, Tian et al. did not pay attention to this matter in their paper. There are many different methods to control type I errors in multiple testing, such as Bonferroni correction by controlling the family-wise error rate and the Benjamini–Hochberg procedure by controlling the false discovery rate.^{2, 3}

The area under the receiver operating characteristic curve (AUROC) was calculated to assess diagnostic performance. Tian et al. found that the diagnostic value of the combined detection of AFP and PIVKA-II was greater than that of AFP and PIVKA-II alone, because the AUROC of the combination of AFP and PIVKA-II (0.975, 95% CI 0.956–0.994) was greater than that of AFP (0.903, 95% CI 0.862–0.944) and PIVKA-II (0.945, 95% CI 0.915–0.975) in comparison with HCC and benign liver disease. This conclusion appears to be arbitrary. It seems to be more rigorous that the statistical significance of difference between two AUROCs was evaluated by the Delong test.⁴

Despite the above-mentioned potential limitations, Tian et al. have made several contributions in assessing the role of serum AFP and PIVKA-II in the diagnosis, treatment, and prognosis of HCC.

亲爱的编辑，

我们饶有兴趣地阅读了 Tian 等人的论文。¹题为“血清AFP和PIVKA-II对肝细胞癌诊断、治疗和预后的临床价值”。我们祝贺田等人。尽管研究非常出色，但研究中的一些问题值得讨论。

在图2a、b中参考文献。[ 1 ] 引入统计学方法，采用秩和检验两两比较五组间血清 AFP 和 PIVKA-II 水平。五组中的多次测试增加了犯第一类错误的机会；然而，田等人。他们的论文中没有关注这个问题。有许多不同的方法可以控制多重测试中的 I 类错误，例如通过控制族错误率的 Bonferroni 校正和通过控制错误发现率的 Benjamini-Hochberg 程序。^{2, 3}

计算受试者工作特征曲线下面积 (AUROC) 以评估诊断性能。田等人。研究发现，AFP 和 PIVKA-II 联合检测的诊断价值大于单独检测 AFP 和 PIVKA-II，因为 AFP 和 PIVKA-II 联合检测的 AUROC（0.975，95% CI 0.956–0.994）为与 HCC 和良性肝病相比，该结果高于 AFP（0.903，95% CI 0.862–0.944）和 PIVKA-II（0.945，95% CI 0.915–0.975）。这个结论似乎是任意的。通过Delong检验评估两个AUROC之间差异的统计显着性似乎更加严格。⁴

尽管存在上述潜在的局限性，Tian 等人。在评估血清 AFP 和 PIVKA-II 在 HCC 诊断、治疗和预后中的作用方面做出了多项贡献。