当前位置: X-MOL 学术J. Clin. Oncol. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Contradiction Between Stated Statistical Analysis Method and Displayed Survival Curves.
Journal of Clinical Oncology ( IF 42.1 ) Pub Date : 2020-01-02 , DOI: 10.1200/jco.19.02527
Fei Liang 1
Affiliation  

In their recent article, Noronha et al1 reported that adding pemetrexed and carboplatin chemotherapy to gefitinib significantly prolonged progression-free survival (PFS) and overall survival (OS). I have some concerns about their statistical analysis methods. As the authors reported in the Patients and Methods section, PFS and OS were estimated using the Kaplan-Meier method, and the 2 arms were compared using the log-rank test. However, the curves in Figure 3 of their article, which showed Kaplan-Meier estimates of PFS and OS in the full analysis set, are highly unlike survival curves of the Kaplan-Meier method. In a Kaplan-Meier survival curve, a turning point represents the occurrence of at least 1 event, and the x-axis of the turning point indicates the actual time of the event. In the OS survival curve of the gefitinib arm in Figure 3B, there were 21 turning points, and the authors reported that 80 patients died in the gefitinib arm. Therefore, it can be inferred that 80 patients died at only 21 time points during the 24 months of follow-up, with an average of 4 deaths per day for these 21 days, and that for the remaining 699 days (assuming 30 days per month), not a single patient died. Furthermore, according to the number of turning points on the x-axis, patients died only on the first day of each month, and no patients died on any other days of the month, which means patients died in a well-organized way at regular 1-month intervals. Similar characteristics can be seen in the OS curve of the gefitinib plus pemetrexed and carboplatin arm and PFS curves of the 2 arms.

中文翻译:

陈述的统计分析方法与显示的生存曲线之间的矛盾。

在最近的文章中,Noronha等人1报道称,在吉非替尼中加入培美曲塞和卡铂化疗可显着延长无进展生存期(PFS)和总体生存期(OS)。我对他们的统计分析方法有些担心。正如作者在“患者和方法”部分所报道的那样,使用Kaplan-Meier方法估算了PFS和OS,并使用对数秩检验比较了这两个组。但是,他们文章的图3中的曲线显示了完整分析集中对PFS和OS的Kaplan-Meier估计,与Kaplan-Meier方法的生存曲线截然不同。在Kaplan-Meier生存曲线中,转折点代表至少1个事件的发生,转折点的x轴指示事件的实际时间。在图3B中的吉非替尼臂的OS生存曲线中,有21个转折点,作者报告说,吉非替尼组有80例患者死亡。因此,可以推断出,在随访的24个月中,只有80个患者在21个时间点死亡,这21天平均每天有4例死亡,其余699天(假设每月30天)死亡),没有一个患者死亡。此外,根据x轴上的转折点数,患者仅在每个月的第一天死亡,而在该月的其他几天没有患者死亡,这意味着患者会定期有组织地死亡每隔1个月。在吉非替尼加培美曲塞和卡铂臂的OS曲线以及两个臂的PFS曲线中可以看到类似的特征。可以推断,在随访的24个月中,只有80个患者在21个时间点死亡,在这21天中平均每天有4例死亡,其余699天(假设每月30天),没有一个病人死亡。此外,根据x轴上的转折点数,患者仅在每个月的第一天死亡,而在该月的其他几天没有患者死亡,这意味着患者会定期有组织地死亡每隔1个月。在吉非替尼加培美曲塞和卡铂臂的OS曲线以及两个臂的PFS曲线中可以看到类似的特征。可以推断,在随访的24个月中,只有80个患者在21个时间点死亡,在这21天中平均每天有4例死亡,其余699天(假设每月30天),没有一个病人死亡。此外,根据x轴上的转折点数,患者仅在每个月的第一天死亡,而在该月的其他几天没有患者死亡,这意味着患者会定期有组织地死亡每隔1个月。在吉非替尼加培美曲塞和卡铂臂的OS曲线以及两个臂的PFS曲线中可以看到类似的特征。根据x轴上的转折点数,患者仅在每月的第一天死亡,而在该月的其他几天没有患者死亡,这意味着患者在正常的1时会以有组织的方式死亡。月间隔。在吉非替尼加培美曲塞和卡铂臂的OS曲线以及两个臂的PFS曲线中可以看到类似的特征。根据x轴上的转折点数,患者仅在每月的第一天死亡,而在该月的其他几天没有患者死亡,这意味着患者在正常的1时会以有组织的方式死亡。月间隔。在吉非替尼加培美曲塞和卡铂臂的OS曲线以及两个臂的PFS曲线中可以看到类似的特征。
更新日期:2020-03-06
down
wechat
bug