抗肿瘤免疫治疗的长期疗效评价

Long-term efficacy evaluation of anti-cancer immunotherapy

抗肿瘤免疫治疗在临床获益方面有显著提升,但是由于其治疗原理与其他抗肿瘤治疗有较大差异,因此当使用常规统计学方法评价免疫治疗的临床获益时,存在不完全适用的情况。尤其在长期疗效评价方面,免疫治疗的时间事件数据不完全符合等比例风险假设,因此使用常规生存数据的评价方法会引入偏差。本文针对此问题,讨论了适用于非比例风险假设数据的新颖长期疗效终点指标,并介绍长期疗效评价中的假设检验和疗效估计方法,同时强调了生物标志物的重要性。当前已有改良的统计学方法解决非比例风险假设,但是这些方法或多或少存在缺陷,需要各方进一步共同努力解决此问题。

In terms of clinical benefits, anti-cancer immunotherapy has been significantly improved. However, since its treatment principle is different from other anti-cancer therapies, common statistical methods are not completely applicable to immunotherapy when evaluating clinical benefit in certain circumstances. Especially in long-term efficacy evaluation, the time-to-event data of immunotherapy does not completely follow the proportional hazard assumption. Therefore, the survival analysis of the proportional hazard assumption may introduce bias. In response to this problem, this paper discusses novel long-term efficacy endpoints which are applicable to the non-proportional hazard hypothesis. We also present hypothesis testing and power estimation in the long-term evaluation, while the importance of biomarkers is emphasized. Currently, there are modified statistical methods that can solve the problem of non-proportional assumptions, but flaws still exist in these methods. Therefore, all parties need to work hard together to solve this problem.