Department of Neurological Surgery, University of Florida, Gainesville, FL, USA.
University of Florida College of Medicine, Gainesville, FL, USA.
J Neurooncol. 2022 Sep;159(2):479-484. doi: 10.1007/s11060-022-04083-8. Epub 2022 Jul 15.
To determine whether participation in a clinical trial was associated with improved survival in patients with glioblastoma (GBM).
Following IRB approval, patients were identified using CPT and ICD codes. Data was collected using retrospective review of electronic medical records. When necessary, death data was obtained from online obituaries. Inverse propensity score matching was utilized to transform the two cohorts to comparable sets. Survival was compared using Kaplan-Meyer curves and Wilcoxon Rank Sum Test.
In this cohort of 365 patients, 89 were enrolled in a clinical trial and 276 were not. Patients enrolled in clinical trials had a significantly higher mean baseline KPS score, higher proportion of surgical resections, and were more likely to receive temozolomide treatment than patients not enrolled in a clinical trial. After inverse propensity score matching, patients enrolled in a clinical trial lived significantly longer than those not enrolled (28.8 vs 22.2 months, p = 0.005). A potential confounder of this study is that patients not in a clinical trial had significantly fewer visits with neuro-oncologists than patients enrolled in a clinical trial (7 ± 8 vs 12 ± 9, p < 0. 0001).
Clinical trials enroll patients with the most favorable prognostic features. Even when correcting for this bias, clinical trial enrollment is an independent predictor of increased survival regardless of treatment arm.
确定参与临床试验是否与胶质母细胞瘤（GBM）患者的生存改善相关。
在获得机构审查委员会（IRB）批准后，使用 CPT 和 ICD 代码来识别患者。通过回顾电子病历来收集数据。必要时，通过在线讣告获取死亡数据。利用逆概率评分匹配将两个队列转化为可比的队列。通过 Kaplan-Meier 曲线和 Wilcoxon 秩和检验比较生存情况。
在这组 365 名患者中，有 89 名患者参加了临床试验，276 名患者未参加。参加临床试验的患者基线 KPS 评分显著较高，手术切除比例更高，并且比未参加临床试验的患者更有可能接受替莫唑胺治疗。经过逆概率评分匹配后，参加临床试验的患者的生存期明显长于未参加临床试验的患者（28.8 个月比 22.2 个月，p=0.005）。本研究的一个潜在混杂因素是，未参加临床试验的患者与神经肿瘤学家的就诊次数明显少于参加临床试验的患者（7±8 次比 12±9 次，p<0.0001）。
临床试验招募的是预后最佳的患者。即使纠正了这种偏差，临床试验的参与仍然是生存改善的独立预测因素，而与治疗组无关。
