基于剂量组学预测肺癌患者放射性肺炎发生的研究

目的 探讨剂量组学在预测肺癌根治性放疗患者放射性肺炎发生中的应用潜能。方法 回顾性收集行根治性放疗的314例肺癌患者的临床资料、放疗剂量文件、定位及随访CT图像,根据临床资料及影像学随访资料对放射性肺炎进行分级,提取全肺的剂量组学特征,构建机器学习模型。应用1000次自助抽样法（bootstrap）的最小绝对值收敛和选择算子嵌套逻辑回归（LASSO‐LR）及1000次bootstrap的赤池信息量准则（AIC）向后法筛选与放射性肺炎相关的剂量组学特征,随机按照7∶3划分为训练集及验证集,应用逻辑回归建立预测模型,并应用ROC曲线及校正曲线评价模型的性能。结果 共提取120个剂量组学特征,经LASSO‐LR降维筛选得到12个特征进入“特征池”,再经过AIC向后法筛选,最终筛选出6个剂量组学特征进行模型构建,训练集AUC为0.77（95%CI为0.65~0.87）,独立验证集AUC为0.72（95%CI为0.64~0.81）。结论 利用剂量组学建立的预测模型具有预测放射性肺炎发生的潜力,但仍需继续纳入多中心数据及前瞻性数据进一步挖掘剂量组学的应用潜能。     

Racial differences in characteristics and prognoses between Asian and white patients with nonsmall cell lung cancer receiving atezolizumab: An ancillary analysis of the POPLAR and OAK studies

This study aimed to compare the differences in characteristics and prognoses between Asian and white patients receiving immunotherapy for nonsmall cell lung cancer (NSCLC). We studied 390 patients who received atezolizumab as part of the POPLAR or OAK trial, and analyzed the differences in baseline characteristics, outcomes and genetic mutations in blood samples between Asian and white patients. Overall survival (OS) was longer in Asian compared to white patients (median OS: 18.7 vs. 11.1 months; p = 0.005). Race was identified as an independent prognostic factor for OS (Asian vs. white: hazard ratio 0.647, 95% confidence interval 0.447–0.936, p = 0.021), together with performance status, histology, baseline sum of the longest tumor diameters (BLSLD) and number of metastatic sites. The two groups also differed in terms of characteristics including smoking history, BLSLD, epidermal growth factor receptor (EGFR) mutation frequency, programmed death-ligand 1 expression and blood-based tumor-mutation burden. Blood mutations of STK11, EGFR, KEAP1, POLE, GRM3, ATM and STAG2 were associated with treatment response, and TP53, KEAP1, APC, RB1, CREBBP, EPHA5 and STAG2 mutations were associated with OS. The blood-based mutation profiles differentiated between Asian and white patients, especially in relation to EGFR (23.8 vs. 8.5%), TP53 (30.2 vs. 46.9%) and STK11 (1.6 vs. 12.3%) mutations (all p < 0.05). The different clinicopathological features and mutation profiles in Asian and white patients may explain the superior outcome following atezolizumab treatment in Asian patients with NSCLC. The results of this study have important implications for further studies on racial disparities in relation to immunotherapy.