胶质瘤进展中放射敏感性指数的稳定性:纵向大块与单细胞分析
本研究利用GLASS联盟214对匹配的原发-复发胶质瘤大块RNA-seq及28例IDH野生型胶质母细胞瘤的6863个恶性细胞Smart-seq2数据,分析放射敏感性指数(RSI)的纵向变化与单细胞异质性。结果显示大块RSI在原发与复发间无显著变化,RSI变化不能独立预测生存;单细胞层面细胞状态间RSI差异显著但效应量小。
要点
- 原发与复发胶质瘤间大块RSI无显著变化,相关性较强。
- RSI变化与生存相关但多变量分析中非独立预后因素。
- 单细胞层面RSI细胞状态间差异显著但生物学效应有限。
核心结论
- 大块RSI在胶质瘤进展中保持稳定,其变化不能独立预测生存。
- 单细胞分析显示细胞状态间RSI变异显著但生物学效应微小,提示RSI特征在单细胞水平分辨率有限。
研究设计与样本
- 回顾性纵向分析214对匹配原发-复发胶质瘤大块RNA-seq(GLASS联盟)及28例IDH野生型胶质母细胞瘤的6863个恶性细胞Smart-seq2数据。
- RSI采用Eschrich 10基因模型计算,纵向变化用Wilcoxon符号秩检验,生存分析用Kaplan-Meier和Cox回归,细胞状态间差异用Kruskal-Wallis检验及Bonferroni校正两两比较。
主要发现
- 大块RSI在原发与复发肿瘤间无显著变化(ΔRSI = +0.009 ± 0.096,p = 0.17),原发-复发相关性较强(r = 0.65,p < 0.001)。
- IDH野生型肿瘤中观察到RSI边缘性升高(ΔRSI = +0.019,p = 0.042)。
- Kaplan-Meier分析提示RSI升高肿瘤生存期较短(中位31 vs. 47个月,log-rank p = 0.040),但多变量Cox回归中RSI变化非独立因素(HR = 0.83,p = 0.82)。
- 单细胞分辨率下,神经祖细胞(NPC)样细胞RSI最低(Kruskal-Wallis p = 3.0 × 10⁻⁵),但效应量小(rank-biserial r = 0.06–0.09)。
关键数值与推荐
- RSI计算采用Eschrich 10基因模型。
临床怎么用
- RSI不宜单独用于预测胶质瘤进展后的放射敏感性变化或生存获益。
- 单细胞水平RSI变异有限,临床解读需结合大块组织转录组及多因素模型。
局限性
- 单细胞分析仅纳入28例IDH野生型胶质母细胞瘤,样本量有限。
- 单细胞RSI差异效应量小(rank-biserial r = 0.06–0.09),生物学意义有限。
- RSI变化在多变量分析中非独立预后因素,提示存在混杂。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
PURPOSE: The radiosensitivity index (RSI) is a transcriptomic predictor of intrinsic radiosensitivity. The longitudinal behavior of the RSI in glioma and its variation across intratumoral cellular states have been unknown. We investigated RSI dynamics between matched primary and recurrent gliomas and examined RSI heterogeneity at single-cell resolution. METHODS: The RSI was calculated using the Eschrich 10-gene model in 214 matched primary–recurrent glioma pairs from the GLASS Consortium (bulk RNA-seq) and in 6,863 malignant cells from 28 IDH-wildtype glioblastomas profiled by Smart-seq2. Longitudinal changes were assessed by Wilcoxon’s signed-rank test. A survival analysis used Kaplan–Meier and Cox regression. Cell state-specific RSI differences were evaluated by the Kruskal–Wallis test with Bonferroni-corrected pairwise comparisons. RESULTS: The bulk RSI showed no significant change between primary and recurrent tumors (ΔRSI = + 0.009 ± 0.096, p = 0.17), with a strong primary–recurrent correlation (r = 0.65, p < 0.001). We observed a marginal RSI increase in the IDH-wildtype tumors (ΔRSI = + 0.019, p = 0.042). The Kaplan–Meier analysis suggested shorter survival in RSI-increasing tumors (median 31 vs. 47 months, log-rank p = 0.040), but the RSI change was not independent in multivariable Cox regression (HR = 0.83, p = 0.82). At single-cell resolution, neural progenitor cell (NPC)-like cells exhibited the lowest RSI (Kruskal–Wallis p = 3.0 × 10⁻⁵), though effect sizes were small (rank-biserial r = 0.06–0.09). CONCLUSION: The bulk RSI remains stable across glioma progression, and the RSI change does not independently predict survival. The single-cell analysis revealed significant but biologically modest RSI variation across cellular states, suggesting limited resolution of the RSI signature at the single-cell level.
引用
- Nakano H, Shiinoki T. Radiosensitivity index stability across glioma progression: a longitudinal bulk and single-cell analysis. J Neurooncol. 2026. PMID: 41984136. doi:10.1007/s11060-026-05570-y
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。