PO.CL01.06 · 临床研究
与可切除胆道癌新辅助度伐利尤单抗联合吉西他滨、顺铂和白蛋白结合型紫杉醇（DURGAP）反应相关的肿瘤免疫微环境特征
Tumor immune microenvironment features associated with response to neoadjuvant durvalumab plus gemcitabine, cisplatin, and nab-paclitaxel (DURGAP) in resectable biliary tract cancer
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
背景：新辅助DurGAP治疗在可切除胆道癌（BTC）中显示出良好疗效。在一个30例患者的更新队列中（2025 The Lancet Summit：中国癌症控制，摘要#TLSU2025_0087；NCT05640791），客观缓解率（ORR）为60%，疾病控制率（DCR）为86.7%，已切除患者中R0切除率为93.8%。然而，反应差异背后的机制仍不清楚。本研究探索了与不同肿瘤消退分级（TRG）相关的肿瘤免疫微环境（TIME）。
方法：分析了DurGAP队列中的四例TRG1（主要反应）和四例TRG3（最小反应）患者。多重免疫荧光定量了CD8⁺ T细胞、CD4⁺ T细胞和调节性T细胞（Tregs）。每组一个代表性病例接受空间转录组测序，以评估细胞组成和空间分布。免疫荧光数据采用Mann-Whitney U检验进行比较。空间转录组数据使用Seurat和sctransform流程处理，通过Harmony进行标准化和批次校正，随后进行PCA和基于UMAP的聚类。使用FindAllMarkers注释空间域以识别域特异性基因。对差异表达基因（p<0.05，|log₂FC|>log₂1.5）进行GO和KEGG通路富集分析。
结果：与TRG3相比，TRG1肿瘤显示出显著更高的CD8⁺和CD4⁺ T细胞浸润，而Tregs在TRG3中富集。空间转录组学证实了这些发现，揭示TRG1中T和B淋巴细胞密度更高、恶性细胞更少，而TRG3表现出更高的肿瘤细胞比例和减少的效应细胞浸润。亚群分析显示，Tregs占总T细胞的比例在TRG3中明显更高，而CD8⁺/Treg比值在TRG1中显著更大。功能富集突出显示TRG1中的免疫激活和抗原呈递通路，与TRG3中代谢和增殖信号富集形成对比。这些空间和转录模式表明，强大的效应T细胞浸润和协调的免疫结构是有效肿瘤消退的基础，而免疫抑制性微环境则限制了治疗益处。
结论：不同的TIME特征区分了BTC中新辅助DurGAP的反应者与非反应者。更高的效应T细胞浸润、增高的CD8⁺/Treg比值和免疫富集的空间域与主要肿瘤消退相关，而Treg主导的微环境对应于不良反应。这些发现凸显了可用于优化患者选择和改善未来治疗策略的潜在免疫生物标志物。
查看英文原文 English abstract
Background: Neoadjuvant DurGAP therapy has shown promising efficacy in resectable biliary tract cancer (BTC). In an updated cohort of 30 patients (2025 The Lancet Summit: Cancer Control in China, Abstract #TLSU2025_0087; NCT05640791), the objective response rate (ORR) was 60%, disease control rate (DCR) 86.7%, and R0 resection rate among resected patients 93.8%. However, mechanisms underlying variable response remain unclear. This study explored the tumor immune microenvironment (TIME) associated with differential tumor regression grades (TRG).
Methods: Four TRG1 (major response) and four TRG3 (minimal response) patients from the DurGAP cohort were analyzed. Multiplex immunofluorescence quantified CD8⁺ T cells, CD4⁺ T cells, and regulatory T cells (Tregs). One representative case per group underwent spatial transcriptomic sequencing to assess cellular composition and spatial distribution. Immunofluorescence data were compared using Mann-Whitney U test. Spatial transcriptomic data were processed with the Seurat and sctransform pipelines, normalized and batch-corrected via Harmony, followed by PCA and UMAP-based clustering. Spatial domains were annotated using FindAllMarkers to identify domain-specific genes. Differentially expressed genes (p<0.05, |log₂FC|>log₂1.5) were subjected to GO and KEGG pathway enrichment analysis.
Results: TRG1 tumors displayed significantly higher infiltration of CD8⁺ and CD4⁺ T cells compared with TRG3, while Tregs were enriched in TRG3. Spatial transcriptomics confirmed these findings, revealing greater T and B lymphocyte density and fewer malignant cells in TRG1, whereas TRG3 exhibited higher tumor cell proportions and reduced effector infiltration. Subset analysis showed that the proportion of Tregs among total T cells was markedly higher in TRG3, while the CD8⁺/Treg ratio was substantially greater in TRG1. Functional enrichment highlighted immune activation and antigen presentation pathways in TRG1, contrasting with metabolic and proliferative signaling enrichment in TRG3. These spatial and transcriptional patterns indicate that robust effector T-cell infiltration and coordinated immune architecture underlie effective tumor regression, whereas immunosuppressive microenvironments limit therapeutic benefit.
Conclusions: Distinct TIME profiles distinguish responders from non-responders to neoadjuvant DurGAP in BTC. Higher effector T-cell infiltration, an increased CD8⁺/Treg ratio, and immune-enriched spatial domains are linked to major tumor regression, whereas Treg-dominant microenvironments correspond to poor response. These findings highlight potential immune biomarkers to refine patient selection and improve future therapeutic strategies.
利益披露 Disclosure
H. Yan, None..
W. Li, None..
L. Xu, None..
R. Li, None..
L. Fu, None..
Q. Qi, None..
X. Zhang, None..
J. Hao, None..
H. Li, None..
Y. Liu, None.