PO.CL05.04 · 临床研究

大型泛癌症队列中多系统免疫相关不良事件的特征与预测因素：一项回顾性分析

Characteristics and predictors of multi-system immune-related adverse events in a large pan-cancer cohort: A retrospective analysis

编号 6539 展板 5 时间 4/21 02:00–05:00 区域 Section 44 主讲 Julia Contini, BS

分会场 Immune Checkpoint Blockade

查看 PDF 下载 PDF 🔒 查看 / 下载完整 PDF 需登录并开通下载套餐 · 查看套餐 / 开通 AACR 官方页面

作者与单位 Authors & Affiliations

Julia G. Contini¹, Min Jung Koh², Charmi Trivedi², Mackenzie Adams², Jacqueline J. Chu², Sapana R. Gupta², Curtis Petruzzelli², Kanika Malani², Rebecca Steuer², Sandeep Jain², William Park², May Min², Matthew J. Hadfield²

¹The Warren Alpert Medical School of Brown University, Providence, RI,²Brown University Health, Providence, RI

摘要 Abstract

中文摘要

背景：免疫检查点抑制剂(ICI)可引起复杂的多系统免疫相关不良事件(irAEs)，而这些事件的描述仍不充分。本研究在一个大型机构队列中刻画其发生频率和临床特征。方法：回顾性队列纳入2015-2025年在布朗大学健康中心接受≥1种ICI治疗的2,134例转移性癌症患者。识别irAEs并按受累器官系统分类。多系统irAEs定义为≥2个受累器官系统。按性别、癌症类型和治疗方案对频率进行分层。采用Fisher精确检验评估亚组差异，采用多变量logistic回归识别独立预测因素，以比值比(OR)及p值报告。结果：在2,134例患者中，24.8%(n=530)发生了irAE；22.4%(n=119)有多系统受累，最常见为两个器官受累(16.8%, n=89)，三个或以上受累较少见(5.7%, n=30)。常见的组合包括皮炎+内分泌(10%, n=12)、结肠炎+皮炎(9.2%, n=11)和结肠炎+肺炎(5.9%, n=7)，并在皮肤、胃肠、肺和内分泌系统中呈聚集性。多系统irAEs在治疗方案间差异显著（单药治疗中5.1% vs. "ICI+化疗/靶向"中4.7% vs. 双ICI中13.9%; p<0.001），但在性别间无差异(p=0.6)。它们在黑色素瘤/皮肤癌(p=0.004)和泌尿生殖系统癌症(p=0.03)中更为常见。在调整治疗方案的多变量logistic回归中，黑色素瘤/皮肤癌仍与多系统irAEs的较高发生几率显著相关(OR: 1.92, p=0.03)，而与泌尿生殖系统癌症的关联减弱(OR: 1.37, p=0.2)。结论：多系统irAEs占所有irAEs的近四分之一，在黑色素瘤/皮肤癌和泌尿生殖系统癌症以及接受双ICI治疗的患者中更为常见，支持对这些群体进行更密切监测的必要性。预测多系统vs.单一/无irAE风险的单变量和多变量logistic回归模型单变量 OR(p值) 多变量 OR(p值) 治疗方案 ICI单药治疗 0.33(<0.001) 0.36(<0.001) 0.37(<0.001) ICI+化疗/靶向 0.30(<0.001) 0.35(0.002) 0.36(0.002) 双ICI(参照组) 性别男性 0.89(0.5) 女性(参照组) 癌症类型胸部(参照：非胸部) 0.80(0.3) 黑色素瘤/皮肤(参照：非黑色素瘤/皮肤) 2.43(0.002) 1.92(0.03) GU(参照：非GU) 1.62(0.02) 1.37(0.2) GI(参照：非GI) 0.58(0.08) 头颈部(参照：非头颈部) 0.66(0.3)

查看英文原文 English abstract

Background: Immune checkpoint inhibitors (ICI) can cause complex, multi-system immune-related adverse events (irAEs) that remain poorly described. This study characterizes their frequency and clinical features in a large institutional cohort. Methods: Retrospective cohort of 2,134 metastatic cancer patients treated with ≥1 ICI at Brown University Health (2015-2025). IrAEs were identified and categorized by affected organ system. Multi-system irAEs were defined as ≥2 affected organ systems. Frequency was stratified by sex, cancer type, and treatment regimen. Fisher's exact tests assessed subgroup differences, and multivariable logistic regression identified independent predictors, reported as odds ratios (OR) with p-values. Results: Of 2,134 patients, 24.8% (n=530) developed an irAE; 22.4% (n=119) had multi-system involvement, most commonly with two organs (16.8%, n=89) and less commonly three or more (5.7%, n=30). Common pairings included dermatitis + endocrine (10%, n=12), colitis + dermatitis (9.2%, n=11), and colitis + pneumonitis (5.9%, n=7), with clustering across dermatologic, gastrointestinal, pulmonary, and endocrine systems. Multi-system irAEs differed significantly by treatment regimen (5.1% in monotherapy vs. 4.7% in “ICI + chemo/targeted” vs. 13.9% in dual-ICI; p<0.001), but not by sex (p=0.6). They were more frequent in melanoma/skin cancers (p=0.004) and genitourinary cancers (p=0.03). On multivariable logistic regression adjusting for treatment regimen, melanoma/skin cancers remained significantly associated with higher odds of multi-system irAEs (OR: 1.92, p=0.03), while the association with genitourinary cancers was attenuated (OR: 1.37, p=0.2). Conclusions: Multi-system irAEs comprised nearly one-quarter of all irAEs and were more frequent in melanoma/skin and genitourinary cancers, and among patients who received dual-ICI therapy, supporting the need for closer monitoring in these groups. Uni- and multivariate logistic regression model predicting risk of multi-system vs. single/no irAE Univariate OR (p-value) Multivariate OR (p-value) Treatment regimen ICI monotherapy 0.33 (<0.001) 0.36 (<0.001) 0.37 (<0.001) ICI + chemo/targeted 0.30 (<0.001) 0.35 (0.002) 0.36 (0.002) Dual ICI (reference group) Sex Male 0.89 (0.5) Female (reference group) Cancer type Thoracic (ref: non-thoracic) 0.80 (0.3) Melanoma/Skin (ref: non-melanoma/skin) 2.43 (0.002) 1.92 (0.03) GU (ref: non-GU) 1.62 (0.02) 1.37 (0.2) GI (ref: non-GI) 0.58 (0.08) Head & Neck (ref: non-head & neck) 0.66 (0.3)

利益披露 Disclosure

J. G. Contini, None.. M. Koh, None.. C. Trivedi, None.. M. Adams, None.. J. J. Chu, None.. S. R. Gupta, None.. C. Petruzzelli, None.. K. Malani, None.. R. Steuer, None.. S. Jain, None.. W. Park, None.. M. Min, None.. M. J. Hadfield, None.

← 返回 AACR 2026 检索