非小细胞肺癌新辅助免疫治疗:CT长径均值变化预测主要病理反应
这项多中心回顾性研究纳入48例接受新辅助免疫治疗的可切除非小细胞肺癌患者,以主要病理反应(残留肿瘤细胞≤10%)为参考标准。结果显示Δ平均直径(长径均值变化)预测主要病理反应的AUC为0.83,敏感度76%、特异度84%、总准确率79.2%,数值上高于RECIST 1.1的72.9%。多变量分析中Δ平均直径是唯一独立相关参数(OR=11.91,95%CI 1.52-93.11,p=0.018)。
要点
- Δ平均直径可作为新辅助免疫治疗后病理反应的简单影像学预测指标。
- 该参数数值准确率高于RECIST 1.1,但无正式优效性统计检验。
- 结果仅具假设生成性,临床实施前需内部和外部验证。
核心结论
- Δ平均直径是预测新辅助免疫治疗病理反应的有前景的影像生物标志物,数值上准确率高于RECIST 1.1。
- 研究结果为假设生成性,临床实施前需要内部和外部验证。
研究设计与样本
- 多中心回顾性研究,纳入2018年9月至2023年7月接受新辅助免疫治疗的48例可切除I-IIIB期非小细胞肺癌患者。
- 治疗前后测量CT密度参数(CT值和相对CT值)与形态学参数(平均直径、体积、表面积),以主要病理反应(残留肿瘤细胞≤10%)为参考标准。
- 统计方法为受试者工作特征曲线分析和逻辑回归评估参数表现。
关键影像表现/征象
- Δ平均直径在非小细胞肺癌新辅助免疫治疗后预测主要病理反应中显示最高预测准确率(AUC=0.83,p<0.001)。
- 形态学和密度参数在治疗前后的变化是评估非小细胞肺癌新辅助免疫治疗反应的影像学观察指标。
诊断标准与关键数值
- 主要病理反应定义为残留肿瘤细胞≤10%。
- Δ平均直径预测主要病理反应的AUC为0.83(p<0.001),敏感度76%,特异度84%,总准确率79.2%。
- RECIST 1.1的准确率为72.9%,Δ平均直径数值上更高,但未进行正式优效性统计检验。
- 探索性多变量分析中Δ平均直径为唯一保持独立相关的参数(OR=11.91,95%CI 1.52-93.11,p=0.018)。
临床应用与阅片要点
- Δ平均直径作为一种简单CT参数,可用于非完全缓解患者中识别治疗疗效,辅助个体化治疗决策。
- 无需复杂影像组学或昂贵PET即可改善免疫治疗疗效评估的准确性。
- 目前证据不足以直接用于临床决策,需等待进一步验证。
局限性
- 回顾性研究设计存在选择偏倚和信息偏倚风险。
- 样本量较小(48例),多变量分析为探索性,结果需谨慎解读。
- 未对Δ平均直径与RECIST 1.1的准确率差异进行正式统计检验。
- 缺乏内部和外部验证。
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)。
英文原文(摘要)· 点击展开
OBJECTIVE: Neoadjuvant immunotherapy for non-small cell lung cancer shows heterogeneous efficacy, and Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1) fails to accurately predict pathological response. The objective was to explore computed tomography parameters associated with neoadjuvant immunotherapy efficacy and compare their performance with RECIST 1.1.
MATERIALS AND METHODS: This multicenter retrospective study included 48 patients with resectable stage I-IIIB non-small cell lung cancer who received neoadjuvant immunotherapy between September 2018 and July 2023. Computed tomography density parameters (computed tomography value and relative computed tomography value) and morphological characteristics (mean diameter, volume, surface area) were measured before and after treatment. Major pathological response (≤ 10% residual tumor cells) served as the reference standard. Receiver operating characteristic analysis and logistic regression evaluated parameter performance.
RESULTS: ΔMean diameter demonstrated the highest predictive accuracy for major pathological response (area under the curve 0.83, p < 0.001), with 76% sensitivity, 84% specificity, and 79.2% overall accuracy, with numerically higher accuracy than Response Evaluation Criteria in Solid Tumors version 1.1 (72.9%), though no formal statistical test for superiority was conducted. In exploratory multivariable analysis, it was the only parameter retaining independent association with major pathological response (odds ratio 11.91, 95% confidence interval 1.52-93.11, p = 0.018).
CONCLUSION: ΔMean diameter is a promising imaging biomarker for predicting pathological response to neoadjuvant immunotherapy. In this cohort, it showed numerically higher accuracy than RECIST 1.1; however, these findings are hypothesis-generating and require both internal and external validation before clinical implementation.
KEY POINTS: Question Neoadjuvant immunotherapy for resectable non-small cell lung cancer shows heterogeneous efficacy, and RECIST 1.1 fails to predict major pathological response, requiring reliable biomarkers. Findings ΔMean diameter was associated with major pathological response (AUC = 0.83), showed numerically higher accuracy than RECIST 1.1, and represented an exploratory independent predictor. Clinical relevance This simple computed tomography parameter shows promise for identifying treatment efficacy in non-complete responders, aids personalized therapy decisions, and improves the accuracy of immunotherapy evaluation without complex radiomics or costly positron emission tomography.
引用
- Liu, Xie, Liu, et al. ΔAverage of long and short axis diameter on non-contrast CT is associated with major pathological response to neoadjuvant immunotherapy in resectable non-small cell lung cancer. Eur Radiol. 2026. PMID: 42595861. doi:10.1007/s00330-026-12716-z
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。