KM Research Hub
← 返回论文库
已完成 入排:符合 图源:可得 入排置信度 0.98

Computationally Derived Spatial Immune Signature Identifies Trastuzumab Responders in HER2+ Breast Cancer: NSABP B-41 Clinical Trial Validation.

Satvika Bharadwaj, Germán Corredor, Hilmi Al-Shakhshir, Sebastian Medina, Sahar N Almahfouz, Rohan Dhamdhere, Tilak Pathak, Pingfu Fu, Yuan Liu, Shipra Gandhi, Sunil Badve, Anant Madabhushi

Clin Cancer Res 32(12) 2508-2518

PMID41915435
PMCIDPMC13266337
DOI10.1158/1078-0432.CCR-25-4185
发表日期2026-03-31
入排依据:原始人体肿瘤 Phase III 随机药物试验,并报告 time-to-event 疗效结局。
图像/No. at risk 复核:KM 曲线已提取;No. at risk 信息待人工复核,不影响当前曲线图像显示。
摘要

摘要

PURPOSE: Trastuzumab-based chemotherapy has improved outcomes in human epidermal growth factor receptor 2 (HER2)-positive breast cancer, but treatment benefit varies among patients. Predictive signatures are needed to identify patients most likely to respond to these therapies. EXPERIMENTAL DESIGN: We developed Density and Spatial architecture of Tumor-Infiltrating Lymphocytes (DeSTIL), a computational signature derived from hematoxylin and eosin slides. The signature captures spatial organization of immune cells and interactions with nonimmune cells. DeSTIL was trained on HER2+ breast cancer slides from The Cancer Genome Atlas (n = 250) and validated in a phase III National Surgical Adjuvant Breast and Bowel Project (NSABP) B-41 randomized clinical trial (n = 221), which compared chemotherapy plus trastuzumab, lapatinib, or combination. The DeSTIL scores were dichotomized into positive and negative groups, and event-free survival (EFS) was assessed using Cox proportional hazards with interaction terms. RESULTS: In NSABP B-41, DeSTIL-positive patients (n = 61) showed significantly improved event-free survival (EFS) with trastuzumab compared with the combination arm [hazard ratio (HR) = 0.09; 95% confidence interval (CI) = 0.01-0.77; P = 0.006] and a significant signature-treatment interaction (P = 0.024). No EFS difference was observed in DeSTIL-negative patients (n = 160). Gene expression analysis supported the image-derived signature stratifying DeSTIL-positive and DeSTIL-negative tumors. In an exploratory pathologic complete response analysis, a classifier trained on University Hospitals Cleveland slides achieved AUCs of 0.70 in the training cohort and 0.63 in the trastuzumab arm of the NSABP B-41 validation cohort. CONCLUSIONS: DeSTIL identifies a subset of HER2+ patients who derive greater benefit from trastuzumab. These findings support the potential of computationally derived immune architecture to inform selection of standard HER2-targeted therapies.

KM 提取结果

1 个 Figure · 4 个 KM 面板

进入结果中心
KM Figure4 个面板

Figure 2.

Treatment outcomes (event) stratified by DeSTIL in the NSABP B-41 clinical trial, with EFS shown on the x -axis. Row 1: Kaplan–Meier curves for ( A ) DeSTIL-positive and ( B ) DeSTIL-negative patient groups comparing trastuzumab and the combination therapy, respectively. Row 2: Kaplan–Meier curves for ( C ) DeSTIL-positive and ( D ) DeSTIL-negative patient groups comparing trastuzumab- and lapatinib-containing regimens (lapatinib alone or combination therapy). The x -axis represents EFS, whereas the y -axis represents survival probability, and patients without events were censored at last follow-up.

A

EFS

Figure 2. · Panel A

complete 待 EXE 对齐 预期 2 条 已提取 2 条

Panel A 是同一 Figure 2. 中的独立 KM 子图;人群、终点或分析条件以本 Panel 图注语义为准。

质量提示曲线 1:flat topline artifact
查看结果
B

EFS

Figure 2. · Panel B

quality_warning 待 EXE 对齐 预期 2 条 已提取 1 条

Panel B 是同一 Figure 2. 中的独立 KM 子图;人群、终点或分析条件以本 Panel 图注语义为准。

质量提示疑似重复轨迹:同一条 KM 曲线可能被重复识别疑似图例或人工水平线:轨迹出现不合理的共同平台曲线缺失:预期 2 条,实际恢复 1 条疑似重复轨迹:同一条 KM 曲线可能被重复识别
查看结果
C

EFS

Figure 2. · Panel C

complete 待 EXE 对齐 预期 2 条 已提取 2 条

Panel C 是同一 Figure 2. 中的独立 KM 子图;人群、终点或分析条件以本 Panel 图注语义为准。

质量提示曲线 1:flat topline artifact
查看结果
D

EFS

Figure 2. · Panel D

quality_warning 待 EXE 对齐 预期 2 条 已提取 1 条

Panel D 是同一 Figure 2. 中的独立 KM 子图;人群、终点或分析条件以本 Panel 图注语义为准。

质量提示疑似重复轨迹:同一条 KM 曲线可能被重复识别疑似图例或人工水平线:轨迹出现不合理的共同平台曲线缺失:预期 2 条,实际恢复 1 条疑似重复轨迹:同一条 KM 曲线可能被重复识别
查看结果
其他 Figure(4,默认折叠)

这些图片未产生 KM 曲线结果,不在主页面重复展示。

Figure 1.DeSTIL analytic pipeline. A, Tissue acquisition and WSI digitization; ( B ) whole-slide quality control, patch extraction, and stain normalization; ( C ) nuclei segmentation; ( D ) identification of TILs; ( E ) extraction of TIL density and spatial architecture features; ( F ) DeSTIL feature selection and model development using TCGA-derived features from ( E ), followed by independent validation in the NSABP B-41 clinical trial cohort for EFS across treatment arms; and ( G ) training of an exploratory classifier using features from the UH cohort to predict pCR, followed by independent validation in the NSABP B-41 clinical trial cohort. RF, random forest; SVM, support vector machine. [Created in BioRender. Bharadwaj, S. (2026) https://BioRender.com/jay0g17 ].
非 KM Figure
Figure 3.Representative WSIs showing differences in TIL cluster density and area between DeSTIL-negative and DeSTIL-positive tumors. A, Heatmaps of TIL cluster density overlaid on WSIs from DeSTIL-negative and DeSTIL-positive cases, with red indicating high and blue indicating low TIL cluster density. Right, corresponding image patches from high- and low-density regions. B, Heatmaps of TIL cluster area on the same WSIs, with blue indicating high and orange indicating low TIL cluster area. Right, patches from representative high- and low-area regions. [Created in BioRender. Bharadwaj, S. (2026) https://BioRender.com/umhdruc ].
非 KM Figure
Figure 4.Immune gene expression differences support spatial features captured by the DeSTIL signature. A, Boxplots of BC360-inferred gene signatures show elevated expression of CD8 T cells, cytotoxic cells, IFNγ, and PD-1 in DeSTIL-negative tumors, whereas B7-H3 is elevated in DeSTIL-positive tumors. These signatures were selected based on a significance of P < 0.05 and did not remain significant after correction for multiple hypothesis testing. B, GSEA across trastuzumab and combination therapy arms demonstrates pathway-level immune activity patterns consistent with image-derived features, including differences in TIL cluster density and area between DeSTIL groups. NES, normalized enrichment score.
非 KM Figure
Figure 5.Performance of the locked pCR classifier across treatment arms in the NSABP B-41 cohort. A, The classifier was trained on the UH cohort comprising patients treated with TCH and applied without further tuning to the NSABP B-41 external test set ( B ). ROC curves and corresponding AUCs are shown separately for trastuzumab, lapatinib, and trastuzumab plus lapatinib arms to assess discrimination across HER2-targeted treatment contexts. C, Features that constitute the pCR classifier and the differences in feature values between patients who achieved pCR and non-pCR; each dot represents an individual patient’s feature value.
非 KM Figure