Multi-modal analysis of breast cancer biopsies before and after T-DXd and Sac-Gov treatment reveals inflamed tumor status as a major correlate of sensitivity to Topo-I ADCs
SITC 2026
Michael A Smith, Keith H Jansson, Anoek Zomer, Mina Khoshdeli, Michela Capello, Merzu K Belete, David Soong, Nora Pencheva, Sriram Sridhar
Background
Topoisomerase I inhibitor antibody-drug conjugates (Topo-I ADCs), including trastuzumab deruxtecan (T-DXd) and sacituzumab govitecan (Sac-Gov), have transformed the treatment of breast cancer, yet clinical features of sensitivity and resistance remain incompletely defined. Therefore, we performed multi-modal profiling of pre-, on-, and post-treatment biopsies from a real-world cohort of breast cancer patients treated with T-DXd or Sac-Gov monotherapy to identify tumor and microenvironmental features associated with clinical benefit.
Methods
RNA, DNA sequencing and digital pathology analysis were performed on unpaired breast cancer biopsies collected ≤6 months before, during, or ≤6 months after treatment with T-DXd or Sac-Gov, excluding samples with intervening therapies between biopsy collection and ADC treatment. The dataset included 98 pre-treatment, 11 on-treatment, and 43 post-treatment biopsies from T-DXd-treated patients, and 57 pre-treatment, 11 on-treatment, and 39 post-treatment biopsies from Sac-Gov-treated patients. Differential gene expression (DESeq2), pathway enrichment and gene set scoring (GSEA, GSVA), immune deconvolution (xCell2), literature-derived immune signature scoring, digital pathology cell quantification (HistoPlus), and survival analyses (Kaplan-Meier, Cox proportional hazards models) were used to identify baseline correlates of response, biomarkers of progression-free survival, and changes associated with ADC exposure.
Results
Across both treatment cohorts, orthogonal measures of inflamed tumor status were associated with sensitivity to T-DXd and Sac-Gov. CD8 T cell enrichment scores correlated with improved PFS to both ADCs, while antigen-presenting cell-related markers and signatures were additionally associated with improved PFS in Sac-Gov-treated patients. Digital pathology analyses confirmed increased lymphocyte infiltration in responders at baseline, supporting the transcriptomic association between an inflamed tumor microenvironment and ADC sensitivity.
Because immune-cold tumors may reflect stromal exclusion, we evaluated fibroblast abundance, extracellular matrix deposition, and angiogenesis using both RNA sequencing and digital pathology. These features were not consistently associated with inflamed tumor status. Instead, immune-cold status was most strongly associated with the Guégan et al. 25-gene exhaustion-associated transcriptomic score, suggesting an immune-resistance program beyond stromal exclusion.
On-treatment Sac-Gov biopsies showed lower CD8 T cell and T-Cell-Inflamed-Gene-Expression-Profile scores than pre-treatment biopsies, whereas post Sac-Gov biopsies were comparable to pre-treatment levels. No significant differences were observed with T-Dxd, although on-treatment samples also trended toward lower inflammation scores in this cohort.
Conclusions
Multi-modal profiling of a large real world breast cancer cohort identified inflamed tumor status as a correlate of sensitivity to T-DXd and Sac-gov and supports further evaluation of immune augmenting strategies in combination with ADC therapy.
Related publications