Abstract P2-21-01: Decoding Inter- and Intra-Tumor Heterogeneity in Lobular Breast Cancer Using Spatial Transcriptomics and Clustering Analysis

Abstract

Abstract Background: Invasive lobular breast carcinoma (ILC) represents 5 to 15% of all invasive breast cancers. Recent studies showed the importance of tumor microenvironment (TME) heterogeneity on patient outcome. Here, we aim to characterize TME spatial heterogeneity by performing clustering analysis on spatial transcriptomics (ST) data. Methods: Frozen tumor samples from 43 primary estrogen receptor positive, HER2-negative ILCs were characterized using ST (Visium, 10x Genomics), each ST slide containing 4992 spots. Hematoxylin/eosin (H&E) stained ST slides were annotated (QuPath software) reaching single cell resolution. After performing normalization, hierarchical clustering (STutility R package) across all samples was carried out on principal components computed using highly variable genes. Clusters were characterized using morphological annotation and gene set enrichment analysis for hallmark gene sets from MSigDB (FGSEA R package). A cluster of spots was defined as tumoral or stromal if the average proportion of pixels annotated as tumor or stroma across its constitutive spots was higher than the average proportion of tumor or stroma pixels across all spots of our cohort. Spatial heterogeneity was assessed by comparing the number of contacts between spots belonging to the same cluster (homo-contacts) and the number of contacts between spots belonging to different clusters (hetero-contacts). Comparisons between groups were assessed using Wilcoxon test. Results: Out of the 43 ILC samples, 19 were T2 or T3, 13 were node-positive and 34 were grade 2. Of note, 9 patients experienced disease relapse. Morphological annotation revealed that an average of 20.4%, 61.12%, 11.5%, 0.45%, 3% of the tissues in our dataset corresponded to tumor, stroma, adipose tissue, immune infiltrate and normal structures (vessels, normal breast), respectively. Bioinformatics analysis revealed 7 tumor, 11 stroma, and 6 normal structures clusters, as well as 8 mixed clusters with no predominant morphological structure, with a median of 22 clusters per sample. Tumor and stroma clusters were either shared across all samples or present only in specific samples. Overall, tumor clusters were characterized by an enrichment in estrogen and androgen response related pathways. Moreover, tumor clusters enriched in oxidative phosphorylation, G2M checkpoint and MYC targets were more present in samples with higher histopathological grade (p=0.016), whereas tumor clusters enriched in interferon alpha/gamma response related pathway were associated with a higher tumor stage (p=0.007). A higher number of hetero-contacts among tumor spots were associated with disease relapse (p=0.02). Similarly, a higher number of hetero-contacts among stroma spots including immune and adipose related clusters was also found in samples from patients who experienced disease relapse (p=0.01). Overall, these findings suggest a role of both tumor and stroma spatial disorganization and heterogeneity in tumor progression. Furthermore, clusters capturing the presence of normal breast and in situ carcinoma were enriched in samples from patients who did not relapse (p< 0.001). Conclusion: Our results revealed the substantial inter- and intra-patient heterogeneity of ILC both at the tumor and microenvironment levels. Different tumor and stroma clusters characterized by specific hallmarks were associated to specific clinical features and disease outcome, unraveling potential new targets for optimizing ILC care. Further validation is needed. Citation Format: Matteo Serra, Mattia Rediti, Frédéric Lifrange, David Venet, Nicola Occelli, Laetitia Collet, Delphine Vincent, Ghizlane Rouas, Ligia Craciun, Denis Larsimont, Miikka Vikkula, Francois P. Duhoux, Françoise Rothé, Christos Sotiriou. Decoding Inter- and Intra-Tumor Heterogeneity in Lobular Breast Cancer Using Spatial Transcriptomics and Clustering Analysis [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P2-21-01.