Introduction:

Research on ductal carcinoma in situ (DCIS) has largely focused on tumour features, while adipose tissue within the tumour microenvironment remains comparatively understudied. Adipocytes actively secrete hormones, influence metabolic signalling, and undergo structural changes with ageing and obesity. These are factors increasingly linked to breast cancer. Enlarged adipocytes and altered fat composition may contribute to more aggressive phenotypes and poorer outcomes. With global obesity rates rising, analysing adipose tissue in whole slide images (WSIs) may clarify metabolic contributions to breast cancer risk.

 

Methods:

We applied a morphology-based machine learning approach to analyse adipocyte structure across Canadian breast cancer cohorts. After artifact removal, adipocyte patches were extracted from WSIs to compute mean and variance of adipocyte size at both slide and patch levels. Using unsupervised clustering, we generated ten distinct adipocyte morphology prototypes, whose stability was validated through bootstrapping. The prototypes were then projected back onto WSIs to evaluate spatial trends relative to tumour boundaries using concentric distance measurements.

​

Results:

Adipocyte size varied with age across all cohorts, with samples grouping into the ten different morphology prototypes. Individuals aged 55 and older displayed the largest adipocytes, while those 45 and younger had the smallest. All age group comparisons showed significant differences, with the strongest separation observed between the youngest and oldest groups. Older patients also exhibited greater variability in adipocyte size, reflecting increased tissue heterogeneity. 

​

Conclusion:

These findings support age‑related adipocyte enlargement as a potential contributor to breast cancer risk and underscore the value of incorporating adipose morphology into DCIS pathology research.

​

References:

[1] Wu C, et al. Cancer-Associated Adipocytes and Breast Cancer: Intertwining in the Tumor Microenvironment and Challenges for Cancer Therapy. Cancers. 2023; 15(3):726. https://doi.org/10.3390/cancers15030726

[2] Zhu Q, et al. Adipocyte mesenchymal transition contributes to mammary tumor progression. Cell Reports. 2022; 40(1). https://doi.org/10.1016/j.celrep.2022.111362

[3] Martel AL, et al. Assessment of Residual Breast Cancer Cellularity after Neoadjuvant Chemotherapy using Digital Pathology [Data set]. The Cancer Imaging Archive. 2019. https://doi.org/10.7937/TCIA.2019.4YIBTJNO

[4] Chen RJ, et al. Towards a general-purpose foundation model for computational pathology. Nat Med. 2024; 30, 850–862. https://doi.org/10.1038/s41591-024-02857-3

[5] Song AH, et al. “Morphological Prototyping for Unsupervised Slide Representation Learning in Computational Pathology.” 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2024; 11566-11578.

[6] Tang L, et al. Corrigendum: Diversity and heterogeneity in human breast cancer adipose tissue revealed at single-nucleus resolution. Front. Immunol. 2023; 14:1213786. doi: 10.3389/fimmu.2023.1213786

[7] Li J, et al. MiCo: Multiple Instance Learning with Context-Aware Clustering for Whole Slide Image Analysis. In: Gee, J.C., et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2025; Lecture Notes in Computer Science, vol 15960. Springer, Cham. https://doi.org/10.1007/978-3-032-04927-8_36

[8] Almekinders MMM, et al. Breast adipocyte size associates with ipsilateral invasive breast cancer risk after ductal carcinoma in situ. npj Breast Cancer. 2021; 7, 31. https://doi.org/10.1038/s41523-021-00232-w

[9] Dodson A, et al. A compendium of adipocyte morphologies across different breast pathologies. Adipocyte. 2025; 14(1). https://doi.org/10.1080/21623945.2025.2568540