Computer Vision–Based Analysis of Adipocyte–Tumour Associations Reveals Tumour Type- and Obesity-Specific Adaptation

Authors: M. E. Lien, G. Hartel, K. A. Tran, N. Waddell, S. S. Brownrigg, K. Bouttle, T. O’Mara, N. H. Halberg.

• Introduced AdiFind, a deep-learning, tumour-aware whole-slide image pipeline for adipocyte segmentation, morphometric quantification, and peritumoural spatial profiling on routine H&E histology.
• Quantified more than 75 million adipocytes across TCGA, the Australian Breast Cancer Tissue Bank, and the Susan G Komen Tissue Bank.
• Identified tumour-proximal adipocyte size as a breast cancer-specific prognostic signal, strongest in postmenopausal patients with overweight or obesity.