bioRxiv [Preprint]. 2026 Mar 2:2026.02.27.708546. doi: 10.64898/2026.02.27.708546. ABSTRACT Extrachromosomal DNA (ecDNA) is a major driver of oncogene amplification, tumour heterogeneity and poor clinical outcomes [1-3], yet its detection relies on specialised genomic assays that are not integrated into routine diagnostics. Here, we show that ecDNA status can be inferred directly from standard haematoxylin and eosin-stained whole-slide pathology images. We develop an end-to-end, weakly supervised deep learning framework that aggregates thousands of high-magnification patches per slide with slide-level augmentation and interpretable attention. Across twelve cancer types from The Cancer Genome Atlas, the approach identifies tumours with genomic amplifications and, critically, distinguishes ecDNA-amplified from chromosomally amplified or non-amplified tumours, with the strongest signal in glioblastoma. Attention maps localise regions enriched for nuclei with altered chromatin intensity and texture, and predicted ecDNA status recapitulates its adverse association with survival. These results indicate that ecDNA amplifications leave reproducible histomorphologic foot-prints detectable by routine pathology, enabling scalable screening to prioritise tumours for confirmatory molecular testing. PMID:41847039 | PMC:PMC12991143 | DOI:10.64898/2026.02.27.708546

Nat Commun. 2025 Dec 10. doi: 10.1038/s41467-025-67257-8. Online ahead of print. ABSTRACT Estrogen receptor-positive breast cancer remains a leading cause of cancer-related death in women, with mortality largely driven by late recurrence of treatment-resistant disease. Loss of MLH1 promotes resistance to estrogen-targeting therapies by uncoupling cell cycle progression from estrogen regulation. Here, we show that even when MLH1 is abundantly expressed, aberrant cytoplasmic localization in a subset of tumor cells drives endocrine therapy resistance by enabling estrogen-independent growth. This resistance arises from failure to undergo robust cell cycle arrest in response to endocrine therapy, creating acute dependency on CDK4/6 activity. Consequently, CDK4/6 inhibitors induce strong regression in cells with cytoplasmic MLH1 compared to cells with nuclear MLH1. As cytoplasmic localization occurs in ~11% of ER+ patients, it represents a contributor to MLH1 dysregulation. Incorporating cytoplasmic MLH1 localization into diagnostics could guide the use of CDK4/6 inhibitors in this hard-to-treat subset. PMID:41372237 | DOI:10.1038/s41467-025-67257-8