AI-guided deep visual proteomics reveals recurrence-associated morpho-molecular signatures within triple-negative breast cancers
Abstract
Outcome informed AI can localize prognostically relevant regions in digital pathology, but directly characterizing the molecular basis of these spatial predictions remains challenging. Here, we developed an outcome informed, AI-guided deep visual proteomics framework that uses recurrence-risk heatmaps to guide physical tissue sampling and mass spectrometry-based spatial proteomics in triple-negative breast cancer (TNBC). In 156 patients, high- and low-risk regions coexisted within the tumor compartment and showed distinct nuclear and architectural features, indicating intratumoral variation beyond broad tissue identity. Guided by these maps, we physically isolated and profiled 46 AI-defined high- and low-risk tumor regions from two recurrent cases. High-risk regions were enriched for mitotic programs, whereas low-risk regions showed immune- and antigen-presentation-associated programs. We used proteins that consistently differed between high- and low-risk regions to construct a 13-protein composite. The composite showed a significant effect to discriminate highly recurrent patients from less recurrent patients in an expanded TNBC cohort, while the corresponding transcript-based score showed a similar trend in METABRIC. Integrating the protein composite with the H&E-derived risk score improved out-of-bag discrimination, increasing the C-index from 0.679 to 0.739 and the 5-year time-dependent AUC from 0.675 to 0.770. These findings establish outcome-informed H&E risk maps as spatial guides for molecular discovery, extending prognostic AI from prediction toward direct biological interrogation of recurrence-associated tissue states.
Related
- Plasma proteomics identifies an endoplasmic reticulum stress-associated molecular signature across cardiovascular-kidney-metabolic syndrome stages
- Salivary prognostic bacterial signatures associated with survival outcomes and recurrence risk in head and neck squamous cell carcinoma
- Identification of Crohn’s disease-associated molecular signatures by integrative bioinformatics analysis and experimental validation
- Integrating physiological and proteomics signatures for understanding drought, salinity and alkalinity stress tolerance mechanisms in lentil
- Integrated Multi-Omics Analysis Reveals Molecular Signatures Associated with Growth Divergence in Mandarin Fish (Siniperca chuatsi).
BibTeX
@misc{cho2026aiguided,
title = {{AI-guided deep visual proteomics reveals recurrence-associated morpho-molecular signatures within triple-negative breast cancers}},
author = {Yesung Cho and Ji Hwan Park and Chan Il Kim and Hyewon Kim and Honglan Li and Yumin Lee and Geongyu Lee and Sujeong Hong and Seong Min Park and Yoonyoung Lee and Hee Sool Rho and Sumin Lee and Amos Chungwon Lee and Changhwan Lee and Hwanyoung Shim and Hyunwook Kim and Hyeji Shin and Sanha Park and Jihoon Yu and Yoon Hee Shin and Sooheon Kim and Hyunjin Park and Seung Min Park and Sangwan Kim and Yujung Kim and Jinwon Jeong and Sung-Im Do and Eun-Young Kim and Dongmyung Shin and Ingu Do and Jongbae Park},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2608.03145},
eprint = {2608.03145},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.03145},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - AI-guided deep visual proteomics reveals recurrence-associated morpho-molecular signatures within triple-negative breast cancers
AU - Yesung Cho
AU - Ji Hwan Park
AU - Chan Il Kim
AU - Hyewon Kim
AU - Honglan Li
AU - Yumin Lee
AU - Geongyu Lee
AU - Sujeong Hong
AU - Seong Min Park
AU - Yoonyoung Lee
AU - Hee Sool Rho
AU - Sumin Lee
AU - Amos Chungwon Lee
AU - Changhwan Lee
AU - Hwanyoung Shim
AU - Hyunwook Kim
AU - Hyeji Shin
AU - Sanha Park
AU - Jihoon Yu
AU - Yoon Hee Shin
AU - Sooheon Kim
AU - Hyunjin Park
AU - Seung Min Park
AU - Sangwan Kim
AU - Yujung Kim
AU - Jinwon Jeong
AU - Sung-Im Do
AU - Eun-Young Kim
AU - Dongmyung Shin
AU - Ingu Do
AU - Jongbae Park
PY - 2026
DA - 2026/10/27
PB - arXiv
DO - 10.48550/arxiv.2608.03145
UR - https://arxiv.org/abs/2608.03145
AB - Outcome informed AI can localize prognostically relevant regions in digital pathology, but directly characterizing the molecular basis of these spatial predictions remains challenging. Here, we developed an outcome informed, AI-guided deep visual proteomics framework that uses recurrence-risk heatmaps to guide physical tissue sampling and mass spectrometry-based spatial proteomics in triple-negative breast cancer (TNBC). In 156 patients, high- and low-risk regions coexisted within the tumor compartment and showed distinct nuclear and architectural features, indicating intratumoral variation beyond broad tissue identity. Guided by these maps, we physically isolated and profiled 46 AI-defined high- and low-risk tumor regions from two recurrent cases. High-risk regions were enriched for mitotic programs, whereas low-risk regions showed immune- and antigen-presentation-associated programs. We used proteins that consistently differed between high- and low-risk regions to construct a 13-protein composite. The composite showed a significant effect to discriminate highly recurrent patients from less recurrent patients in an expanded TNBC cohort, while the corresponding transcript-based score showed a similar trend in METABRIC. Integrating the protein composite with the H&E-derived risk score improved out-of-bag discrimination, increasing the C-index from 0.679 to 0.739 and the 5-year time-dependent AUC from 0.675 to 0.770. These findings establish outcome-informed H&E risk maps as spatial guides for molecular discovery, extending prognostic AI from prediction toward direct biological interrogation of recurrence-associated tissue states.
ER -