Sleep patterns and breast cancer risk: a prospective cohort study integrating multi-omics and machine learning
- 1Department of Breast Surgery, The Second Hospital of Jilin University
Abstract
Background Sleep is a potentially modifiable behavioural factor that may influence breast cancer risk. This study examined the association between sleep patterns and incident breast cancer in relation to genetic susceptibility and plasma proteomic profiles. Methods This prospective cohort study included 159,452 UK Biobank women free of breast cancer at baseline. Sleep patterns were constructed from five sleep characteristics. Associations with incident breast cancer were assessed using Cox proportional hazards models and time-varying Cox proportional hazards models adjusted for demographic, socioeconomic, lifestyle, breast cancer-related, and clinical covariates. Kaplan-Meier curves, restricted cubic splines, and PRS-based joint and stratified analyses were performed. In the proteomic subcohort, differential expression, enrichment, and machine-learning-based feature prioritisation analyses were used to identify overlapping proteomic signatures related to sleep pattern and breast cancer status. Results During follow-up, 7,291 incident breast cancer cases were identified. Compared with a healthy sleep pattern, an unhealthy sleep pattern was associated with a higher risk of breast cancer in the fully adjusted model (HR = 1.10, 95% CI = 1.05–1.15). This association was consistent in time-varying Cox proportional hazards models (HR = 1.34, 95% CI = 1.02–1.77). Joint analyses showed that participants with both high PRS and an unhealthy sleep pattern had the highest breast cancer risk. Proteomic analyses identified overlapping protein signatures related to both sleep pattern and breast cancer status, with enrichment mainly in immune regulation, inflammatory signalling, vascular function, and extracellular matrix-related biological processes. Machine-learning-based feature prioritisation identified CRISP3, PADI2, PLA2G2A, SUSD4, TNFRSF6B, and UPK3A as key candidate proteins. Conclusions Unhealthy sleep patterns were associated with increased incident breast cancer risk. Integrated genetic and proteomic analyses suggest that immune-inflammatory, vascular, and extracellular matrix-related processes may provide biological context for this association.
Related
- Integrating disulfidptosis patterns and machine learning identifies a novel diagnostic signature for endometriosis
- Explainable machine learning for breast cancer prediction in resource-constrained settings: A multi-algorithmic framework integrating shap-based transparency with clinical decision support
- Association of the metabolic signature of sleep patterns with breast cancer risk: a prospective cohort study
- Letter to the Editor: Integrating Machine Learning Into Stroke Risk Stratification in AF—Authors' Reply
- Development and interpretable validation of a machine-learning model for edentulism risk in US adults : A cross-sectional study integrating five algorithms and dietary data from NHANES 2009–2014
BibTeX
@article{chang2026sleep,
title = {{Sleep patterns and breast cancer risk: a prospective cohort study integrating multi-omics and machine learning}},
author = {Xiwen Chang and Shuting Zuo and Zhenyu Wang and Zhen Wang and Yafang Gao and Yan Zhang},
journal = {BMC Cancer},
year = {2026},
month = jul,
doi = {10.1186/s12885-026-16641-8},
url = {https://doi.org/10.1186/s12885-026-16641-8},
note = {CC BY},
}
RIS
TY - JOUR
TI - Sleep patterns and breast cancer risk: a prospective cohort study integrating multi-omics and machine learning
AU - Xiwen Chang
AU - Shuting Zuo
AU - Zhenyu Wang
AU - Zhen Wang
AU - Yafang Gao
AU - Yan Zhang
PY - 2026
DA - 2026/07/01
JO - BMC Cancer
DO - 10.1186/s12885-026-16641-8
UR - https://doi.org/10.1186/s12885-026-16641-8
AB - Background Sleep is a potentially modifiable behavioural factor that may influence breast cancer risk. This study examined the association between sleep patterns and incident breast cancer in relation to genetic susceptibility and plasma proteomic profiles. Methods This prospective cohort study included 159,452 UK Biobank women free of breast cancer at baseline. Sleep patterns were constructed from five sleep characteristics. Associations with incident breast cancer were assessed using Cox proportional hazards models and time-varying Cox proportional hazards models adjusted for demographic, socioeconomic, lifestyle, breast cancer-related, and clinical covariates. Kaplan-Meier curves, restricted cubic splines, and PRS-based joint and stratified analyses were performed. In the proteomic subcohort, differential expression, enrichment, and machine-learning-based feature prioritisation analyses were used to identify overlapping proteomic signatures related to sleep pattern and breast cancer status. Results During follow-up, 7,291 incident breast cancer cases were identified. Compared with a healthy sleep pattern, an unhealthy sleep pattern was associated with a higher risk of breast cancer in the fully adjusted model (HR = 1.10, 95% CI = 1.05–1.15). This association was consistent in time-varying Cox proportional hazards models (HR = 1.34, 95% CI = 1.02–1.77). Joint analyses showed that participants with both high PRS and an unhealthy sleep pattern had the highest breast cancer risk. Proteomic analyses identified overlapping protein signatures related to both sleep pattern and breast cancer status, with enrichment mainly in immune regulation, inflammatory signalling, vascular function, and extracellular matrix-related biological processes. Machine-learning-based feature prioritisation identified CRISP3, PADI2, PLA2G2A, SUSD4, TNFRSF6B, and UPK3A as key candidate proteins. Conclusions Unhealthy sleep patterns were associated with increased incident breast cancer risk. Integrated genetic and proteomic analyses suggest that immune-inflammatory, vascular, and extracellular matrix-related processes may provide biological context for this association.
SN - 1471-2407
ER -