Integration of Insomnia-Related Network Perturbation Reveals Distinct Molecular Subtypes and Therapeutic Vulnerabilities in Breast Cancer

Abstract

<title>Abstract</title> Core circadian regulatory genes orchestrate vital physiological processes through temporal coordination of organ function, cellular metabolism, and mitotic regulation, critically influencing oncogenic transformation. Emerging evidence implicates circadian disruption as both prognostic determinant and therapeutic modulator in carcinogenesis. Leveraging genomic insights from 7,868 breast cancer patients, we established a sleep perturbation network incorporating 956 insomnia-associated genes, identifying 443 prognostic gene pairs (P-value&lt;0.05) through integrative matrix analysis. Our molecular stratification system delineated three clinically distinct subtypes with differential outcomes: Cluster 1 demonstrated poorest prognosis with paradoxical immunoevasive hyper-infiltration, whereas Cluster 3 exhibited favorable survival despite "immune cold" characteristics. The transitional Cluster 2 phenotype showed extracellular matrix remodeling signatures. The XGBoost-derived classifier achieved superior predictive performance (Macro-AUC=0.9476; Micro-AUC=0.9453) across validation cohorts, confirming robust generalizability. Through systematic evaluation of 76 machine learning ensembles, we developed a precision risk model stratifying patients into distinct prognostic groups. Low-risk patients demonstrated significantly prolonged survival, validated through multi-center cohorts. The model's discriminant capacity persisted across molecular subtypes and treatment regimens.

Affiliated Institutions

• Third Affiliated Hospital of Harbin Medical University CN
• Second Military Medical University CN
• Harbin Medical University CN
• Changhai Hospital CN
• Fourth Affiliated Hospital of Harbin Medical University CN

Related Publications