Breast cancer is the most prevalent malignancy and the second-highest contributor to cancer-related mortality among women worldwide [1]. Tumor hypoxia is observed in approximately 40 % of breast cancer cases and is strongly associated with therapeutic resistance and poor prognosis [[2], [3], [4]]. Prolonged exposure to a hypoxic environment results in partial cell death and promotes genetic and adaptive changes in surviving cells [5]. Under hypoxic conditions, hypoxia-inducible factor-1α (HIF-1α) is activated and drives angiogenesis, metabolic reprogramming, and metastasis [[6], [7], [8]]. Therefore, evaluating the degree of hypoxia and monitoring the expression of HIF-1α in breast cancer are critical for predicting prognosis and assessing treatment response.

HIF-1α expression is traditionally measured using biopsy samples; however, sampling bias due to intratumoral hypoxia heterogeneity is a significant limitation [9]. Magnetic resonance imaging (MRI) provides versatile multi-sequence imaging options and excellent soft-tissue contrast, and it is thus extensively utilized in controlled studies focusing on imaging pathology [[10], [11], [12]]. In previous studies, MRI and its derived techniques have been used to assess hypoxia in malignant tumors [13]. Conventional MRI provides information on the morphological characteristics of tumors [14]. Dynamic Contrast Enhanced MRI (DCE-MRI) enables quantitative assessment of tumor hemodynamics, including blood flow, volume, and vascular permeability [15]. Diffusion Kurtosis Imaging (DKI) captures microstructural complexity by assessing non-Gaussian water diffusion patterns [16]. Additionally, although radiomics has been applied to predict HIF-1α expression, most studies have been limited to single-modality imaging data and exhibit poor clinical interpretability [17]. However, tumor hypoxia arises from dynamic interactions among cellular proliferation, angiogenesis, and metabolic adaptation [18]. Single parameters are inadequate as sensitive and reliable markers of tumor hypoxia status. By contrast, integrating multimodal MRI techniques with clinical-pathological examinations offers a comprehensive approach to analyze tumor hypoxia status [19,20].

This study examined the association of clinical-pathological and multimodal MRI features with the expression levels of HIF-1α in breast cancer. These associations were used to develop and validate a comprehensive combined model integrating imaging biomarkers with clinical-pathological parameters, thus providing a novel and non-invasive method for accurately assessing hypoxia status in breast tumors. Such a model could potentially facilitate the design of personalized therapeutic strategies and improve prognostic prediction in clinical practice.