Breast cancer, with 1.7 million newly diagnosed patients each year [1], is the leading cause of cancer-related deaths among women worldwide [2] and has become a public health problem threatening women's health. Therefore, the accurate diagnosis of breast cancer is particularly important. For therapeutic purposes, the four immunohistochemical (IHC) biomarkers that define the clinicopathological subtypes in breast cancer cells include the estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), and Ki-67, as recommended by the 12th International Breast Cancer Conference [3]. In accordance with gene expression measurements, breast cancer can be classified into different subtypes: luminal A, luminal B, human epidermal growth factor receptor 2 (HER2)-enriched, and triple-negative breast cancer (TNBC) [4]. As a heterogeneous disease, breast cancer has different etiologies, prognoses and treatment responses [5]. The various molecular subtypes have distinct individual treatment, recurrence and survival rates. Therefore, accurate diagnosis of breast cancer and obtaining breast cancer subtypes and immunohistochemical factor status for breast cancer patients are conducive to selecting the optimal individualized treatment plan.

IVIM imaging can sensitively reflect both molecular diffusion in tissues and random flow of blood in capillaries [6]. In addition, DKI considers the non-Gaussian diffusion of water molecules inside tissues and helps to show the interaction of water molecules with tissue features, especially cell membranes, and can accurately assess the microstructural complexity of tissues [7]. It has been reported to identify breast cancer and predict prognostic factors by combining IVIM and DKI sequences [8,9]. Nevertheless, breast cancer is a complex and heterogeneous disease [3], and the mean or median values of IVIM and DKI parameters based on a region of interest (ROI) of breast lesions may dilute or even cover up small but important differences between different lesion entities and do not reflect the overall heterogeneity of the lesion. Histogram analysis is a mathematical method used to evaluate tumor heterogeneity without additional imaging. Histogram analysis of the spatial distribution of magnetic resonance (MR) parameters can obtain more information on tumor heterogeneity, such as skewness and kurtosis [10,11]. However, there are few studies using histogram analysis of these advanced diffusion sequences, and the study on breast cancer molecular subtypes is not comprehensive.

Chemical exchange saturation transfer (CEST) imaging reflects a novel contrast mechanism, relying on the exchange between mobile protons in amide (-NH), amine (−NH2) and hydroxyl (-OH) groups and bulk water. APTW, a subtype of CEST imaging, generates the CEST effect at 3.5 ppm away from the water signal [12]. Recently, several preliminary studies that applied APTWI to breast malignancies also demonstrated the potential for tumor detection, characterization, and treatment assessment [[13], [14], [15], [16], [17]].

Based on the above, we hypothesized that the pathological types of breast lesions can be distinguished using advanced indicators of histogram analysis based on APTWI parameters. Accordingly, the purpose of this study was to evaluate the histogram features of APTWI, DKI and IVIM parameters to distinguish malignant from benign breast lesions and to identify the molecular subtypes of breast cancers.