Kinetic heterogeneity is associated with axillary lymph node metastasis in cN0 breast cancer based on dynamic contrast-enhanced magnetic resonance imaging radiomics nomogram.

BackgroundDetection of metastases in axillary lymph nodes (ALNs) is of vital significance for determining appropriate therapeutic strategies and prognosis for breast cancer patients. Studies combining multiparametric magnetic resonance imaging (MRI) and pathological biomarkers for predicting ALN metastasis in breast cancer are rarely reported. This study aimed to evaluate the predictive value of conventional MRI features, intravoxel incoherent motion (IVIM), quantitative dynamic contrast-enhanced MRI (DCE-MRI), and pathological biomarkers for ALN metastasis in breast cancer patients.MethodsIn total, 149 subjects with breast cancer confirmed via pathology were recruited for study. Among the participants, patients were randomly allocated to the training cohort (42 and 62 presented with ALN and non-ALN metastasis) or validation cohort (18 and 27 presented with ALN and non-ALN metastasis), respectively. All participants underwent both IVIM and DCE-MRI. The analysis focused on the clinicopathological characteristics along with conventional MRI features, in addition to assessment of a range of quantitative parameters, including DCE-MRI derived parameters (Ktrans, Kep and Ve), and the IVIM-derived parameter [apparent diffusion coefficient (D), fast apparent diffusion coefficient (D*), perfusion fraction (f)]. To evaluate diagnostic efficacy in predicting ALN metastasis, multivariate logistic regression and receiver operating characteristic (ROC) curve assessments were conducted. A nomogram for the combined model was created on the basis of the findings derived from the multivariate logistic regression model.ResultsIn the training and validation cohorts, patients with ALN metastasis had significantly higher Ki-67 (P=0.01, P=0.03) and hypoxia-inducible factor-1 alpha (HIF-1α) expression (P<0.001, P=0.04). Lymphovascular invasion (LVI) and programmed death ligand-1 (PD-L1) expression were significantly more common in the metastatic group (P=0.002, P=0.003, respectively) in the training cohort. In the training and test cohorts, compared to the non-metastatic group, patients with ALN metastasis exhibited significantly lower D values (all P<0.001) and significantly higher values of D* (P=0.02, P=0.04), Ktrans (all P<0.001), and Kep (all P<0.001). Multivariate analysis identified PD-L1 [odds ratio (OR) =82.55, P=0.045], lesion margin (OR =21.08, P=0.048), D (OR <1,000, P=0.01), and Ktrans (OR >1,000, P=0.01) as independent predictors. Calibration curves confirmed excellent agreement between predicted and observed outcomes (P=0.99). Furthermore, in both training and test validations, the combined model achieved significantly enhanced the areas under the ROC curve (AUCs) compared with the pathologic, conventional MRI, IVIM, and DCE-MRI models (Z=2.083–4.402, P<0.05).ConclusionsCombining MRI parameters (lesion margin, D, Ktrans) with pathological biomarker PD-L1 significantly improves prediction accuracy for ALN metastasis in breast cancer. This integrated model has considerable clinical potential, enabling precise preoperative assessment and potentially reducing unnecessary lymph node biopsies.

Objective To investigate the value of preoperative prediction of breast cancer axillary lymph node metastasis based on intratumoral and peritumoral dynamic contrast enhancement magnetic resonance imaging (DCE-MRI) radiomics nomogram. Material and Methods. In this study, a radiomics model was developed based on a training cohort involving 250 patients with breast cancer (BC) who had undergone axillary lymph node (ALN) dissection between June 2019 and January 2021. The intratumoral and peritumoral radiomics features were extracted from the second postcontrast images of DCE-MRI. Based on filtered radiomics features, the radiomics signature was built by using the least absolute shrinkage and selection operator method. The Support Vector Machines (SVM) learning algorithm was used to construct intratumoral, periatumoral, and intratumoral combined periatumoral models for predicting axillary lymph node metastasis (ALNM) in BC. Nomogram performance was determined by its discrimination, calibration, and clinical value. Multivariable logistic regression was adopted to establish a radiomics nomogram. Results The intratumoral combined peritumoral radiomics signature, which was composed of fifteen ALN status-related features, showed the best predictive performance and was associated with ALNM in both the training and validation cohorts (P < 0.001). The prediction efficiency of the intratumoral combined peritumoral radiomics model was higher than that of the intratumoral radiomics model and the peritumoral radiomics model. The AUCs of the training and verification cohorts were 0.867 and 0.785, respectively. The radiomics nomogram, which incorporated the radiomics signature, MR-reported ALN status, and MR-reported maximum diameter of the lesion, showed good calibration and discrimination in the training (AUC = 0.872) and validation cohorts (AUC = 0.863). Conclusion The intratumoral combined peritumoral radiomics model derived from DCE-MRI showed great predictive value for ALNM and may help to improve clinical decision-making for BC.

Previous studies have explored the potential on radiomics features of primary breast cancer tumor to identify axillary lymph node (ALN) metastasis. However, the value of deep learning (DL) to identify ALN metastasis remains unclear. To investigate the potential of the proposed attention-based DL model for the preoperative differentiation of ALN metastasis in breast cancer on dynamic contrast-enhanced MRI (DCE-MRI). Retrospective. A total of 941 breast cancer patients who underwent DCE-MRI before surgery were included in the training (742 patients), internal test (83 patients), and external test (116 patients) cohorts. A 3.0 T MR scanner, DCE-MRI sequence. A DL model containing a 3D deep residual network (ResNet) architecture and a convolutional block attention module, named RCNet, was proposed for ALN metastasis identification. Three RCNet models were established based on the tumor, ALN, and combined tumor-ALN regions on the images. The performance of these models was compared with ResNet models, radiomics models, the Memorial Sloan-Kettering Cancer Center (MSKCC) model, and three radiologists (W.L., H.S., and F. L.). Dice similarity coefficient for breast tumor and ALN segmentation. Accuracy, sensitivity, specificity, intercorrelation and intracorrelation coefficients, area under the curve (AUC), and Delong test for ALN classification. The optimal RCNet model, that is, RCNet-tumor+ALN , achieved an AUC of 0.907, an accuracy of 0.831, a sensitivity of 0.824, and a specificity of 0.837 in the internal test cohort, as well as an AUC of 0.852, an accuracy of 0.828, a sensitivity of 0.792, and a specificity of 0.853 in the external test cohort. Additionally, with the assistance of RCNet-tumor+ALN , the radiologists' performance was improved (external test cohort, P< 0.05). DCE-MRI-based RCNet model could provide a noninvasive auxiliary tool to identify ALN metastasis preoperatively in breast cancer, which may assist radiologists in conducting more accurate evaluation of ALN status. 3 TECHNICAL EFFICACY: Stage 2.

This study aims to develop and validate a radiomics nomogram based on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to noninvasively predict axillary lymph node (ALN) metastasis in breast cancer. This retrospective study included 263 patients with histologically proven invasive breast cancer and who underwent DCE-MRI examination before surgery in two hospitals. All patients had a defined ALN status based on pathological examination results. Regions of interest (ROIs) of the primary tumor and ipsilateral ALN were manually drawn. A total of 1,409 radiomics features were initially computed from each ROI. Next, the low variance threshold, SelectKBest, and least absolute shrinkage and selection operator (LASSO) algorithms were used to extract the radiomics features. The selected radiomics features were used to establish the radiomics signature of the primary tumor and ALN. A radiomics nomogram model, including the radiomics signature and the independent clinical risk factors, was then constructed. The predictive performance was evaluated by the receiver operating characteristic (ROC) curves, calibration curve, and decision curve analysis (DCA) by using the training and testing sets. ALNM rates of the training, internal testing, and external testing sets were 43.6%, 44.3% and 32.3%, respectively. The nomogram, including clinical risk factors (tumor diameter) and radiomics signature of the primary tumor and ALN, showed good calibration and discrimination with areas under the ROC curves of 0.884, 0.822, and 0.813 in the training, internal and external testing sets, respectively. DCA also showed that radiomics nomogram displayed better clinical predictive usefulness than the clinical or radiomics signature alone. The radiomics nomogram combined with clinical risk factors and DCE-MRI-based radiomics signature may be used to predict ALN metastasis in a noninvasive manner.

Difference of DCE-MRI Parameters at Different Time Points and Their Predictive Value for Axillary Lymph Node Metastasis of Breast Cancer

Background To construct and validate a radiomic-based model for estimating axillary lymph node (ALN) metastasis in patients with breast cancer by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). Methods In this retrospective study, a radiomic-based model was established in a training cohort of 236 patients with breast cancer. Radiomic features were extracted from breast DCE-MRI scans. A method named the least absolute shrinkage and selection operator (LASSO) was applied to select radiomic features based on highly reproducible features. A radiomic signature was built by a support vector machine (SVM). Multivariate logistic regression analysis was adopted to establish a clinical characteristic-based model. The performance of models was analysed through discrimination ability and clinical benefits. Results The radiomic signature comprised 6 features related to ALN metastasis and showed significant differences between the patients with ALN metastasis and without ALN metastasis (P < 0.001). The area under the curve (AUC) of the radiomic model was 0.990 and 0.858, respectively, in the training and validation sets. The clinical feature-based model, including MRI-reported status and palpability, performed slightly worse, with an AUC of 0.784 in the training cohort and 0.789 in the validation cohort. The radiomic signature was confirmed to provide more clinical benefits by decision curve analysis. Conclusions The radiomic-based model developed in this study can successfully diagnose the status of lymph nodes in patients with breast cancer, which may reduce unnecessary invasive clinical operations.

The aim of this study was to evaluate the clinical value of 18F-fluorodeoxyglucose (FDG) uptake and the clinicopathological or immunohistochemical findings of the primary tumor to predict axillary lymph node (ALN) metastasis in breast cancer with clinically negative ALN. This study retrospectively enrolled 104 women (49.43 ± 9.9 years) having breast cancer with clinically negative ALN using all types of preoperative imaging modalities including ultrasonography, FDG positron emission tomography, and magnetic resonance imaging. All cases of breast cancer in this study were proven as invasive ductal carcinoma with ≥1 cm in size. The final diagnosis of ALN status was confirmed by permanent pathology after operation. Among 104 breast cancers with clinically negative ALN, 21 breast cancers (20.2 %) were proven to have ALN metastasis. The ROC curve analysis showed that the best cut-off value of SUVmax for identifying ALN metastasis was 9.8 with 33.3 % sensitivity and 92.8 % specificity (AUC = 0.656; p = 0.027). The multivariable analysis revealed that primary tumors with SUVmax >9.8 (p = 0.011) and D2-40 positivity (p = 0.027) were independently associated with ALN metastasis with odds ratios of 5.516 (CI 1.475-20.6333) and 3.409 (CI 1.154-10.072), respectively. Our study demonstrates that the incidence of ALN metastasis in even rigorously clinically evaluated breast cancer without suspiciously positive ALN is still not negligible, and while a high SUVmax of the primary tumor may be associated with a higher incidence of ALN metastasis in breast cancer with clinically negative ALN, a low SUVmax does not exclude ALN metastasis.

Objective:To establish a radiomics nomogram by integrating clinical risk factors and radiomics features extracted from digital mammography (MG) images for pre-operative prediction of axillary lymph node (ALN) metastasis in breast cancer.Methods:216 patients with breast cancer lesions confirmed by surgical excision pathology were divided into the primary cohort (n = 144) and validation cohort (n = 72). Radiomics features were extracted from craniocaudal (CC) view of mammograms, and radiomics features selection were performed using the methods of ANOVA F-value and least absolute shrinkage and selection operator; then a radiomics signature was constructed with the method of support vector machine. Multivariate logistic regression analysis was used to establish a radiomics nomogram based on the combination of radiomics signature and clinical factors. The C-index and calibration curves were derived based on the regression analysis both in the primary and validation cohorts.Results:95 of 216 patients were confirmed with ALN metastasis by pathology, and 52 cases were diagnosed as ALN metastasis based on MG-reported criteria. The sensitivity, specificity, accuracy and AUC (area under the receiver operating characteristic curve of MG-reported criteria were 42.7%, 90.8%, 24.1% and 0.666 (95% confidence interval: 0.591–0.741]. The radiomics nomogram, comprising progesterone receptor status, molecular subtype and radiomics signature, showed good calibration and better favorite performance for the metastatic ALN detection (AUC 0.883 and 0.863 in the primary and validation cohorts) than each independent clinical features (AUC 0.707 and 0.657 in the primary and validation cohorts) and radiomics signature (AUC 0.876 and 0.862 in the primary and validation cohorts).Conclusion:The MG-based radiomics nomogram could be used as a non-invasive and reliable tool in predicting ALN metastasis and may facilitate to assist clinicians for pre-operative decision-making.Advances in knowledge:ALN status remains among the most important breast cancer prognostic factors and is essential for making treatment decisions. However, the value of detecting metastatic ALN by MG is very limited. The studies on pre-operative ALN metastasis prediction using the method of MG-based radiomics in breast cancer are very few. Therefore, we studied whether MG-based radiomics nomogram could be used as a predictive biomarker for the detection of metastatic ALN.

Objective To investigate the risk factors of axillary lymph node metastasis in patients with invasive breast cancer. Methods This study retrospectively included 122 cases of invasive breast cancer patients admitted to the First Medical Center of PLA General Hospital from January 2019 to September 2020. According to postoperative pathological results, axillary lymph node metastasis was divided into axillary lymph node metastasis (ALNM) group (n =40) and non-axillary lymph node metastasis (NALNM) group (n =82). General demographic information was collected and compared between the two groups. Collected pathological results included lymphovascular invasion (LVI) and the expression of estrogen receptor (ER), progestogen receptor (PR), human epidermal growth factor receptor 2 (HER-2), and Ki-67 detected by immunohistochemistry. Imaging parameters of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) including apparent diffusion coefficient (ADC), early enhanced rate, and time-intensity curve (TIC) were also included into univariate analysis. The variables with differences between the two groups were compared by univariate analysis, and the related factors of axillary lymph node metastasis were analyzed by logistic regression model. Results There was no significant difference in general demographic information between the two groups. No significant differences were found in the positive rates of HER-2, ER, PR, Ki-67, pathological types, and clavicular lymph node metastasis and skin chest wall invasion between the two groups (P > 0.05). The proportion of LVI in ALNM group was significantly higher than that in NALNM group (37.50% vs. 6.10%, P < 0.001). The proportion of breast cancer on the left side in the ALNM group was higher than that in the NALNM group, and the difference was statistically significant (70.00% vs. 47.56%, P = 0.019). There were no significant differences in the imaging parameters obtained by DCE-MRI between the two groups. Binary logistics regression analysis showed that LVI (OR =12.258, 95% CI =3.681-40.812, P < 0.001) and left breast cancer (OR =3.598, 95% CI =1.404-9.219, P = 0.008) were risk factors for axillary lymph node metastasis in patients with invasive breast cancer. Conclusion The formation of vascular tumor thrombi in breast cancer tissue and left breast cancer are risk factors for axillary lymph node metastasis in invasive breast cancer and might be helpful for preoperative detailed assessment of the patient's condition.

This study aims to establish and validate a radiomics nomogram based on contrast-enhanced spectral mammography (CESM) for prediction of axillary lymph node (ALN) metastasis in breast cancer. This retrospective study included 394 patients with breast cancer who underwent CESM examination in two hospitals. The least absolute shrinkage and selection operator (LASSO) logistic regression was established for feature selection and utilized to construct radiomics signature. The nomogram model included the radiomics signature and independent clinical factors. The receiver operating characteristic (ROC) curves were used to confirm the performance of the nomogram in training and validation sets. The nomogram model, which includes the radiomics signature and the CESM-reported lymph node status, has areas under the ROC curves of 0.774 (95% confidence interval (CI) 0.689-0.858), 0.767 (95% CI 0.583-0.857), and 0.79 (95% CI 0.63-0.94) in the training, internal validation, and external validation sets, respectively. We identified the cutoff score in the radiomics nomogram as - 1.49, which corresponded to a total point of 49 that could diagnose ALN metastasis with a sensitivity of > 95%. The CESM-based radiomics nomogram is a noninvasive predictive tool that shows good application prospects in the preoperative prediction of ALN metastasis in breast cancer. • The CESM-based radiomics nomogram shows good performance in predicting ALN metastasis in breast cancer. • The application of radiomics nomogram in this study provides a new approach for establishing a prediction model with multiple characteristics. • The nomogram has good application prospects in assisting clinical decision makers.

Background: Axillary lymph node (ALN) metastasis in early breast cancer affects the prognosis and accurate identification of ALN metastasis in early breast cancer is important for determining the treatment. Conventional ultrasonography (cUS) of breast cancer is used worldwide because it is inexpensive, simple, and does not involve radiation exposure and is not invasive. However, cUS is insufficient to evaluate ALN status accurately. Contrast-enhanced US (CEUS) improves the performance of predicting ALN metastasis, so at our facility, CEUS is performed routinely. However, there are many diagnostic indicators, and standard imaging and image interpretation methods have not yet been established, resulting in difficulty of use. To make CEUS universal and effective, we evaluate the ability of predicting ALN metastasis and importance features, using deep learning (DL) models or weighted decision tree (WDT) model with ALN CEUS imaging data and tabular formatted data.Methods:In this retrospective study, 788 CEUS images of ALNs were collected from 788 patients (mean age 59.7 [range, 25–90] years) , who underwent breast surgery between January 2013 and December 2021. First, CEUS images were inputted into the DL models, EfficientNet B0, B4, and B8 and VisionTransformer Base/16, which yielded an image-based predictive covariate (iP : 0–1) for ALN metastasis. Second, the tabular formatted data of primary tumor and ALN of cUS and CEUS was inputted into light gradient boosting machine (LightGBM), one of the WDT models. This data was including diameter of primary tumor and ALN evaluated by cUS and CEUS, 10 findings of primary tumor which could be evaluated by only CEUS, 9 findings of ALN evaluated by both cUS and CEUS, and 6 findings of ALN evaluated by only CEUS. These findings were evaluated by two breast surgery specialists. We also used iP for lightGBM models. Labeling data was histopathologic diagnosis of ALN surgical specimen divided into two groups pN0 or pN1mi and more. Results: In the analysis of DL models, the area under the receiver operating characteristic curve (AUC) was highest for the EfficientNet B8 model at 0.69, which was used as the iP. In the analysis of LightGBM, the AUC for CEUS was 0.93 (0.88-0.98), significantly higher than that for cUS alone of 0.76 (0.68-0.85) (p &amp;lt; 0.001, DeLong test). Top feature importance for predicting of ALN metastasis was “heterogeneous enhancement pattern, “diffuse cortical thickening”, and “eccentric cortical thickening &amp;gt;3 mm”.Conclusion:LightGBM model, an AI model, was able to extract CEUS findings that are important in predicting ALN metastasis in early breast cancer. Citation Format: Tomohiro Oshino, Ken Enda, Hirokazu Shimizu, Megumi Sato, Mutsumi Nishida, Fumi Kato, Mitsuchika Hosoda, Kohsuke Kudo, Norimasa Iwasaki, Masato Takahashi. Tomohiro Oshino, Ken Enda, Hirokazu Shimizu, Megumi Sato, Mutsumi Nishida, Fumi Kato, Mitsuchika Hosoda, Kohsuke Kudo, Norimasa Iwasaki, Masato Takahashi [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P2-04-16.

This study sought to develop a radiomics model capable of predicting axillary lymph node metastasis (ALNM) in patients with invasive breast cancer (IBC) based on dual-sequence magnetic resonance imaging(MRI) of diffusion-weighted imaging (DWI) and dynamic contrast enhancement (DCE) data. The interpretability of the resultant model was probed with the SHAP (Shapley Additive Explanations) method. Established inclusion/exclusion criteria were used to retrospectively compile MRI and matching clinical data from 183 patients with pathologically confirmed IBC from our hospital evaluated between June 2021 and December 2023. All of these patients had undergone plain and enhanced MRI scans prior to treatment. These patients were separated according to their pathological biopsy results into those with ALNM (n = 107) and those without ALNM (n = 76). These patients were then randomized into training (n = 128) and testing (n = 55) cohorts at a 7:3 ratio. Optimal radiomics features were selected from the extracted data. The random forest method was used to establish three predictive models (DWI, DCE, and combined DWI + DCE sequence models). Area under the curve (AUC) values for receiver operating characteristic (ROC) curves were utilized to assess model performance. The DeLong test was utilized to compare model predictive efficacy. Model discrimination was assessed based on the integrated discrimination improvement (IDI) method. Decision curves revealed net clinical benefits for each of these models. The SHAP method was used to achieve the best model interpretability. Clinicopathological characteristics (age, menopausal status, molecular subtypes, and estrogen receptor, progesterone receptor, human epidermal growth factor receptor 2, and Ki-67 status) were comparable when comparing the ALNM and non-ALNM groups as well as the training and testing cohorts (P > 0.05). AUC values for the DWI, DCE, and combined models in the training cohort were 0.793, 0.774, and 0.864, respectively, with corresponding values of 0.728, 0.760, and 0.859 in the testing cohort. The predictive efficacy of the DWI and combined models was found to differ significantly according to the DeLong test, as did the predictive efficacy of the DCE and combined models in the training groups (P < 0.05), while no other significant differences were noted in model performance (P > 0.05). IDI results indicated that the combined model offered predictive power levels that were 13.5% (P < 0.05) and 10.2% (P < 0.05) higher than those for the respective DWI and DCE models. In a decision curve analysis, the combined model offered a net clinical benefit over the DCE model. The combined dual-sequence MRI-based radiomics model constructed herein and the supporting interpretability analyses can aid in the prediction of the ALNM status of IBC patients, helping to guide clinical decision-making in these cases.

This retrospective study aimed to evaluate the predictive value of the preoperative dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and diffusion-weighted imaging (DWI) findings of mass lesions for predicting sentinel lymph node (SLN) metastasis in early breast cancer. A total of 310 patients with suspicious mass lesions detected in preoperative MRI who subsequently underwent surgery and SLN biopsy (SLNB) between September 2015 and September 2022 were analysed. The relationship between DCE-MRI and DWI findings and SLNB positivity was analysed. SLNB was positive for SLN metastasis in 108 of 310 lesions. Younger age (P = 0.001) and larger lesion size (P < 0.001) were found to be associated with SLNB positivity. Findings associated with SLN metastasis included peritumoural oedema in 53%, adjacent vessel sign (AVS) in 81%, and increased whole-breast vascularity (WBV) in 58% of patients with positive SLNB (P < 0.001). The SLNB positivity rate was higher in mass lesions with DCE-MRI findings of heterogenous enhancement pattern (P = 0.003), medium or rapid initial phase enhancement (P = 0.001), and washout delayed phase kinetic curve (P = 0.001). It was found that lower tumoural apparent diffusion coefficient (ADC) values (P = 0.003) and higher peritumoural/tumoural ADC ratios (P = 0.018) increased the probability of encountering SLN metastasis. Patient age, presence of peritumoural oedema, presence of AVS, increased WBV, and initial phase kinetic curve of the lesions on MRI were found to be associated with SLN metastasis. We found that younger age and MR findings obtained from the perilesional area of breast cancer may be helpful in the preoperative prediction of SLN metastasis.

Predicting axillary lymph node metastasis in small breast cancers using convolutional neural networks for multiparametric magnetic resonance imaging (MRI).

We aimed to develop a multi-modality model to predict axillary lymph node (ALN) metastasis by combining clinical predictors with radiomic features from magnetic resonance imaging (MRI) and mammography (MMG) in breast cancer. This model might potentially eliminate unnecessary axillary surgery in cases without ALN metastasis, thereby minimizing surgery-related complications. We retrospectively enrolled 485 breast cancer patients from two hospitals and extracted radiomics features from tumor and lymph node regions on MRI and MMG images. After feature selection, three random forest models were built using the retained features, respectively. Significant clinical factors were integrated with these radiomics models to construct a multi-modality model. The multi-modality model was compared to radiologists' diagnoses on axillary ultrasound and MRI. It was also used to assist radiologists in making a secondary diagnosis on MRI. The multi-modality model showed superior performance with AUCs of 0.964 in the training cohort, 0.916 in the internal validation cohort, and 0.892 in the external validation cohort. It surpassed single-modality models and radiologists' ALN diagnosis on MRI and axillary ultrasound in all validation cohorts. Additionally, the multi-modality model improved radiologists' MRI-based ALN diagnostic ability, increasing the average accuracy from 70.70 to 78.16% for radiologist A and from 75.42 to 81.38% for radiologist B. The multi-modality model can predict ALN metastasis of breast cancer accurately. Moreover, the artificial intelligence (AI) model also assisted the radiologists to improve their diagnostic ability on MRI. The multi-modality model based on both MRI and mammography images allows preoperative prediction of axillary lymph node metastasis in breast cancer patients. With the assistance of the model, the diagnostic efficacy of radiologists can be further improved. • We developed a novel multi-modality model that combines MRI and mammography radiomics with clinical factors to accurately predict axillary lymph node (ALN) metastasis, which has not been previously reported. • Our multi-modality model outperformed both the radiologists' ALN diagnosis based on MRI and axillary ultrasound, as well as single-modality radiomics models based on MRI or mammography. • The multi-modality model can serve as a potential decision support tool to improve the radiologists' ALN diagnosis on MRI.