Tao Ai

Background: Accurate identification of molecular subtypes and prognostic factors is crucial for personalized breast cancer management. Multi-parametric magnetic resonance imaging (mp-MRI) offers comprehensive, non-invasive characterization of tumors by capturing complementary information on morphology, perfusion, and tissue microstructure, but conventional radiologic assessment is often inadequate for reliably inferring molecular subtype or prognostic factors. Existing deep learning (DL) approaches depend on manually delineated tumor regions to define inputs, which is difficult to scale in routine practice. We therefore propose an automated end-to-end DL pipeline that jointly performs tumor segmentation and biomarker prediction directly from mp-MRI. Methods: In this multicenter study, mp-MRI data from 484 patients with breast cancer were collected from three institutions. The pipeline first used a three-dimensional U-Net (3D U-Net) for automated tumor segmentation and then a two-dimensional Residual Network (2D ResNet)-based classifier for predicting molecular subtypes and prognostic factors. Segmentation performance was quantified using mean intersection over union (mIoU) and mean dice similarity coefficient (mDice). Classification performance was assessed through area under the receiver operating characteristic curve (AUC), complemented by accuracy, sensitivity, specificity, and F1-score metrics. The DeLong test was used to compare AUCs between models based on automated segmentation regions of interest (ROIs) and manually delineated ROIs. Results: The model achieved robust segmentation performance in the validation set (mIoU =0.772, mDice =0.864) and external cohorts (mIoU =0.732–0.746, mDice =0.838–0.848). For molecular subtyping, triple-negative (TN) differentiation achieved the best performance (AUC: validation =0.839, external =0.729–0.749). Axillary lymph node (ALN) status prediction yielded AUCs of 0.773, 0.715, and 0.700, respectively. Ki67 expression prediction demonstrated high generalizability (AUC: validation =0.788, external =0.749–0.784). No statistically significant differences were observed in classification performance between models based on automated segmentation ROIs and those based on manually delineated ROIs (all P>0.05). Conclusions: The automated end-to-end pipeline integrates 3D U-Net-based tumor segmentation and 2D ResNet-based biomarker classification, supporting its potential as a complementary tool for diagnostics, treatment planning, and prognostic assessment in breast cancer.

Radiomics-driven prediction of pathologic complete response in non-mass breast cancer using post-neoadjuvant chemotherapy preoperative dynamic contrast-enhanced magnetic resonance imaging - Diagnostic and Interventional Radiology

To investigate the potential of combining time-dependent diffusion MRI (td-dMRI) with intravoxel incoherent motion (IVIM) to predict the Nottingham Prognos

Objective: This study aims to evaluate the diagnostic performance of diffusion-weighted imaging (DWI) with a fractional order calculus (FROC) model for differentiating breast lesions and to explore the associations between FROC/apparent diffusion coefficient (ADC)-derived diffusion metrics and prognostic biomarkers and molecular subtypes in breast cancer. Methods: This retrospective study included 147 patients with 159 histopathology-confirmed lesions who underwent multi-b DWI using simultaneous multi-slice (SMS) readout-segmented echo-planar imaging (rs-EPI) at 3.0 T. Whole-lesion histograms were computed for mono-exponential ADC and FROC parameters (D, β, μ). The Mann–Whitney U test was used to compare the histogram metrics of each diffusion parameter between the benign and malignant groups and between groups with different prognostic biomarkers and molecular subtypes. The Kruskal–Wallis test was used to compare the histogram metrics of each DWI-derived parameter among the different molecular subtypes. The Spearman rank correlation analysis was employed to characterize correlations between diffusion parameters and prognostic biomarkers. The diagnostic performance of each DWI-derived parameter in differentiating breast lesions was assessed using receiver operating characteristic (ROC) analysis. Results: Interobserver reproducibility was excellent (intra-class correlation coefficient 0.827–0.928). Central tendency histogram metrics (10th, 90th percentiles, mean, median) of ADC and FROC parameters were higher in benign than malignant lesions, whereas skewness (all models) and entropy/kurtosis (ADC, D, μ) were lower in benign lesions (all p < 0.05, except β-skewness). The histogram metrics of ADC-median, D _FROC -mean, and D _FROC -median showed similar diagnostic performance. The values of ADC-mean, D _FROC -10%, D _FROC -mean, D _FROC -median, β _FROC -10%, β _FROC -mean, and β _FROC -median were significantly lower in the estrogen receptor (ER)-positive group compared with those in the ER-negative group. The tumors with progesterone receptor (PR)-negative status showed significantly higher β _FROC -10%, β _FROC -mean, and β _FROC -median values than those of tumors with PR-positive status. The values of D _FROC -skewness, β _FROC -10%, and β _FROC -mean exhibited significant differences in differentiating the triple-negative and luminal subtypes. Conclusions: FROC-based histogram analysis yields diagnostic performance comparable to ADC for benign vs. malignant classification, while providing richer associations with ER/PR status, proliferation, and nodal involvement, reflecting microstructural heterogeneity not captured by mono-exponential diffusion.

This study aimed to develop a predictive model integrating multi-sequence MRI radiomics, deep learning features, and habitat imaging to forecast pathological complete response (pCR) in breast cancer patients undergoing neoadjuvant therapy (NAT). A retrospective analysis included 203 breast cancer patients treated with NAT from May 2018 to January 2023. Patients were divided into training (n = 162) and test (n = 41) sets. Radiomics features were extracted from intratumoral and peritumoral regions in multi-sequence MRI (T2WI, DWI, and DCE-MRI) datasets. Habitat imaging was employed to analyze tumor subregions, characterizing heterogeneity within the tumor. We constructed and validated machine learning models, including a fusion model integrating all features, using Receiver Operating Characteristic (ROC) and Precision-Recall (PR) curves, decision curve analysis (DCA), and confusion matrices. Shapley Additive Explanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) analyses were performed for model interpretability. The fusion model achieved superior predictive performance compared to single-region models, with AUCs of 0.913 (95% CI: 0.770–1.000) in the test set. PR curve analysis showed improved precision-recall balance, while DCA indicated higher clinical benefit. Confusion matrix analysis confirmed the model’s classification accuracy. SHAP revealed DCE_LLL_DependenceUniformity as the most critical feature for predicting pCR and PC72 for non-pCR. LIME provided patient-specific insights into feature contributions. Integrating multi-dimensional MRI features with habitat imaging enhances pCR prediction in breast cancer. The fusion model offers a robust, non-invasive tool for guiding individualized treatment strategies while providing transparent interpretability through SHAP and LIME analyses.

This study aims to comprehensively evaluate the clinical utility of five diffusion models, including conventional mono-exponential (Mono), intravoxel incoherent motion (IVIM), diffusion kurtosis imaging (DKI), stretched exponential (SEM), and continuous-time random-walk (CTRW), for preoperatively predicting of breast lesion pathology, prognostic biomarkers, and molecular subtypes. We retrospectively analyzed 132 patients with pathologically verified breast lesions (41 benign and 91 malignant) who underwent a full protocol preoperative breast MRI protocol, including a diffusion-weighted imaging (DWI) sequence with nine b values (0 to 2000 s/mm²) on a 3.0T MR scanner. The diffusion parameters from each model—Mono (ADC), IVIM (D, D*, f), DKI (MD, MK), SEM (DDC, α) and CTRW (D_m, α, β)—were quantitatively calculated and compared between benign and malignant breast lesions, as well as across different prognostic biomarker statuses in breast cancer, using Mann–Whitney U-tests. For molecular subtypes comparisons, we employed the Kruskal-Wallis test followed by Bonferroni. All parameters, except IVIM-D*, significantly differentiated benign from malignant lesions. Notably, IVIM-D and DKI-MK values were significantly different between estrogen receptor (ER)-positive and ER-negative tumors. Progesterone receptor (PR)-positive cancers exhibited lower Mono-ADC, IVIM-D, DKI-MD, SEM-DDC, CTRW-D_m, and CTRW-α values, alongside higher DKI-MK value compared to PR-negative cancers (p < 0.05). Significant differences in IVIM-D, IVIM-D*, and DKI-MK values were observed between human epidermal growth factor receptor 2 (HER2)-negative and HER2-positive tumors. Furthermore, higher SEM-α and CTRW-β values, along with lower DKI-MD and SEM-DDC values, were noted in the high Ki-67 expression group compared to the low Ki-67 group (p < 0.05). All five diffusion models proved valuable for breast cancer diagnosis, with the CTRW model exhibiting the highest diagnostic performance, although the difference was not statistically significant. The diffusion parameters derived from these models can effectively assist in distinguishing prognostic factors and molecular subtypes of breast cancer.

Purpose: To investigate what pre-treatment clinical-pathological features and MRI characteristics influence the performance of breast MRI in assessing the pathologic complete response (pCR) of breast cancer patients to Neoadjuvant Chemotherapy (NAC).Methods: A total of 225 patients with pathologically-confirmed breast cancer who underwent pre- and post-NAC breast MRI between January 2020 and April 2023 were retrospectively analyzed. All patients were categorized into radiologic complete response (rCR) and non-rCR groups based on pre-operative MRI. Univariable and multivariable logistic regression were used to identify independent clinicopathological and imaging features associated with imaging-pathological discordance. The performance of pre-operative MRI for predicting pCR to NAC was assessed according to the baseline characteristics of the clinicopathological data and pre-NAC MRI. In addition, the discrepancy between the pre-operative MRI and post-operative pathological findings was further analyzed by a case-control approach.Results: Among 225 patients, 99 (44.0%) achieved pCR after NAC. MRI showed the overall sensitivity of 97.6%, specificity of 58.6%, accuracy of 80.4%, a positive predictive value (PPV) of 75.0%, and a negative predictive value (NPV) of 95.1% in identifying pCR. Of baseline features, presence of ductal carcinoma in situ (DCIS) (OR, 3.975 [95% CI: 1.448–10.908], p = 0.007), luminal B (OR, 5.076 [95% CI: 1.401–18.391], p = 0.013), HER2-enriched subtype (OR, 10.949 [95% CI: 3.262–36.747], p < 0.001), multifocal or multicentric lesions (OR, 2.467 [95% CI: 1.067–5.706], p = 0.035), segmental or regional distribution of NME (OR, 8.514 [95% CI: 1.049–69.098], p = 0.045) and rim enhancement of mass (OR, 4.261 [95% CI: 1.347–13.477], p = 0.014) were significantly associated with the discrepancy between MRI and pathology.Conclusion: Presence of DCIS, luminal B or HER2-enriched subtype, multicentric or multifocal lesions, segmental or regional distribution of NME and rim enhancement of mass may lead to a decrease in diagnostic accuracy of MRI in patients of breast cancer treated with NAC.

Background: To compare the compartmentalized diffusion-weighted models, intravoxel incoherent motion (IVIM) and restriction spectrum imaging (RSI), in characterizing breast lesions and normal fibroglandular tissue. Methods: This prospective study enrolled 152 patients with 157 histopathologically verified breast lesions (41 benign and 116 malignant). All patients underwent a full-protocol preoperative breast MRI, including a multi-b-value DWI sequence. The diffusion parameters derived from the mono-exponential model (ADC), IVIM model (Dt, Dp, f), and RSI model (C₁, C₂, C₃, C₁C₂, F₁, F₂, F₃, F₁F₂) were quantitatively measured and then compared among malignant lesions, benign lesions and normal fibroglandular tissues using Kruskal-Wallis test. The Mann-Whitney U-test was used for the pairwise comparisons. Diagnostic models were built by logistic regression analysis. The ROC analysis was performed using five-fold cross-validation and the mean AUC values were calculated and compared to evaluate the discriminative ability of each parameter or model. Results: Almost all quantitative diffusion parameters showed significant differences in distinguishing malignant breast lesions from both benign lesions (other than C₂) and normal fibroglandular tissue (all parameters) (all P < 0.0167). In terms of the comparisons of benign lesions and normal fibroglandular tissues, the parameters derived from IVIM (Dp, f) and RSI (C₁, C₂, C₁C₂, F₁, F₂, F₃) showed significant differences (all P < 0.005). When using individual parameters, RSI-derived parameters-F₁, C₁C₂, and C₂ values yielded the highest AUCs for the comparisons of malignant vs. benign, malignant vs. normal tissue and benign vs. normal tissue (AUCs = 0.871, 0.982, and 0.863, respectively). Furthermore, the combined diagnostic model (IVIM + RSI) exhibited the highest diagnostic efficacy for the pairwise discriminations (AUCs = 0.893, 0.991, and 0.928, respectively). Conclusions: Quantitative parameters derived from the three-compartment RSI model have great promise as imaging indicators for the differential diagnosis of breast lesions compared with the bi-exponential IVIM model. Additionally, the combined model of IVIM and RSI achieves superior diagnostic performance in characterizing breast lesions.

Background: In our previous study, we developed a combined diagnostic model based on time-intensity curve (TIC) types and radiomics signature on contrast-enhanced magnetic resonance imaging (CE-MRI) for non-mass enhancement (NME). The model had a high diagnostic ability for differentiation without the additional diffusion-weighted imaging (DWI) sequence. In this study, we aimed to compare the diagnostic performance of the combined clinical-radiomics model based on CE-MRI and DWI in discriminating Breast Imaging-Reporting and Data System (BI-RADS) 4 NME breast lesions, ductal carcinoma in situ (DCIS), and invasive carcinoma. Methods: This retrospective study enrolled 364 NME lesions (343 patients). Of these, 183 malignant and 84 benign breast lesions classified as BI-RADS 4 NMEs by the initial diagnosis were reclassified based on the combined clinical–radiomics model and DWI, respectively. The nomogram score (NS) values for malignancy risk derived from the combined clinical-radiomics model and the minimal apparent diffusion coefficient (ADC) values from DWI were calculated and compared. The percentage of false positives were estimated in comparison with the original classification. Receiver operating characteristic (ROC) curve analysis was performed to determine the diagnostic value of the NS and minimal ADC values in distinguishing benign and malignant lesions, DCIS, and invasive breast carcinoma. An ablation experiment was used to test the value of the additional DWI sequence. Results: The diagnostic value of the NS values [area under curve (AUC) =0.843; 95% CI: 0.789–0.896] for discriminating the 267 NME breast lesions categorized as BI-RADS 4 was significantly higher than the minimal ADC values (AUC =0.662; 95% CI: 0.590–0.735). The NS values showed higher sensitivity, specificity, and accuracy compared with the minimal ADC values (sensitivity: 80.3% vs. 65.6%; specificity: 79.8% vs. 65.5%; accuracy: 80.1% vs. 65.5%). The NS values and minimal ADC values did not achieve high diagnostic accuracy in discriminating between DCIS and invasive cancer. However, the diagnostic performance of the combined NS-ADC model (AUC =0.731; 95% CI: 0.655–0.806) was higher than that of the NS values alone (P=0.008) and comparable to that of the minimal ADC values (P=0.440). Conclusions: The combined clinical-radiomics model based on CE-MRI could improve the diagnostic performance in discriminating the BI-RADS 4 NME lesions without an additional DWI sequence. However, DWI may improve the diagnostic performance in discriminating DCIS from invasive cancer.