The value of breast ultrafast dynamic contrast-enhanced magnetic resonance imaging in diagnosing axillary lymph node metastasis in mass-type invasive ductal carcinoma of the breast

Introduction

Breast cancer (BC) is the most common malignant tumor among women worldwide, surpassing lung cancer in 2020 to become the most prevalent malignancy, and significantly affecting women’s health. Accurate assessment of axillary lymph node (ALN) status plays a crucial role in BC staging, treatment decision-making, and overall survival evaluation (1-3). Currently, physical examination, ultrasound, mammography, computed tomography (CT), conventional magnetic resonance imaging (MRI), and positron emission tomography-CT (PET-CT) can all identify suspicious lymph nodes (4-6), although with varying accuracy due to multiple influencing factors (7). ALN dissection (ALND) and sentinel lymph node biopsy (SLNB) are performed with the primary goal of accurately targeting metastatic lymph nodes. However, ALND often results in complications such as upper limb edema and restricted shoulder movement, impacting postoperative quality of life (8,9), and SLNB carries a higher risk of anesthesia and a relatively high false-negative rate (10). It is particularly important to improve the accurate prediction of ALN metastasis in clinical work. MRI, capable of both anatomical and functional imaging, is widely accepted in clinical breast examinations. Dynamic contrast-enhanced MRI (DCE-MRI) is an essential component, offering morphological and hemodynamic characteristics of lesions (11). Ultrafast DCE-MRI (UF-DCE MRI) is a new technique that captures the early hemodynamic information (within 2 minutes post-contrast injection) of lesions with high temporal resolution while maintaining reasonable spatial resolution. It not only avoids the influence of background parenchymal enhancement (BPE) (12) but also provides new semi-quantitative parameters by analyzing the detailed time and shape of the rising phase in the time-intensity curve (TIC) (13).

Research on UF-DCE MRI in the context of ALN metastasis in BC is currently limited. This study represents the first attempt to employ the golden-angle radial sparse parallel (GRASP) sequence for continuous scanning, with UF-DCE MRI analysis performed on the first 120 seconds of post-contrast data. The TICs derived from GRASP’s high temporal resolution demonstrate superior smoothness and better reflect the true perfusion characteristics of lesions. Consequently, we did not perform standard DCE-MRI scans (14), as prior studies have demonstrated comparable diagnostic performance between GRASP and standard DCE-MRI (15).

This study aimed to retrospectively analyze preoperative UF-DCE MRI parameters in surgically confirmed mass-type invasive ductal carcinoma (IDC) cases at our hospital, noninvasively and one-stop evaluate their diagnostic efficacy for ALN metastasis, and attempt to construct a nomogram for ALN metastasis risk assessment in combination with clinical characteristics. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-703/rc).

Methods

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Ethics Committee of The Fourth Hospital of Hebei Medical University (No. 2022037), and the requirement for individual consent for this analysis was waived due to the retrospective nature. We retrospectively analyzed the imaging and medical records of patients with surgically confirmed mass-type IDC of the breast treated at our hospital from August 2023 to May 2024. The inclusion criteria were as follows: (I) female, >18 years of age; (II) underwent breast UF-DCE MRI within two weeks before surgery, with complete imaging data; (III) imaging showed a mass-type BC lesion >8 mm; (IV) preoperative biopsy or postoperative pathology confirmed IDC, with pathological confirmation of ALN status. The exclusion criterion was BC patients who were lactating or pregnant. A total of 96 patients were included, with 38 cases of ALN metastasis and 58 cases without ALN metastasis (Figure 1).

Figure 1 Flowchart. ALN, axillary lymph node; IDC, invasive ductal carcinoma; UF-MRI, ultrafast magnetic resonance imaging.

MRI image acquisition

All examinations were performed using a 3 T MRI scanner (MAGNETOM Vida, Siemens Healthineers, Forchheim, Germany) with an 18-channel phased-array breast coil. Patients were positioned prone with both breasts naturally suspended, and the scanning range included both breasts, both axillae, and extended posteriorly to the anterior edge of the thoracic spine. The routine scanning sequences were as follows: axial fast spin-echo T1-weighted imaging (T1WI) sequence [repetition time (TR)/echo time (TE) =6.04/2.46 ms, field of view (FOV) =360 mm × 448 mm, slice thickness =4.5 mm, matrix =385 mm × 448 mm], T2-weighted imaging (T2WI) Dixon sequence (TR/TE =5,000/84 ms, FOV =360 mm × 448 mm, slice thickness =4.5 mm, matrix =448 mm × 448 mm), T1-mapping quantitative measurement sequence (TR/TE =5.3/2.34 ms, FOV =380 mm × 256 mm, slice thickness =2.5 mm, matrix =385 mm × 385 mm), and diffusion-weighted imaging (DWI) sequence (TR/TE =6,700/67 ms, FOV =340 mm × 170 mm, slice thickness =4.5 mm, matrix =88 mm × 170 mm, b-values =0, 1,200 s/mm², averages =1 and 3). Gadodiamide (Omniscan, GE Healthcare Ireland Ltd., Cork, Ireland) was injected via the cubital vein at 0.2 mmol/kg with a rate of 3.0 mL/s, followed by an immediate 10 mL flush of 0.9% sodium chloride solution. UF-DCE MRI scanning was performed using a GRASP sequence for continuous multi-phase acquisition. The total scan time was 8 minutes 25 seconds with scanning parameters of TR/TE =3.5/1.35 ms, FOV =380 mm × 320 mm, slice thickness =2.5 mm, matrix =380 mm × 380 mm, and radial views =2,165. First, mask images were scanned (total scan time: 25 s, temporal resolution: 12.5 s, 2 phases, radial views =55), and contrast agent was injected after 25 seconds, initiating the ultrafast phase scanning (total scan time: 180 s, temporal resolution: 3 s, 60 phases), followed by the delayed phase scanning (total scan time: 300 s, temporal resolution: 12.5 s, 24 phases) (Figure 2).

Figure 2 Schematic diagram of ultrafast segmented scanning parameters.

Image analysis

Analysis of clinical, pathological, and routine imaging data

All images were uploaded to the Picture Archiving and Communication System (PACS) and reviewed by two radiologists with 5 and 10 years of respective experience in breast MRI diagnosis. Together, they evaluated and recorded the qualitative and quantitative data of all cases’ MRI images and clinical data. In cases of disagreement, consensus was reached through consultation. Categorical data parameters included menopausal status, fibroglandular tissue (FGT) type, BPE, and TIC type. Quantitative data included patient age, tumor size (measured in axial, sagittal, and coronal views, with the largest diameter recorded), and apparent diffusion coefficient (ADC) value. Regions of interest (ROI) were drawn on the largest layer of the lesion, highlighting the area with the most intense enhancement, avoiding vessels, hemorrhage, and necrotic areas, with an ROI area ≥5 mm². Measurements were recorded accordingly.

UF-DCE MRI analysis

Data were processed using the research software MR-DCE (version 1.1.2, Siemens Healthineers, Erlangen, Germany), with the following steps: GRASP, T2WI, and T1-mapping sequences were imported into the MR-DCE software [with 60 phases of GRASP sequence ultrafast scans used as dynamic data, fat-suppressed T2-turbo spin echo (TSE) as morphology data, and the T1-mapping sequence as personalized T1 values to fit the enhancement concentration curve]. After dynamic motion correction, T2WI anatomical images and pre-enhanced multi-flip angle T1-volumetric interpolated breath-hold examination (VIBE) images were registered and aligned with dynamic enhanced images to reduce fusion calculation errors. For the arterial input function (AIF), the descending aorta was uniformly selected as the ROI to construct a reference TIC. The breast model (14,16) was then selected, and an ROI was drawn to obtain parameters, including maximum slope (MS), time-to-enhancement (TTE, defined as the time to reach 20% of total enhancement), time-to-peak (TTP) after enhancement, relative peak enhancement (PEI), wash-out slope for 45 seconds after peak (WOS45), absolute peak enhancement index (APEI), area under the curve 60 seconds after enhancement (iAUC60), and time to center of maximum slope (TTMS), as shown in Figure 3.

Figure 3 Female, 45 years old, right breast mass. (A) T1 enhancement, showing lobulation in the right breast mass; (B) MIP image, displaying axial vascular distribution in the right breast mass; (C) TIC curve; (D-I) ultrafast scan parameter pseudocolor images for TTE, MS, iAUC60, TTP, PEI, and TTMS. Parameter values are shown below (A) and (B). APEI, absolute peak enhancement index; iAUC60, area under the curve, 60 sec after enhancement; MIP, maximum intensity projection; MS, maximum slope; PEI, relative peak enhancement; ROI, region of interest; TIC, time-intensity curve; TTE, time-to-enhancement; TTP, time-to-peak; TTMS, time to center of maximum slope; WOS45, wash-out slope for 45 sec after peak.

Conventional DCE-MRI analysis

Using a post-processing software (syngo.MR Tissue 4D, Siemens Healthineers, Erlangen, Germany), a pharmacokinetic Tofts’ model (17) was established. GRASP, T2WI (as anatomical images), flip-angle images, and T1-map sequences were imported into the Tissue 4D workflow, performing motion correction during dynamic scanning. Pre-contrast, anatomical, and dynamic reference images were registered (using T2WI as anatomical reference and images with different flip angles (10°, 15°) to calculate T1 values). Pharmacokinetic arrival time was defined to calculate volume transfer constant (K^trans), extravascular extracellular volume (V_e), rate constant (K_ep) and the initial area under the time-to-signal intensity curve (iAUC) measured during the first 60 seconds. The similar ROIs were drawn on the at the same anatomical location on the same imaging plane and maintaining similar ROI sizes. The average values of all pixel points within the ROI were recorded (Figure 4).

Figure 4 The same patient as in Figure 3. (A) Axial enhanced image showing a mass in the outer quadrant of the right breast with lobulated edges; (B) axial MIP image showing vascular enhancement in the breast and tumor; (C) TIC curve; (D-F) K^trans, K_ep, V_e pseudocolor maps, with values shown below (B). K_ep, rate constant; K^trans, volume transfer constant; MIP, maximum intensity projection; ROI, region of interest; TIC, time-intensity curve; V_e, extravascular extracellular volume.

Statistical analysis

Statistical analysis was performed using the software SPSS 26.0 (IBM Corp., Armonk, NY, USA) and R (R Foundation for Statistical Computing, Vienna, Austria). For continuous variables, independent samples t-tests were applied when the data satisfied normality and homogeneity of variance assumptions; otherwise, the non-parametric Mann-Whitney U test (Wilcoxon rank-sum test) was used. For categorical data, Fisher’s exact test was employed for 2×2 contingency tables with binary outcomes, whereas the Chi-squared test was used when the measurement data did not meet the requirement. A P value <0.05 was considered statistically significant, and 95% confidence interval (CI) (18) sensitivity, and specificity were calculated. Univariate analysis and Pearson correlation test were used for feature screening, and the selected features were used to construct a logistic regression model. Receiver operating characteristic (ROC) curves were generated for significant features and the combined diagnostic model, and the area under the curve (AUC) was calculated. A nomogram was also created based on parameters with significant intergroup differences to predict the risk of ALN metastasis in mass-type BC.

Results

Differences in clinical, conventional imaging, and pathological parameters between groups

The mean age of the group without ALN metastasis was 51 years (range, 27–72 years, median: 52 years), whereas the mean age of the group with ALN metastasis was 49 years (range, 33–71 years, median: 49 years, P>0.05). The tumor size in the group without ALN metastasis was significantly smaller than that in the group with ALN metastasis (2.2±1.1 vs. 3.0±1.5 cm, P<0.001) (Table 1). Although the difference in ADC values between the two groups was not statistically significant, the ADC value in the ALN metastasis group (0.83±0.17 mm²/s) was slightly lower than that in the without ALN metastasis group (0.91±0.21 mm²/s) (Table 2). No significant differences were observed between the two groups in menopausal status, gland type, background enhancement, or TIC (Table 1).

Differences in UF-DCE and conventional DCE imaging parameters between groups

The TTE, TTP, and TTMS values were significantly lower in the ALN metastasis group (10.05±4.91 vs. 15.59±15.04 s, 85.89±33.93 vs. 208.27±102.28 s, 19.05±22.25 vs. 19.81±9.29 s), with statistically significant differences (P<0.05). Among UF-DCE parameters, MS, iAUC60, PEI, APEI, and WOS45 showed no statistically significant differences between groups (P>0.05). Conventional DCE parameters K^trans, K_ep, and V_e were slightly lower in the ALN metastasis group but without statistical significance between groups (Table 2).

Diagnostic performance of MRI and clinical parameters for lymph node metastasis

Among UF-DCE parameters, TTP had the highest diagnostic efficacy (Figure 5), with an AUC of 0.865 [95% CI: 0.794–0.937, specificity: 0.974, sensitivity: 0.672, positive predictive value (PPV): 0.975, negative predictive value (NPV): 0.661], followed by TTMS: 0.667 (95% CI: 0.555–0.779, specificity: 0.763, sensitivity: 0.552, PPV: 0.780, NPV: 0.527), and TTE: 0.664 (95% CI: 0.550–0.777, specificity: 0.421, sensitivity: 0.914, PPV: 0.707, NPV: 0.762). The AUC for tumor size was 0.694 (95% CI: 0.588–0.801, specificity: 0.448, sensitivity: 0.921, PPV: 0.522, NPV: 0.896). A combined predictive model was constructed using logistic regression with TTP, ADC value, and tumor size, yielding an AUC of 0.919 (95% CI: 0.864–0.974, specificity: 0.759, sensitivity: 0.947, PPV: 0.720, NPV: 0.957).

Figure 5 ROC curves for imaging and clinical parameters. The AUC values for each parameter are as follows—TTP: 0.865 (95% CI: 0.794–0.937, specificity: 0.974, sensitivity: 0.672, PPV: 0.975, NPV: 0.661), TTE: 0.664 (95% CI: 0.550–0.777, specificity: 0.421, sensitivity: 0.914, PPV: 0.707, NPV: 0.762), TTMS: 0.667 (95% CI: 0.555–0.779, specificity: 0.763, sensitivity: 0.552, PPV: 0.780, NPV: 0.527), tumor size: 0.694 (95% CI: 0.588–0.801, specificity: 0.448, sensitivity: 0.921, PPV: 0.522, NPV: 0.896), combined logistic prediction probability: 0.919 (95% CI: 0.864–0.974, specificity: 0.759, sensitivity: 0.947, PPV: 0.720, NPV: 0.957). AUC, area under the curve; CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value; ROC, receiver operating characteristic; TTE, time-to-enhancement; TTMS, time to center of maximum slope; TTP, time-to-peak.

Feature selection and joint model construction

The univariate analysis demonstrated that tumor size, TTE, TTP, and TTMS were statistically significant in differentiating ALN metastasis in mass-type IDC. However, both practical considerations and subsequent correlation analyses revealed that TTP, TTE, and TTMS were highly correlated with each other. Therefore, we retained ADC (P=0.065) despite its univariate P value exceeding 0.05 (but remaining below 0.1), as it still held some discriminative value for identifying ALN metastasis in mass-type IDC. Since TTP, TTE, and TTMS are all UF-DCE MRI-derived parameters and correlation analysis confirmed their strong interdependence, only TTP was retained (Figure 6). Consequently, tumor size, ADC, and TTP were incorporated into the construction of the nomogram for predicting ALN metastasis (Figure 7). TTP had the greatest influence on ALN metastasis, followed by tumor size (approximately half the impact of TTP), and ADC had the lowest influence. The lower the TTP and ADC values and the larger the tumor size, the higher the probability of predicting ALN metastasis. Score = 0.70 * tumor size − 2.49 * ADC − 0.03 * TTP + 3.78, metastasis probability = e^score/1 + e^score. When using the nomogram, the TTP value, tumor size, and ADC value of the patient’s tumor were first obtained, and then the score of each parameter was obtained by aligning the corresponding parameter on the horizontal line of the nomogram according to the specific value. The total score of the patient was obtained by summing the scores of each parameter. Then, the predicted probability of ALN metastasis was obtained by aligning the Total score of the patient downward on the horizontal line of Total Points.

Figure 6 Feature correlation heatmap. Significant correlations were observed among TTP, TTE, and TTMS (P<0.05), indicating strong relationships between them. Consequently, only TTP was included in the final model. ADC, apparent diffusion coefficient; TTE, time-to-enhancement; TTMS, time to center of maximum slope; TTP, time-to-peak.

Figure 7 Nomogram for predicting the risk of axillary lymph node metastasis. TTP has the greatest impact on ALN metastasis risk, followed by tumor size, with ADC having the lowest impact. ADC, apparent diffusion coefficient; ALN, axillary lymph node; TTP, time-to-peak.

Discussion

Although Zhang et al. found that analyzing the ROI of the ALN area has better diagnostic efficiency than the ROIs containing the breast tumor area (19), some lymph nodes may be located at the edge of the coil during breast MRI scanning due to patient body type, and some may not even be within the coil’s coverage area. Additionally, matching lymph nodes on imaging with intraoperative pathology requires more manual effort. This study examined the characteristics of clinical, conventional MRI parameters, UF-DCE, and traditional DCE-MRI parameters in cases of mass-type IDC, aiming to predict ALN metastasis directly through these parameters related to the breast tumor itself, with the hope of noninvasively predicting ALN status ahead of time. This study used a new free-breathing ultrafast GRASP technique for continuous multi-phase DCE imaging of the breast, thereby achieving both ultrafast and traditional DCE-MRI quantitative analysis in a single scan, greatly saving scanning resources and time. Previous studies have verified the potential of GRASP in breast imaging. Heacock et al. found that, under the same parameters, the image quality of GRASP breast images is superior to that of VIBE sequences (15). This is because, although traditional DCE sequences offer higher spatial resolution, their temporal resolution is relatively low, leading to the loss of important temporal information. The GRASP sequence combines compressed sensing and parallel imaging, achieving both high spatial and temporal resolution and providing continuous time-frame data.

In our study, there were no statistically significant differences in age or menopausal status between the two groups with and without lymph node metastasis. Tumor size was statistically significant between the groups, with larger tumors more likely to have lymph node metastasis, consistent with the findings of Chen et al. (20). This may be because larger tumors have a higher probability of involving lymphoid tissues. The ADC value of tumors in the lymph node metastasis group was slightly lower than that in the non-metastasis group, but there was no statistically significant difference between the groups. No statistical differences were observed in gland type, background enhancement, or TIC between the two groups. Tumor angiogenesis can increase pathological vascular density, wall permeability, and venous return. These vascular characteristics are thought to precede morphological changes and can be reflected in the hemodynamics of the tumor. In conventional MRI, DCE reflects the hemodynamics of both tumor and normal glandular tissues. Kang et al. found that DCE perfusion parameters are associated with poor prognosis in BC (21). However, this study found no significant differences in the pharmacokinetic parameters of DCE-MRI, including K^trans, K_ep, and V_e, between the groups. This may be due to the lack of subtyping for IDC of the breast in our study. However, in our study, TTE, TTP, and TTMS in UF-DCE MRI showed statistically significant differences between the groups. TTE indicates changes in tumor vascular permeability, reflecting the time at which the lesion begins to enhance (22). In our study, the TTE enhancement in the ALN metastasis group started earlier than in the non-ALN metastasis group (10.05±4.91 vs. 15.59±15.04), possibly due to more abundant neovascularization and poorer basement membrane integrity in the ALN metastasis group, resulting in relatively early enhancement in the dynamic enhancement curve (23). Tumor TTP in the ALN metastasis group was significantly earlier than that in the non-ALN group. This may be due to an increase in the number and diameter of blood vessels within the tumor, causing the peak time to arrive sooner. TTMS indicates microcirculation perfusion and early leakage of contrast agent from the blood vessels to the extracellular extravascular space, expressed as a slope. Lee et al. found that it has statistical differences in distinguishing between benign and malignant breast lesions (24). Tumor enlargement and increased malignancy can lead to an increase in tumor intravascular arteriovenous shunts, an increase in vascular wall integrity issues, and consequently leakage of contrast agent in the tumor in the circulatory system. These pathophysiological differences often appear early after administration, and UF-DCE MRI can effectively capture these differences. The diagnostic efficiency of TTP reaches 0.865, which is close to the diagnostic efficiency of Zhang et al. (0.89) and Chen et al. (0.899) using conventional DCE-MRI combined with radiomics and machine learning, respectively (19,25). However, in this study, the combined diagnostic efficiency of clinical and imaging parameters was higher, reaching 0.919, which is close to the diagnostic efficacy of habitat analysis applied by Wu et al. (the AUC in the training and test sets was 0.977 and 0.873, respectively) (26) with lower time and labor costs.

The nomogram, as a more intuitive visualization method, can reflect the influence of potential predictors on disease risk and prediction trends. Among the predictors, TTP has the greatest impact on ALN metastasis, followed by tumor size and ADC. The lower the TTP value, the lower the ADC value, and the larger the tumor, the greater the probability of predicting ALN metastasis.

This study has some limitations. First, this study involved a retrospective analysis, which requires further validation through prospective studies. Second, this study only included the most common subtype of BC, mass-type IDC. Further research on non-mass-type IDC and other types of BC requires a more extensive dataset. Third, this was as a single-center study; we plan to conduct multicenter validation in the future to further verify the reliability of the formula combining UF-DCE MRI with clinical parameters in predicting ALN metastasis.

Conclusions

Preoperative multiparametric UF-DCE MRI can support the prediction of ALN metastasis in BC, and combining clinical and pathological features can help assessing the risk of lymph node metastasis in BC patients. This method provides clinicians with more information about ALNs, aiding in clinical decision-making.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-703/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-703/dss

Funding: This study was supported by the Medical Science Research Project of Hebei (No. 20201073) and 2024 Government-Funded Clinical Medicine Outstanding Talents Training Program (No. ZF2024116).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-703/coif). R.G. is from Siemens Healthineers AG. M.W. is from Siemens Healthineers Ltd. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Ethics Committee of The Fourth Hospital of Hebei Medical University (No. 2022037) and individual consent for this analysis was waived due to the retrospective nature.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Hu J, Xia X, Yang H, Yu Y. Dissection of Level III Axillary Lymph Nodes in Breast Cancer. Cancer Manag Res 2021;13:2041-6. [Crossref] [PubMed]
2. Murata T, Watase C, Shiino S, Kurita A, Ogawa A, Jimbo K, Iwamoto E, Yoshida M, Takayama S, Suto A. Development and validation of a pre- and intra-operative scoring system that distinguishes between non-advanced and advanced axillary lymph node metastasis in breast cancer with positive sentinel lymph nodes: a retrospective study. World J Surg Oncol 2022;20:314. [Crossref] [PubMed]
3. Yi C, Li L, Ma J. Tumor enhancement and shrinkage pattern in dynamic contrast-enhanced magnetic resonance imaging for predicting pathologic complete response after human epidermal growth factor receptor 2 (HER2)-targeted therapy in breast cancer. Quant Imaging Med Surg 2024;14:6734-44. [Crossref] [PubMed]
4. Zhu Y, Ma Y, Zhang Y, Liu A, Wang Y, Zhao M, Li H, He N, Wu Y, Ye Z. Radiomics nomogram for predicting axillary lymph node metastasis-a potential method to address the limitation of axilla coverage in cone-beam breast CT: a bi-center retrospective study. Radiol Med 2023;128:1472-82. [Crossref] [PubMed]
5. Li Z, Gao Y, Gong H, Feng W, Ma Q, Li J, Lu X, Wang X, Lei J. Different Imaging Modalities for the Diagnosis of Axillary Lymph Node Metastases in Breast Cancer: A Systematic Review and Network Meta-Analysis of Diagnostic Test Accuracy. J Magn Reson Imaging 2023;57:1392-403. [Crossref] [PubMed]
6. Sasada S, Masumoto N, Kimura Y, Kajitani K, Emi A, Kadoya T, Okada M. Identification of Axillary Lymph Node Metastasis in Patients With Breast Cancer Using Dual-Phase FDG PET/CT. AJR Am J Roentgenol 2019;213:1129-35. [Crossref] [PubMed]
7. Pesapane F, Mariano L, Magnoni F, Rotili A, Pupo D, Nicosia L, Bozzini AC, Penco S, Latronico A, Pizzamiglio M, Corso G, Cassano E. Future Directions in the Assessment of Axillary Lymph Nodes in Patients with Breast Cancer. Medicina (Kaunas) 2023;59:1544. [Crossref] [PubMed]
8. Rafn BS, Christensen J, Larsen A, Bloomquist K. Prospective Surveillance for Breast Cancer-Related Arm Lymphedema: A Systematic Review and Meta-Analysis. J Clin Oncol 2022;40:1009-26. [Crossref] [PubMed]
9. Davies JM, Knopman J, Mokin M, Hassan AE, Harbaugh RE, Khalessi A, et al. Adjunctive Middle Meningeal Artery Embolization for Subdural Hematoma. N Engl J Med 2024;391:1890-900. [Crossref] [PubMed]
10. Bharath S, Sharma D, Yadav SK, Shekhar S, Jha CK. A Systematic Review and Meta-analysis of Touch Imprint Cytology and Frozen Section Biopsy and Their Comparison for Evaluation of Sentinel Lymph Node in Breast Cancer. World J Surg 2023;47:478-88. [Crossref] [PubMed]
11. Ya G, Wen F, Xing-Ru L, Zhuan-Zhuan G, Jun-Qiang L. Difference of DCE-MRI Parameters at Different Time Points and Their Predictive Value for Axillary Lymph Node Metastasis of Breast Cancer. Acad Radiol 2022;29:S79-86. [Crossref] [PubMed]
12. Nissan N, Anaby D, Mahameed G, Bauer E, Moss Massasa EE, Menes T, Agassi R, Brodsky A, Grimm R, Nickel MD, Roccia E, Sklair-Levy M. Ultrafast DCE-MRI for discriminating pregnancy-associated breast cancer lesions from lactation related background parenchymal enhancement. Eur Radiol 2023;33:8122-31. [Crossref] [PubMed]
13. Gao Y, Heller SL. Abbreviated and Ultrafast Breast MRI in Clinical Practice. Radiographics 2020;40:1507-27. [Crossref] [PubMed]
14. Mann RM, Mus RD, van Zelst J, Geppert C, Karssemeijer N, Platel B. A novel approach to contrast-enhanced breast magnetic resonance imaging for screening: high-resolution ultrafast dynamic imaging. Invest Radiol 2014;49:579-85. [Crossref] [PubMed]
15. Heacock L, Gao Y, Heller SL, Melsaether AN, Babb JS, Block TK, Otazo R, Kim SG, Moy L. Comparison of conventional DCE-MRI and a novel golden-angle radial multicoil compressed sensing method for the evaluation of breast lesion conspicuity. J Magn Reson Imaging 2017;45:1746-52. [Crossref] [PubMed]
16. Mus RD, Borelli C, Bult P, Weiland E, Karssemeijer N, Barentsz JO, Gubern-Mérida A, Platel B, Mann RM. Time to enhancement derived from ultrafast breast MRI as a novel parameter to discriminate benign from malignant breast lesions. Eur J Radiol 2017;89:90-6. [Crossref] [PubMed]
17. Tofts PS, Brix G, Buckley DL, Evelhoch JL, Henderson E, Knopp MV, Larsson HB, Lee TY, Mayr NA, Parker GJ, Port RE, Taylor J, Weisskoff RM. Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging 1999;10:223-32. [Crossref] [PubMed]
18. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45.
19. Zhang J, Li L, Zhe X, Tang M, Zhang X, Lei X, Zhang L. The Diagnostic Performance of Machine Learning-Based Radiomics of DCE-MRI in Predicting Axillary Lymph Node Metastasis in Breast Cancer: A Meta-Analysis. Front Oncol 2022;12:799209. [Crossref] [PubMed]
20. Chen Y, Wang L, Dong X, Luo R, Ge Y, Liu H, Zhang Y, Wang D. Deep Learning Radiomics of Preoperative Breast MRI for Prediction of Axillary Lymph Node Metastasis in Breast Cancer. J Digit Imaging 2023;36:1323-31. [Crossref] [PubMed]
21. Kang SR, Kim HW, Kim HS. Evaluating the Relationship Between Dynamic Contrast-Enhanced MRI (DCE-MRI) Parameters and Pathological Characteristics in Breast Cancer. J Magn Reson Imaging 2020;52:1360-73. [Crossref] [PubMed]
22. Mann RM, Cho N, Moy L. Breast MRI: State of the Art. Radiology 2019;292:520-36. [Crossref] [PubMed]
23. Goto M, Sakai K, Yokota H, Kiba M, Yoshida M, Imai H, Weiland E, Yokota I, Yamada K. Diagnostic performance of initial enhancement analysis using ultra-fast dynamic contrast-enhanced MRI for breast lesions. Eur Radiol 2019;29:1164-74. [Crossref] [PubMed]
24. Lee SJ, Ko KH, Jung HK, Koh JE, Park AY. The additional utility of ultrafast MRI on conventional DCE-MRI in evaluating preoperative MRI of breast cancer patients. Eur J Radiol 2020;124:108841. [Crossref] [PubMed]
25. Chen M, Kong C, Lin G, Chen W, Guo X, Chen Y, Cheng X, Chen M, Shi C, Xu M, Sun J, Lu C, Ji J. Development and validation of convolutional neural network-based model to predict the risk of sentinel or non-sentinel lymph node metastasis in patients with breast cancer: a machine learning study. EClinicalMedicine 2023;63:102176. [Crossref] [PubMed]
26. Wu PQ, Guo FL, Wang J, Gao Y, Feng ST, Chen SL, Ma J, Liu YB. Development and validation of a dynamic contrast-enhanced magnetic resonance imaging-based habitat and peritumoral radiomic model to predict axillary lymph node metastasis in patients with breast cancer: a retrospective study. Quant Imaging Med Surg 2024;14:8211-26. [Crossref] [PubMed]