Can unenhanced MRI of the breast replace contrast-enhanced MRI in assessing response to neoadjuvant chemotherapy?

Abstract

Background

The goals of neoadjuvant chemotherapy (NAC) are to reduce tumor volume and to offer a prognostic indicator in assessing treatment response. Contrast-enhanced magnetic resonance imaging (CE-MRI) is an established method for evaluating response to NAC in patients with breast cancer.

Purpose

To validate the role of unenhanced MRI (ue-MRI) compared to CE-MRI for assessing response to NAC in women with breast cancer.

Material and Methods

Seventy-one patients with ongoing NAC for breast cancer underwent MRI before, during, and at the end of NAC. Ue-MRI was performed with T2-weighted sequences with iterative decomposition of water and fat and diffusion-weighted sequences. CE-MRI was performed using three-dimensional T1-weighted sequences before and after administration of gadobenate dimeglumine. Two blinded observers rated ue-MRI and CE-MRI for the evaluation of tumor response. Statistical analysis was performed to compare lesion size and ADC values changes during therapy, as well as inter-observer agreement.

Results

There were no statistically significant differences between ue-MRI and CE-MRI sequences for evaluation of lesion size at baseline and after every cycle of treatment (P > 0.05). The mean tumor ADC values at baseline and across the cycles of NAC were significantly different for the responder group.

Conclusion

Ue-MRI can achieve similar results to CE-MRI for the assessment of tumor response to NAC. ADC values can differentiate responders from non-responders.

Keywords

Breast neoplasms magnetic resonance imaging MRI chemotherapy

Introduction

Patients with locally advanced primary breast cancer are at increased risk of metastatic disease. These patients often receive neoadjuvant chemotherapy (NAC), even if the tumor is primarily operable (1). The principal goals of NAC are to reduce tumor volume and to offer a prognostic indicator in assessing treatment response (2). The possibility of reducing in size large tumors is frequently important in facilitating breast conservation surgery which might otherwise not be achievable, while early assessment of treatment response impacts positively on overall survival and outcome (3,4). Given the potential toxicity of chemotherapeutic drugs, early assessment of tumor response is particularly beneficial for patient management decisions, especially in cases of progressive disease. A further benefit of NAC is in providing an early opportunity to treat micrometastatic disease (5).

Currently, contrast-enhanced magnetic resonance imaging (CE-MRI) is an established method for evaluating response to NAC in patients with breast cancer. Several studies have demonstrated high correlation between tumor size measured at CE-MRI using RECIST 1.1 criteria (Response Evaluation Criteria in Solid Tumors) (6) and histopathological specimens obtained after NAC (7 –9). The European Society of Breast Imaging (EUSOBI) recommends CE-MRI before, during, and after NAC (10). Apart from providing morphologic information, the value of dynamic CE-MRI is that it permits evaluation of the functional properties of tumors, allowing detection of reduced enhancement washout independently from their reduction in size (11). Unfortunately, contrast agent use increases both examination time and costs; moreover, they cannot be administered in patients with gadolinium intolerance/allergies.

In recent years, the potential role of unenhanced MRI (ue-MRI) based only on diffusion-weighted imaging (DWI) and T2-weighted (T2W) sequences, has been described for mass lesion detection (12,13). Findings from these studies suggest that differentiation of benign from malignant breast lesions may be achievable with high sensitivity and specificity through a combination of DWI with T2W sequences. The potential advantages of this approach, apart from cost and time savings, are that breast MRI might still be feasible in patients not suitable for contrast agent administration and that patient comfort is increased as intravenous line positioning is no longer required.

Recently, T2-IDEAL (Iterative Decomposition of water and fat with Echo Asymmetry, GE Healthcare, Milwaukee, WI, USA) has been proposed for breast imaging (14). This approach is a chemical shift-based fat–water separation method that can be applied to a variety of MR pulse sequences and which allows the acquisition of four complete imaging datasets (T2W in-phase images, T2W out-of-phase images, T2 water-only images, and T2 fat-only images).

Ue-MRI using a combination of DWI and T2-IDEAL sequences may be a promising alternative to CE-MRI in breast imaging for the assessment of lesion size and tumor conspicuity. With this in mind, we compared ue-MRI using combined DWI and T2-IDEAL with CE-MRI for the assessment of response to NAC.

Material and Methods

Consecutive female patients with locally advanced breast cancer with indication to NAC were prospectively enrolled between January 2010 and September 2014. Women with contraindications for CE-MRI were excluded from the study. The study was approved by the ethical and protocol review committees of our institution and all patients provided written informed consent.

Patient treatment

Eighty-three patients with advanced breast cancer, proven by core biopsy, underwent breast CE-MRI before and at the end of each cycle of NAC; type and cycles number of NAC were scheduled on the basis of the molecular profile and tumor histology.

All patients underwent surgery within 20 days after the last MRI examination. Histopathological analysis of the surgical specimens was performed in all cases.

MRI response to NAC was assessed using RECIST 1.1 criteria (6). Complete pathological response was defined as the absence of invasive cancer within the specimen. Incomplete pathological response was defined as the presence of residual invasive carcinoma of any size.

Imaging procedures

Breast MRI was performed on a 3-T magnet (Discovery MR750, GE Healthcare) using an eight-channel bilateral breast surface coil with the patient in the prone position.

The MRI protocol included FSE T2-IDEAL sequences, DWI sequences, and Vibrant three-dimensional (3D) T1-weighted (T1W) (GE Healthcare) sequences acquired before and after the intravenous administration of gadolinium contrast agent (gadobenate dimeglumine [MultiHance]; Bracco Imaging SpA, Milan, Italy). The T2-IDEAL sequence was acquired with: TR/TE = 9128/120 ms; flip angle (FA) = 30°; matrix = 242 × 256 pixels; field of view (FOV) =340 mm; slice thickness = 3 mm; acquisition time =200 s. The multi-echo single-shot EPI-DWI sequence was acquired with: TR/TE = 6840–6236/51 ms; ETL = 37; matrix = 128 × 87; FOV = 36 × 36 cm; section thickness = 5 mm; intersection gap = 1.5 mm; NEX = 2; parallel imaging reduction factor = 2; b-values = 0, 800, 1000 mm²/s. The Vibrant fast 3D FS T1W sequence was acquired with: TR/TE = 7.7/4.2 ms; TI = 20; FA = 10°; matrix = 256 × 512; FOV =350 × 490 mm; slice thickness = 1.2 mm; acquisition time = 120 s before and 2, 4, 6, and 8 min after injection of gadobenate dimeglumine at a dose of 0.1 mmol/kg bodyweight. Gadobenate dimeglumine was injected at a flow rate of 2 mL/s through an antecubital venous access by means of a power injector (Spectris, Medrad, Warrendale, Palo Alto, CA, USA) and was followed by 20 mL of saline flush administered at the same rate.

Image analysis

Image analysis was performed on an offline dedicated workstation separately by two independent readers (with five and nine years of experience in breast MRI, respectively) who were fully blinded to all patient demographic and clinical information. In order to prevent a recall bias, ue-MR and CE-MR images of patients were presented randomly to the two independent readers. Initial evaluation was performed of combined ue-MRI images for tumor response which was measured in terms of tumor size, tumor appearance, and apparent diffusion coefficient (ADC) values. After an intervening period of at least two weeks, all CE-MRI images were evaluated using the same assessment criteria used for evaluation of the ue-MRI images.

Lesion size

Lesion size was determined separately on the T2-IDEAL sequence, on b-1000 DWI images and on VIBRANT T1W CE-MRI images (second post-contrast dynamic acquisition). Tumor size on T2-IDEAL sequence was expressed as the mean of the lesion sizes determined on the evaluation of the four imaging datasets. If multicentric or multifocal disease was present, the lesion with greatest dimensions was considered for size measurements.

Assessment of MR images was based on the Breast Imaging Reporting and Data System (BI-RADS) MRI lexicon (15). Lesions were considered as demonstrating mass or non-mass enhancement.

Apparent diffusion coefficient

ADC maps were generated from the DWI datasets of each study using a semi-automatic software tool (Functool; GE Healthcare) on an offline workstation (ADW 4.4; GE Healthcare) with a gray-scale indicating low ADC values in black. The ADC value was measured by placing a circular regions of interest (ROI) with a diameter of 6 mm within the darkest portion of individual lesions by two readers in consensus.

Evaluation of tumor response

The response to NAC was evaluated separately on the ue-MRI and CE-MRI sequences for each patient. Tumor response to NAC was classified using RECIST 1.1 criteria as: (i) complete response (absence of residual mass); (ii) partial response ( ≥ 30% reduction in tumor size); (iii) stable disease ( < 30% reduction or < 20% increase in tumor size); and (iv) progressive disease ( ≥ 20% increase in tumor size).

Based on these criteria, patients were classified as responders, partial responders, or non-responders (the latter including stable disease and progressive disease) on T2-IDEAL + DWI images and on dynamic CE-MRI images. Imaging results from the last MRI were then compared with findings of surgery obtained histological specimens. Pathological evaluation of response to NAC was conducted using Residual Cancer Burden Score of the MD Anderson Cancer Center.

In the last MR examination, for partial responders group, the shrinkage patterns of residual disease was also classified as concentric (without surrounding lesion) or scattered (shrinkage with residual multinodular lesions).

Statistical analysis

Data were collected using an electronic datasheet. The statistical analysis was performed using a dedicated statistical software package (SPSS Statistics 19 for Macintosh; IBM Company, Armonk, NY, USA) and Matlab 9 (the MathWorks, Natick, MA, USA) as deemed appropriate. Bland–Altman analysis was done using an additional tool by Ran Klein, the Bland–Altman and Correlation Plot v1.3 (https://uk.mathworks.com/matlabcentral/fileexchange/45049-bland-altman-and-correlation-plot?requestedDomain=www.mathworks.com).

All continuous values were expressed as mean ± SD. 95% confidence intervals (CI) were identified with statistical significance set at P < 0.05 for all tests performed. The Kolmogorov–Smirnov test was used to test data distribution. Bonferroni was used to correct multiple pairwise comparisons.

Lesion size on T2-IDEAL, DWI, combined T2-IDEAL + DWI sequences, and pathology were compared across cycles using Spearman's correlation coefficient. Moreover, Bland–Altman plots were used to test differences between baseline and after the second and sixth cycles of therapy. One-way ANOVA was used to test differences in percentage changes between the baseline scan and the follow-up scans. Inter-observer agreement concerning lesion size measurements was analyzed using Cohen κ coefficients in which κ values of 0.21–0.40 indicate poor agreement, 0.41–0.60 moderate agreement, 0.61–0.80 good agreement, and 0.81–1.00 excellent agreement.

A comparison of quantitative tumor ADC values at baseline and after NAC was likewise performed using a paired sample t-test. These results were stratified by dividing patients into two groups based on histopathological findings: group A (responders) and group B (partial responders, stable disease, and non-responders).

Results

Of 83 enrolled patients, 12 underwent surgery before completing all six cycles of NAC. These patients were then excluded from the analysis in order to homogenize as much as possible the patient population. Seventy-one consecutive female patients (mean age = 44 ± 10 years; age range = 34–74) that completed all six cycles of NAC were then considered the final study population. Nine (12.9%) of these patients were post-menopausal, 18 (25.8%) were peri-menopausal, and 43 (61.3%) were pre-menopausal. Histological analysis determined 64 (94%) lesions to be invasive ductal carcinoma (IDC) and seven (10%) to be invasive lobular carcinoma (ILC). Multicentric and multifocal disease was observed in 15 and 11 patients, respectively.

Image analysis

Evaluation of tumor response

Based on histopathological assessment after surgery, 36/71 (51%) patients were classified as complete responders to NAC (Figs. 1 and 2), while 32/71 (45%) patients were considered partial responders (Figs. 3 and 4) and 3/71 (4%) non-responders. In the partial responders group, the shrinkage pattern of the lesion was classified as “scattered” in 8/32 patients (25%) and “concentric” in 24/32 patients (75%).

Fig. 1.

A 48-year-old woman with IDC of the left breast. Baseline breast MRI (before NAC). (a) DWI image (b-1000) shows a 2.8-cm area of hyperintensity in the retro-areolar region. In the homolateral axilla, another area of hyperintensity measuring 1.3 cm is detected. (b) T2W-IDEAL image with fat saturation reveals a parenchymal distortion in the retro-areolar region and an abnormal lymph node in the homolateral axilla. (c) T1W CE-MRI subtracted image shows a mass enhancement lesion of 2.9 cm in the retro-areolar region.

Fig. 2.

Same patient as in Fig. 1 after six cycles of NAC. (a) DWI image (b-1000) shows a reduction of the hyperintense retro-areolar area and the lymph node. (b) T2W-IDEAL sequence with fat saturation shows that the parenchymal distortion is no longer visible and the size of the lymphoadenopathy is reduced. (c) On T1W CE-MRI subtracted image, no enhancement is visible at the end of NAC. After surgery, pathology confirmed the complete response.

Fig. 3.

A 67-year-old woman with biopsy-proven IDC of the right breast. Baseline MRI. (a, b) DWI and water-only T2W-IDEAL images show a 50-mm neoplastic mass located between upper quadrants. Water-only images show diffuse edema surrounding the lesion and at the level of the skin and the pectoral muscle. (c) T1W CE-MRI subtracted image confirm a 53-mm enhancing mass.

Fig. 4.

Same patient as in Fig. 3 after six cycles of NAC. (a, b) DWI and T2W-IDEAL images show the persistence of a 33-mm tumor; water-only images show a reduction of the edema surrounding the lesion and within the pectoral muscle, with persistent edema at the level of the skin. (c) T1W CE-MRI subtracted image confirms the 53-mm enhancing mass. Histopathology confirmed a 40% reduction of the lesion (partial responder).

Lesion size and characteristics

Overall, 56/71 (79%) lesions were classified as mass lesions. The remaining 15/71 (21%) lesions demonstrated non-mass enhancement.

The mean size of lesions on different sequences at baseline and during NAC is shown in Table 1. The Kolmogorov–Smirnov test showed that data distributions were not Gaussian. Figure 5 illustrates the results of the one-way ANOVA for T2-IDEAL, CE-MRI, DWI, and T2-IDEAL + DWI; there were no statistically significant differences between those modalities across all the time points (P > 0.05). The correlation between the MRI derived parameters at baseline was good (r² = 0.65) after the second cycle and moderate (r² = 0.46) after the sixth cycle. The Bland–Altman analysis showed a good agreement between the measures (Fig. 6).

Table 1.

The mean size of lesions (mm) on different sequences at baseline and during NAC is shown.

	Baseline	2 cycles of NAC	4 cycles of NAC	6 cycles of NAC
CE-MRI	29 ± 18	18.7 ± 15.9	13 ± 14	8.7 ± 11
T2-IDEAL	28 ± 17.9	17.7 ± 15.7	11.45 ± 11.6	8.2 ± 11.3
DWI	29.5 ± 18.7	19.8 ± 17.1	13.25 ± 12.8	8.7 ± 12
T2 + DWI	28.8 ± 18.2	18.8 ± 15.9	13.26 ± 14	8.4 ± 11.5

NAC, neoadjuvant chemotherapy.

Fig. 5.

Analysis of differences expressed in percentage across the time points.

Fig. 6.

Correlation between MRI parameters at baseline after the second cycle (a) and the six cycle (c); (b, d) Bland–Altman plots of the same time points.

ADC quantitative analysis

Quantitative analysis of DWI ADC values was performed at time zero and after two, four, and six cycles of NAC. The mean ADC values at time zero and across the cycles of NAC for the whole study group were significantly different (Table 2).

Table 2.

The mean ADC values before and after NAC are shown for the whole study group as well as for responders and non-responders group.

	ADC values (mm²/s)
	Whole study group	Responders	Non- responders
Before NAC	1.2 ± 0.2	1.3 ± 0.3	1.2 ± 0.2
After 2 cycles of NAC	1.2 ± 0.6	1.7 ± 0.2	1.1 ± 0.6
After 4 cycles of NAC	1.0 ± 0.7	2.0 ± 0.2	1.0 ± 0.7
After 6 cycles of NAC	0.9 ± 0.7	1.7 ± 0.3	0.9 ± 0.7
P value	0.007	0.003	0.08

NAC, neoadjuvant chemotherapy.

Stratification of ADC findings into group A (responders) and group B (non-responders plus partial responders) revealed a significant difference before and after NAC for group A (P = 0.003), but not for group B (P = 0.08; Table 2).

Discussion

Breast MRI is the most accurate diagnostic method for the assessment of breast disease (17 –19). Breast MRI is considered the most reliable method for monitoring tumor response to chemotherapy and predicting the extent of residual disease after NAC (20 –22). One of the main issues of the technique at this time is related to the use of contrast agent. Even if gadolinium-based contrast agents are widely used and have been considered to have a wide safety margin in patients with normal renal function, a deposition of gadolinium in different tissues have been described, such as kidneys (with consequent potential development of nephrogenic systemic fibrosis), bone matrix, skin, and brain (23,24). This deposition could increase with repeated gadolinium contrast administrations, despite an intact blood–brain barrier and normal renal function (25). Moreover, there are other advantages in avoiding repeated contrast agent administrations. Indeed, a reduction in the use of contrast agents will determine a significant shortening in both acquisition time and reading time, with consequential decrease in costs. Although it may not be as relevant in this setting as it may be for example for screening purposes, this is still an advantage of this approach on a macro-scale. Lastly, patient comfort may be increased as IV line positioning is no longer required and, therefore, patient compliance could increase.

In recent years, the potential role of ue-MRI based only on DWI and T2W sequences has been described for mass lesion detection (12,13). Available data suggest that a combination of DWI and T2W TSE may offer potential for the differentiation of benign from malignant breast lesions with high sensitivity and specificity.

It is well-known that CE-MRI is essential for the evaluation of size and the vascular patterns of breast lesions. Nevertheless, our findings suggest that the morphological information from T2W images together with the cellular information available on DWI images is able to provide similarly useful information for the accurate evaluation of tumor response.

Analysis of the four different images obtained with the T2-IDEAL sequence gives important data concerning lesion size measurements and the detection of associated features (e.g. the presence of tissue edema). The main limitation of T2W imaging is often seen in the evaluation of non-mass enhancement (NME), frequently represented by ductal carcinoma in situ (DCIS). However, Rahbar et al. (26) showed that T2W images combined with DWI can play an important role in detecting DCIS. In their study, Rahbar et al. retrospectively reviewed 74 pure DCIS and observed that 96% of these lesions demonstrated greater qualitative and quantitative DWI intensity and lower ADC values than that of normal breast tissue. The study reported a 95% detection rate of DCIS supporting DWI as a potential method for early breast cancer detection without contrast material injection. In our study, four cases of invasive carcinomas presenting as NME were underestimated compared to CE-MRI when reading the T2-IDEAL alone; in particular, T2-IDEAL was able to detect the tumor at baseline, but it failed to identify the residual pathologic tissue at MRI before surgery. This limit was overcome by the association of T2-IDEAL and DWI that correctly depicted the residual tumor in all cases. When evaluating T2-IDEAL images together with DWI for lesion size assessment, we obtained better results at all time points.

We also demonstrated good inter-observer reproducibility of size measurements suggesting that this technique may be appropriate for clinical routine for the follow-up of patients undergoing NAC.

In the present study, we reported a change of ADC values in patients undergoing NAC, with a significant difference between ADC values before and after NAC (Table 2). When stratifying our patients into two groups based on response to NAC treatment (responders in group A, non-responders in group B), the difference in ADC values between the two groups was even more relevant. These data correlate well with literature reports (27 –29) and confirm that ADC values are a useful indicator for the evaluation of response to NAC.

Despite these promising results, our study has some limitations. First, ADC measurements require manual placement of the ROI which makes the method highly operator-dependent (30). Moreover, lesion heterogeneity and the larger spatial distribution of non-mass enhancement could decrease the reproducibility of ADC values (31).

Second, our study population was relatively small which prevented us from analyzing potential differences related to different chemotherapeutic drugs and different tumor's histotypes. For example, despite the large number of complete response to therapy, we had only 2/7 cases of complete response in ILCs. It is true that, in general, invasive lobular breast cancers are less likely to experience Pathological Complete Response (pCR) after chemotherapy and that physicians currently give much importance to hormone receptor assay more than histological type of tumor. Anyway, given the very small number of lobular cancers in the present study, we cannot state whether response to therapy of these cases depended on the histotype more than the hormone receptor assay.

Nevertheless, the promising results obtained in this study deserve validation in a larger prospective study.

In conclusion, the combination of T2-IDEAL for lesion morphology and size assessment and DWI for assessment of tumor cellularity and cell membrane integrity is potentially a useful approach for the follow-up of tumor response during NAC, while CE-MRI remains indispensable for depiction and first staging of the disease. The combination of T2-IDEAL and DWI sequences allows assessment of the response to chemotherapy with MRI. The similar results obtained for ue-MRI and CE-MRI suggest that the contrast agent-free approach may be a good option in patients with locally advanced breast cancer undergoing NAC, saving time and costs and preventing potential tissue damage due to repeated injections of contrast agents.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Fangberget

Nilsen

Hole

et al.

Neoadjuvant chemotherapy in breast cancer-response evaluation and prediction of response to treatment using dynamic contrast-enhanced and diffusion-weighted MR imaging. Eur Radiol 2011; 21: 1188–1199.

Woodhams

Kakita

Hata

et al.

Identification of residual breast carcinoma following neoadjuvant chemotherapy: diffusion -weighted imaging - comparison with contrast-enhanced MR imaging and pathologic findings. Radiology 2010; 254: 357–366.

Sharma

Danishad

Seenu

et al.

Longitudinal study of the assessment by MRI and diffusion-weighted imaging of tumor response in patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy. NMR Biomed 2009; 22: 104–113.

Nilsen

Fangberget

Geier

et al.

Diffusion-weighted magnetic resonance imaging for pretreatment prediction and monitoring of treatment response of patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy. Acta Oncol 2010; 49: 354–360.

Fisher

Brown

Mamounas

et al.

Effect of preoperative chemotherapy on local-regional disease in women with operable breast cancer: findings from National Surgical Adjuvant Breast and Bowel Project B-18. J Clin Oncol 1997; 15: 2483–2493.

Eisenhauer

Therasse

Bogaerts

et al.

New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009; 45: 228–247.

Denis

Desbiez-Bourcier

Chapiron

et al.

Contrast enhanced magnetic resonance imaging underestimates residual disease following neoadjuvant docetaxel based chemotherapy for breast cancer. Eur J Surg Oncol 2004; 30: 1069–1076.

Liu

. Effectiveness of dynamic contrast-enhanced magnetic resonance imaging in evaluating clinical responses to neoadjuvant chemotherapy in breast cancer. Chin Med J (Engl) 2011; 124: 194–198.

Rosen

Blackwell

Baker

et al.

Accuracy of MRI in the detection of residual breast cancer after neoadjuvant chemotherapy. Am J Roentgenol 2003; 181: 1275–1282.

10.

Mann

Kuhl

Kinkel

et al.

Breast MRI: guidelines from the European Society of Breast Imaging. Eur Radiol 2008; 18: 1307–1318.

11.

Tudorica

Chui

et al.

Early prediction and evaluation of breast cancer response to neoadjuvant chemotherapy using quantitative DCE-MRI. Transl Oncol 2016; 9: 8–17.

12.

Yabuuchi

Matsuo

Sunami

et al.

Detection of non-palpable breast cancer in asymptomatic women by using unenhanced diffusion-weighted and T2-weighted MR imaging: comparison with mammography and dynamic contrast-enhanced MR. imaging. Eur Radiol 2011; 21: 11–17.

13.

Baltzer

Benndorf

Dietzel

et al.

Sensitivity and specificity of unenhanced MR mammography (DWI combined with T2-weighted TSE imaging, ueMRM) for the differentiation of mass lesions. Eur Radiol 2010; 20: 1101–1110.

14.

Tagliafico

Bignotti

Tagliafico

et al.

Breast density assessment using a 3T MRI system: comparison among different sequences. PLoS One 2014; 9: e99027.

15.

D'Orsi

Sickles

Mendelson

et al.

Breast Imaging Reporting and Data System: ACR BI-RADS Breast Imaging Atlas, 5th ed. Reston, VA: ACR, 2013.

16.

Huang

Rooney

. Signal-to-noise ratio, contrast-to-noise ratio and pharmacokinetic modeling considerations in dynamic contrast enhanced magnetic resonance imaging. Magn Reson Imaging 2012; 30: 1313–1322.

17.

Kuhl

. The current status of breast MR imaging. Part I. Choice of technique, image interpretation, diagnostic accuracy, and transfer to clinical practice. Radiology 2007; 244: 356–378.

18.

Kuhl

. Current status of breast MR imaging. Part 2. Clinical applications. Radiology 2007; 244: 672–691.

19.

Sardanelli

Podo

D'Agnolo

et al.

Multicenter comparative multimodality surveillance of women at genetic-familial high risk for breast cancer (HIBCRIT study): interim results. Radiology 2007; 242: 698–715.

20.

Marinovich

Houssami

Macaskill

et al.

Meta-analysis of magnetic resonance imaging in detecting residual breast cancer after neoadjuvant therapy. J Natl Cancer Inst 2013; 105: 321–333.

21.

Chen

. Clinical application of magnetic resonance imaging in management of breast cancer patients receiving neoadjuvant chemotherapy. Biomed Res Int 2013; 2013: 348167.

22.

Prati

Minami

Gornbein

et al.

Accuracy of clinical evaluation of locally advanced breast cancer in patients receiving neoadjuvant chemotherapy. Cancer 2009; 115: 1194–1202.

23.

Fraum

Ludwig

Bashir

et al.

Gadolinium-based contrast agents: a comprehensive risk assessment. J Magn Reson Imaging 2017; 46: 338–353.

24.

Stojanov

Aracki-Trenkic

Benedeto-Stojanov

. Gadolinium deposition within the dentate nucleus and globus pallidus after repeated administrations of gadolinium-based contrast agents-current status. Neuroradiology 2016; 58: 433–441.

25.

Ramalho

Semelka

Ramalho

et al.

Gadolinium-based contrast agent accumulation and toxicity: an update. AJNR Am J Neuroradiol 2016; 37: 1192–1198.

26.

Rahbar

Partridge

Eby

et al.

Characterization of ductal carcinoma in situ on diffusion weighted breast MRI. Eur Radiol 2011; 21: 2011–2019.

27.

Iacconi

Giannelli

Marini

et al.

The role of mean diffusivity (MD) as a predictive index of the response to chemotherapy in locally advanced breast cancer: a preliminary study. Eur Radiol 2010; 20: 303–308.

28.

Woodhams

Matsunaga

Kan

et al.

ADC mapping of benign and malignant breast tumors. Magn Reson Med Sci 2005; 4: 35–42.

29.

Hatakenaka

Soeda

Yabuuchi

et al.

Apparent diffusion coefficients of breast tumors: clinical application. Magn Reson Med Sci 2008; 7: 23–29.

30.

Clauser

Marcon

Maieron

et al.

Is there a systematic bias of apparent diffusion coefficient (ADC) measurements of the breast if measured on different workstations? An inter- and intra-reader agreement study. Eur Radiol 2016; 26: 2291–2296.

31.

Imamura

Isomoto

Sueyoshi

et al.

Diagnostic performance of ADC for Non-mass-like breast lesions on MR imaging. Magn Reson Med Sci 2010; 9: 217–225.