Accepted Manuscript

Title: Magnetic resonance imaging in breast cancer

management in the context of neo-adjuvant chemotherapy

Authors: P. Taourel, E. Pages, I. Millet, C. Bourgier, P.

Rouanet, W. Jacot, P. Crochet, D. Azria

PII: S1040-8428(18)30035-0
DOI: https://doi.org/10.1016/j.critrevonc.2018.09.012
Reference: ONCH 2622

To appear in: Critical Reviews in Oncology/Hematology

Received date: 23-1-2018

Revised date: 31-8-2018
Accepted date: 19-9-2018

Please cite this article as: Taourel P, Pages E, Millet I, Bourgier C, Rouanet P, Jacot
W, Crochet P, Azria D, Magnetic resonance imaging in breast cancer management in
the context of neo-adjuvant chemotherapy, Critical Reviews in Oncology / Hematology
(2018), https://doi.org/10.1016/j.critrevonc.2018.09.012

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
1.1 Magnetic resonance imaging in breast cancer management in the

context of neo-adjuvant chemotherapy

P. Taourel1,*, E. Pages1, I. Millet1, C. Bourgier2, P. Rouanet2, W. Jacot2, P. Crochet1, D.

Azria2

T
IP
1 2
CHU Montpellier, Univ Montpellier, Montpellier, France ; ICM, Univ Montpellier,

R
INSERM U1194, Montpellier, France

SC
*Corresponding Author:

U
Patrice Taourel, CHU Montpellier, 34000 Montpellier, France.
N
Phone: + 33 4 67 33 86 01, Fax: + 33 4 67 33 22 92
A
E-mail: [email protected]
M

Abstract
ED

This review discusses the clinical applications of magnetic resonance imaging (MRI) for the

assessment of neo-adjuvant chemotherapy (NAC) indication, axillary lymph node status,

PT

preNAC cancer prognosis, early and intermediate response to NAC, and post-NAC residual

disease in patients with breast cancer. Contrast-enhanced MRI with analysis of the tumor
E

morphological features and qualitative enhancement kinetics must be considered as the

CC

standard method for pre-NAC breast cancer staging and post-NAC residual disease
A

assessment. Diffusion-weighted imaging (DWI) is easy to perform and may increase the

specificity of breast MRI for tumor staging, and also for the assessment of tumor multifocality

and multicentricity and lymph node status. It also provides an ancillary added value in the

early and post-NAC response evaluation. Changes in the functional tumor volume are the

1
main criterion for the early response analysis. Other MRI methods, such as quantitative

perfusion analysis, MR spectroscopy and texture analysis, are still under study.

Keywords: MRI; neo-adjuvant chemotherapy; breast cancer

Neo-adjuvant chemotherapy (NAC) in breast cancer is usually indicated in two situations:

T
Downstaging locally advanced disease or inflammatory breast cancer and make it eligible for

IP
surgery (usually mastectomy) or decreasing the size of breast cancer to allow breast

R
conservative. Currently, NAC is performed in about 10% of patients with invasive breast

SC
cancer. As it represents an important alternative treatment modality for breast cancer, the

selection of patients eligible for NAC has become part of the routine patient management by a

multidisciplinary team. U
N
Magnetic resonance imaging (MRI) is recommended before and after NAC. Conversely, the
A

relevance of additional MRI examinations during NAC to assess the early and intermediate
M

tumor responses to therapy is still under scientific evaluation [1].

ED

In the present review, five major issues concerning MRI use in the context of NAC are

discussed: assessment of NAC indications, baseline evaluation of the axillary lymph node
PT

status, baseline disease prognosis estimation, assessment of the early response to NAC, and

identification of residual tumor disease after NAC.

E
CC

2. MRI for breast cancer staging in patients with scheduled neoadjuvant chemotherapy
A

2.1 Background

One of the two major indications of NAC is when conventional imaging shows that the breast

tumor is unicentric, but too large to be treated by breast-conserving therapy. In this context,

MRI has two roles: i) to accurately measure the tumor size, because an unfavorable tumor

2
size/breast size ratio could jeopardize the breast-conserving therapy, and ii) to detect the

presence of multicentric lesions that are sometimes a contraindication to breast-conserving

surgery, even after NAC.

2.2 Overview of evidences

The superiority of MRI, compared with mammography or sonography, for measuring the real

clinical size of breast tumors is acknowledged [2, 3]. However, there are still discrepancies

T
between the MRI data and the pathological tumor size at surgery, especially in the case of

IP
large tumors. Tumor size overestimation is more common than underestimation with a wrong

R
estimation rate that depends on the chosen threshold of discordance between MRI and

SC
pathology size. When the discordance is defined as a difference greater than 10 mm between

U
MRI and pathology size, the tumor size underestimation and overestimation rates are 4.6%
N
and 7.5%, respectively. These rates increase to 19% and 36%, respectively, when the
A
threshold is set at 5 mm [4]. This could increase the mastectomy rate [5], and could lead to
M

more (not required) indications for NAC. However, the correlation between pre-operative

MRI and pathology tumor size in patients referred for NAC cannot be assessed. And, more
ED

importantly, the efficiency of aesthetically satisfactory breast-conserving surgery is linked not

only to the tumor size, but also to the breast size, the tumor location and the surgeon’s
PT

proficiency. This underlines the need for the conversion of the MRI tumor data to obtain a 3D
E

volume representation of the tumor within the breast (Fig. 1).

CC

As reported by the EUSOMA study group [1], single [6, 7] and multi-center studies [8] and

meta-analyses [9-11] have confirmed that MRI is more sensitive than conventional imaging
A

for the detection of multifocal and multicentric lesions. In a recent meta-analysis [11] that

reviewed 50 articles and included 10811 women, MRI detected additional ipsilateral lesions

in 20% of women with a summary positive predictive value (PPV) of 66%. However, due to

the high (34%) false-positive rate, the identified lesions should always be pathologically

3
verified. For instance, in the case of scheduled NAC, any additional lesion should be biopsied

before starting NAC because multicentric breast cancer might require mastectomy, regardless

of NAC effectiveness.

MRI for routine staging of patients with breast cancer is not recommended, and preoperative

MRI is restricted to specific patient subgroups, for instance patients scheduled for NAC with

operable breast cancer. The EUSOMA working group [1] recommends that in patients with a

T
large potentially operable breast cancer, breast MRI should be performed prior to NAC on

IP
condition that this examination does not markedly delay its initiation. This recommendation is

R
based on the fact that MRI can improve tumor staging through a better assessment of tumor

SC
size, multifocality (Fig. 2) and multicentricity (possible NAC contraindications), the

U
identification of a contralateral breast cancer (Fig. 3) (in about 3% of patients) [10], and a
N
better identification of the pectoral muscle or the skin involvement in patients with a big
A
tumor (Fig. 4). This strategy is also supported by the potential difficulty in interpreting the
M

post-NAC breast MRI data to assess the tumor response. Indeed, as noted by the EUSOMA

working group [1], even a very low-level contrast enhancement at the primary tumor site
ED

should be considered as a sign of residual disease. Thus, the comparison of pre- and post-

NAC MRI data is necessary for the detection and interpretation of subtle findings in the post-
PT

NAC MRI images.

E

2.3 Research issues

CC

In the clinical setting of a scheduled NAC, it is essential to assess MRI impact on breast

cancer staging (tumor size, multicentricity and bilaterality assessment) and generally on
A

changes in tumor management, despite the difficulty to accurately determine this effect when

feasible correlations with the pathology data are lacking.

The value of computer-aided diagnosis systems for the automated determination of the tumor

and breast volumes has not been assessed yet. Moreover, the usefulness of standardizing the

4
3D volume representation in patients with big tumors should also be evaluated to define

guidelines for providing images with the MRI report.

Finally, the potential benefits of performing diffusion-weighted imaging (DWI) as a

supplement to dynamic contrast-enhanced MRI (DCE-MRI) should be investigated. A

metaanalysis showed that DWI diagnostic value in terms of sensitivity to predict malignancy

was comparable to that of DCE-MRI [12]. However DWI could increase MRI specificity in

T
breast cancer staging by detecting additional enhanced areas (Fig. 2).

R IP
3. MRI for assessing the regional lymph node status

SC
3.1 Background

U
The percentage of patients with positive lymph nodes is high among those with NAC
N
indication. In a prospective multicenter study that included 2234 patients scheduled for NAC,
A
40% of patients had clinically node-positive breast cancer, and 34% of patients with clinically
M

negative-node cancer had a positive sentinel lymph node biopsy (SLND) [13]. This is

particularly true in patients with inoperable breast cancer as 55 to 85% of them will present
ED

with axillary or supraclavicular lymph nodes metastases [14]. In this setting, the main

advantage of MRI is to permit a regional lymph node assessment.

PT

It may be useful to take into account the axillary lymph node status before NAC for two main
E

reasons: (i) after NAC, the false-negative rate of SLND is higher than that of axillary lymph
CC

node dissection (ALND) [13], (ii) pre-NAC lymph node status assessment provides useful

information to guide the choice of NAC regimen and particularly for decision-making on
A

regional radiation [15], although the benefit of complementary lymph node radiotherapy on

initially positive lymph nodes but sterilized by neo adjuvant treatment is still controversial.

Nevertheless, the use of SLND before NAC with women with positive SLND treated by

axillary node dissection after chemotherapy is no longer a standard. The SENTINA and

5
ACOSOG Z1071 trials [13, 16] have shown that SNB post-NAC could be achieved with an

acceptable false negative rate providing that more than two sentinel lymph nodes were

sampled and a dual tracer was used. Post-NAC MRI assessment should be helpful in this

setting to identify metastatic lymph nodes allowing surgeons to proceed directly with ALND.

MRI evaluation of axillary lymph nodes has two potential advantages. First, the MRI

identification of suspicious lymph nodes that could be sampled by less invasive image-guided

T
biopsy techniques could avoid SLND. On the contrary, it may also motivate multidisciplinary

IP
teams to extend the axillary dissection before radiotherapy. Second, MRI can indicate

R
precisely the anatomical location of invaded lymph nodes using the Berg’s [17] categorization

SC
of axillary lymph nodes relative to the pectoralis minor muscle. Specifically, lymph nodes are

U
categorized in three surgical levels: level I (lymph nodes located below the lower edge of the
N
pectoralis minor muscle), level II (lymph nodes posterior/underneath the pectoralis minor
A
muscle), and level III (lymph nodes medial/above the medial margin of the pectoralis minor
M

muscle) (Fig. 5). As most axillary metastases involves level I and level II lymph nodes,

axillary lymph node dissection typically does not involve removal of level III nodes.
ED

Therefore, preoperative imaging assessment of suspicious level III nodes is of particular

clinical interest [18].

PT

Lastly, the detection of internal mammary adenopathy may require internal mammary chain
E

radiation therapy, a treatment that increases cardiac toxicity but improves overall survival at
CC

ten years. The prevalence of internal mammary adenopathy increases with the tumor size, the

presence of axillary adenopathy, lymphovascular invasion, younger patient age and inner
A

quadrant tumors [19, 20]. Generally, internal mammary adenopathy is more common in

patients with advanced breast tumors who often receive NAC. In a series of 90 consecutive

patients with T2-T3 N0 or T1-T3 N1-N3 breast cancer referred for NAC, the internal

mammary adenopathy prevalence revealed by MRI was 16% [21].

6
3.2 Overview of evidences

When an axillary or internal mammary adenopathy is detected, sentinel lymph node biopsy is

replaced by imaging-guided biopsy, and internal mammary chain radiation therapy is

prescribed. Therefore, the presence of abnormal lymph nodes must be evaluated when MRI is

used for breast cancer staging. However, benign axillary and even internal mammary

adenopathy are very common. Irregular borders, high contrast enhancement, round hila with

T
abnormal thickened cortices and lack of fatty hilum are among the features that best

IP
distinguish malignant from benign lymph nodes [22]. In a group of 74 patients evaluated

R
before NAC, Javid et al. [23] considered a lymph node size bigger than 1 cm, absence of fatty

SC
hilum, thickened cortex, and irregular node contours as malignant lymph node criteria.

U
According to these authors, the MRI-based prediction of axillary lymph node status using
N
these criteria shows moderate sensitivity (65%) and very high specificity (100%). Similarly,
A
de Felice et al. [24] reported moderate sensitivity (76%) and very high specificity (99%) of
M

breast MRI by using very simple features, such as a >5 mm short axis and the absence of

hilum, as malignancy criteria (Fig. 6). Differently from morphologic criteria, contrast kinetic
ED

features are not useful for lymph node discrimination because many normal and malignant

lymph nodes exhibit similar quantitative kinetic features, particularly contrast washout [22].
PT

By comparison to Ultrasound, MRI has advantages and drawbacks. The advantages are the
E

possibility to precisely compare the homo and contro-lateral axilla and the fact that it is less
CC

operator-dependent, particularly to determine the anatomical level of involved lymph nodes

and to detect N2 and N3 lymph nodes. The advantages of US are its cost-effectiveness and the
A

possibility of US-guided core biopsy if a suspicious lymph node is found. Comparison of

diagnostic performances between US, MRI and PET/CT in the evaluation of lymph nodes

status showed that MRI had a higher sensitivity and a lower specificity than US [25]. So, MRI

and ultrasonography may be considered as complementary modalities for axillary node

7
evaluation. Adding MRI to ultrasonography and ultrasound-guided fine-needle aspiration can

reduce the false-negative rate in axillary nodal metastasis diagnosis in patients with breast

cancer (particularly N2 and N3). On the other hand, targeted ultrasonography examination and

ultrasound-guided core biopsy must be performed after identification of a suspicious axillary

lymph node on MRI. PET/CT had the lowest sensitivity but the best specificity [26]. Its main

advantage is the evaluation of regional lymph nodes, including N3 axillary, internal mammary

T
and supraclavicular nodes [27].

IP
3.3 Research issues

R
Preliminary studies have shown DWI potential for the detection of metastatic disease in

SC
lymph nodes [27, 28]; however, contradictory results have been published on the usefulness

U
of apparent diffusion coefficient (ADC) measurements for discriminating lymph nodes
N
identified on breast MRI [29, 30]. Therefore, DWI cannot be considered as a standard
A
approach in regional lymph node evaluations in patients selected for NAC until more
M

consistent data are published. Moreover, although a study published 10 years ago showed that

ultra-small superparamagnetic iron oxide-enhanced MRI improves the accuracy of axillary

ED

lymph node metastasis detection compared with non-enhanced MRI [31], this contrast agent

is not yet commonly used for axillary staging in patients with breast cancer.
PT

4. Prognosis estimation before neo-adjuvant chemotherapy

E

4.1 Background
CC

Tumor prognosis estimation, in terms of recurrence and overall survival, is currently based on

traditional indicators, such as disease stage that includes tumor size and lymph node status. In
A

patients selected for NAC, the response to this treatment also is a major prognostic indicator.

Over the last decade, tumor immunopathological features have been used to classify breast

cancers in four main tumor subtypes (luminal A, luminal B, HER2 and basal types) that are

characterized by different prognoses and responses to chemotherapy [32, 33]. Anatomical and

8
functional MRI assessments, like immunopathological evaluations, could contribute to refine

the different prognostic profiles, on the basis of morphologic, perfusion diffusion and

spectroscopic parameters.

4.2 Overview of evidences

The value of different MRI parameters as pre-NAC prognostic factors has been evaluated

using different MRI methods. However, the results are considered controversial, thus

T
complicating their use for determining the optimal NAC indications.

IP
3.2.1 Morphologic tumor factors

R
Some tumor morphologic features have been associated with the treatment response.

SC
Particularly, absence of mass effect and very high intratumoral signal intensity on T2

U
weighted MR images, which corresponds to intratumoral necrosis, are significantly associated
N
with chemoresistance [34]. However, in a systematic review of 15 studies, BI-RADS
A
descriptors were not predictive of the response to NAC [35]. It would surely be interesting to
M

identify morphological predictive factors of the response to NAC in specific cancer subtypes,

particularly because the MRI morphological pattern of breast cancers varies according to the
ED

molecular subtype. For instance, it was shown that an irregularly shaped mass and the

presence of intratumoral necrosis are significantly associated with poor responses to NAC in
PT

triple-negative breast cancers (Fig. 7) [36].

E

3.2.2 Contrast enhancement kinetic parameters

CC

Tumor prognosis could rely on the identification of contrast enhancement kinetic parameters

that are associated with a higher likelihood of metastases. The main hypothesis is that higher
A

tumor angiogenesis leads to higher risk of cancer cell dissemination via the systemic

circulation. Contrast enhancement kinetics is typically evaluated on the basis of three features:

wash-in phase, maximal enhancement, and wash-out phase. A recent study [37] showed that

the maximal enhancement index is a predictive indicator of reduced survival, providing

9
independent prognostic information that could complement traditional and histological

indicators. Furthermore, in pre-NAC DCE-MRI breast images, increased volume washout and

total enhancing volume in the tumor are predictors of poor survival [38]. As higher tumor

angiogenesis is characterized by increased microvasculature with contrast agent leakage

through the vessel wall into the interstitium, two-compartment models that include the

vascular and interstitial spaces could help to describe the quantitative pharmacokinetic

T
parameters. Specifically, the Ktrans exchange rate constant describes leakage from the

IP
vasculature to the interstitial space, while the kep rate constant describes leakage from the

R
interstitial space back to the vascular space, and Ve represents the distribution volume in the

SC
interstitial space (Fig. 8). Although theoretically promising, contradictory results were

U
obtained by using the Ktrans, Kep and Ve values to differentiate between NAC responders and
N
non-responders [35]. This could be explained by different factors: tumor type, problems
A
linked with the arterial input function measurement, dosage of the contrast agent and
M

variations in the used MRI sequences. Furthermore, this analysis requires high temporal

resolution sequences that are not used in the daily practice.

ED

3.2.3 Motion of water molecules measured by diffusion-weighted magnetic resonance

imaging (DW-MRI)
PT

DW-MRI is a technique where the differences in the diffusion rate of water molecules in vivo
E

are exploited to create images with different contrast. Tumors with high cell density, and
CC

therefore with very low ADC, respond better to NAC [39, 40]. Conversely, tumor necrosis,

which may reduce the delivery of chemotherapeutic drugs to the tumor, leads to higher ADC
A

values (Fig. 7). Although these results are promising, not all studies agree that the

pretreatment ADC value is predictive of the clinical response to NAC [41].

3.2.4 Texture analysis

10
Texture analysis is used to describe the MRI appearance of tissues and its changes in terms of

fineness, coarseness, smoothness, granularity, homogeneity or periodicity [42]. These

attributes are related to the local spatial distribution of grey levels in the image matrix,

resulting from the underlying ultrastructural properties of the tissues affected by the disease.

They are not perceptible by naked eye, but can be captured by calculating texture parameters

that are currently not incorporated in the software used in the clinical practice. A preliminary

T
report showed that some pretreatment texture parameters (based on high order statistics) could

IP
help to predict breast tumor response to NAC [43].

R
4.3 Research issues

SC
There are three main limitations to overcome before recommending MRI as a tool for breast

U
cancer prognosis. First, there is no consensus on the MRI parameters that should be used, for
N
instance very simple parameters, such as the tumor size or necrosis level in triple-negative
A
cancer, or more sophisticated perfusion or texture parameters. Moreover, most published
M

studies did not use multivariate analysis to assess the value of MRI parameters relative to

robust and fully recognized anatomical and histochemical parameters. Finally, the impact of
ED

an MRI prognostic factor on the management of patients with breast cancer scheduled for

NAC has never been demonstrated. Therefore, large-scale homogeneous studies are required
PT

to identify/validate MRI parameters with added value for breast cancer prognosis, and
E

especially to determine the true role of breast MRI in the indication and choice of NAC.
CC

5. Early response assessment

A

5.1 Background

NAC main objectives are tumor downsizing and achieving a pathological complete response

(pCR). About 80% of patients with breast cancer respond to NAC with an overall pCR of

19% in a metaanalysis including more than 11000 patients [44]. pCR is higher in selected

11
groups according to cancer subtypes reaching 31% and 39% in patients with triple-negative

and HER2-positive breast cancers, respectively [44]. Patients who do not respond to NAC

(20%) have poorer disease-free and overall survival rates compared with responders [45, 46].

As a complete course of NAC usually takes several months with different jointly or

sequentially administered chemotherapy regimens, the ability to identify patients who do not

respond early to NAC could influence the outcome in two directions. First, the oncologist

T
could change the type of chemotherapy early in the therapeutic program, based on the

IP
hypothesis that cancer cells resistant to a specific compound might be more sensitive to a

R
different one. However, this assumption overlooks the possibility that some breast cancers

SC
might be intrinsically non-sensitive to chemotherapy.

U
Second, the early detection of non-responders could avoid unnecessary toxicity and costs.
N
5.2 Overview of evidences
A
The clinical impact of assessing the early response and the potential of MRI to produce
M

morphological and functional information on tumors has led many researchers to investigate

the accuracy of different MRI techniques, such as contrast-enhanced MRI, perfusion MRI,
ED

DWI and MR spectroscopy, in monitoring the early response in patients with breast cancer

undergoing NAC.
PT

4.2.1 Volume and contrast enhancement analysis

E

Convincing results were obtained in the ACRIN 6657 trial [46], the imaging component of the
CC

multicenter Investigation of Serial Studies to Predict Your Therapeutic Response with

Imaging and Molecular Analysis (I-SPY) study. The ACRIN 6657 trial was a multicenter
A

study designed to assess MRI ability to measure the primary response to NAC in patients with

breast cancer and to predict recurrence-free survival (RFS). The main parameter tested in this

trial was the functional tumor volume calculated by semi-automated computer analysis based

on the signal enhancement ratio with an enhancement threshold set at 70%, knowing that the

12
measured tumor volume depends on the chosen threshold (Fig. 9). As this volume reflects the

actual size and the microvascular features of the tumor, it could be considered as a functional

measurement. The study found that changes in the functional tumor volume after one cycle of

NAC better predict pCR than the change in the tumor diameter, also measured by MRI (Fig.

10). More recently, these authors showed that in the same population, the functional tumor

volume after one cycle of NAC could predict RFS, even more accurately than pCR [47].

T
Moreover, the absolute functional tumor volume value is a stronger RFS predictor than the

IP
change from baseline after one cycle [47].

R
4.2.2 Advanced perfusion parameters

SC
Data concerning the interest of using the inflow and outflow transfer constants (Ktrans and Kep)

U
to evaluate the early response to NAC and predict the patient outcome are contradictory.
N
Some studies have shown that the percentage change in Kep is an independent predictor of
A
pCR, with high specificity but not enough sensitivity for the clinical practice [48]. Moreover,
M

high Ktrans and Kep are predictors of overall survival [45] in early NAC evaluation. However,

several other studies did not report similar clear-cut results and suggested additional research
ED

[35, 49]. A more recent study [50] found that the perfusion pharmacokinetic parameters did

not predict the early pathologic response to NAC after the first chemotherapy cycle,
PT

differently from the functional tumor volume changes. These contradictory results and the fact
E

that the measurement of these pharmacokinetic parameters is labor-intensive explain why they
CC

are not assessed in the clinical practice, and they are recommended only in translational

cancer research (http://dctd.cancer.gov/ProgramPages/cip/clinical_trials_imaging.htm). The

A

kinetic analysis of the tumor pixel distribution in the initial and delayed phase of contrast

enhancement is easier to perform. It shows a modification of the kinetics when the treatment

is efficient (Fig. 11), but the added value of this finding must be assessed relative to the tumor

volume changes.

13
 4.2.3 Motion of water molecules measured by diffusion-weighted imaging

Several data indicate that an increase in the mean ADC after one NAC cycle is an early

predictor of response to therapy. A meta-analysis that included six studies [51] reported a

DWI sensitivity of 93% (95% CI: 82%-97%) and a specificity of 82% (95% CI: 70%-90%)

for the early prediction of the pathological response in patients undergoing NAC. The early

increase in ADC in NAC responders could be the explained by successful apoptosis, leading

T
to loss of cell membrane integrity, which in turn increases free water diffusion [52]. Although

IP
it has been shown that ADC increase occurs earlier than tumor size reduction [53, 54] (Fig.

R
12), no data are available concerning DWI additional benefits in the early prediction of patient

SC
outcome, particularly in comparison with the functional tumor volume.

4.2.4 Choline concentration measurement

U
N
MR spectroscopy can be used to detect elevated levels of total choline-containing metabolites
A
(tCho) in breast cancer, as a result of increased phospholipid metabolism and cell membrane
M

proliferation. High tCho level is a marker of cell membrane proliferation, therefore it could be

used to monitor the very early tumor response to NAC, but findings are inconsistent. An early
ED

pilot study using quantitative 1H-MR spectroscopy at 4 Tesla found that the change in tCho

level in the tumor within 24h after the first treatment is significantly different between
PT

responders and non-responders [55]. However, this result was not confirmed [56], and in the
E

clinical practice, some technological issues must be overcome before using tCho as a
CC

quantitative imaging biomarker for early NAC response prediction [52]. These issues include:

low in vivo tCho concentration, high variability in breast fat content, non-uniform magnetic
A

field, difficulties in metabolite quantification, and artifacts due to the placement of marker

clips before NAC that can bias tCho measurements. In addition, fewer tumors are available

for tCho quantification when the lesion shrinks after NAC (Fig. 13).

14
5.3 Research issues

The aim of the ACRIN 6657 trial is to investigate the usefulness of 1H MR spectroscopy

measurement of tCho early in the course of treatment. The ACRIN 6698 study is designed to

determine how an early change in tumor ADC is predictive of pCR, and to investigate the

predictive effectiveness of changes in tumor ADC value compared with changes in tumor

functional volume. These studies will identify the additional benefits of more advanced

T
quantitative methods, such as DWI and 1H MR spectroscopy, for the early prediction of the

IP
response to NAC. Furthermore, changes in the background parenchymal enhancement are

R
also an interesting field of research. For instance, a high stromal enhancement ratio may

SC
reflect high microvessel density that favors chemotherapy delivery to the tumor [57].

U
In addition, it should be determined whether metabolic pathway imaging should replace MRI
N
for predicting the response to NAC, because metabolic changes may occur early in the course
A
of therapy, before tumor size reduction. Indeed, it has been shown that positron emission
M

tomography with 2-deoxy-2[fluorine-18]fluoro-D-glucose associated with computer

tomography (FDG PET/CT) is more accurate than DCE-MRI as an interim modality for pCR
ED

prediction [58].

Lastly, it is essential to determine the clinical value of therapeutic strategies (e.g., earlier
PT

surgery) or of NAC switching on the basis of MRI data. However, this will require
E

randomized controlled trials, as clearly stated by the EUSOMA working group [1]. This is
CC

even more important than identifying the most accurate parameter(s) for early tumor response.
A

6. Evaluation after neo-adjuvant chemotherapy

6.1 Background

The pathologic response after NAC in patients with breast cancer is a key factor for assessing

the prognosis in terms of overall, disease-free and recurrence-free survival [58]. Yet, the

15
strategies used to evaluate this parameter are considerably different [59]. The standard

definition of pCR includes the absence of invasive cancer on pathological examination,

irrespective of the presence of residual ductal carcinoma in situ (DCIS) that does not

adversely affect the patient outcome [59]. However, in some studies, pCR was defined as the

absence of any residual cancer that should be included in the surgical excision. Clearly, the

standard definition of pCR increases the sensitivity and decreases the specificity of MRI for

T
residual disease detection. Other definitions of pCR exist. For instance, near-pCR is used

IP
when only small clusters of microscopic invasive cells are still present. The estimated residual

R
cancer burden (RCB), a pathologic tumor response grading index, is also used. RCB is a

SC
continuous histopathologic variable that reflects the size and cellularity of the residual tumor

U
disease in the breast and/or axillary lymph nodes [60]. The use of different criteria to evaluate
N
the pathologic response could be explained by its variability in function of the breast cancer
A
immunohistochemical profile. In patients with low grade, estrogen receptor- or progesterone
M

receptor-positive, or HER2-negative breast cancer, the standard pCR rate is very low [61], and

this criterion is not a suitable surrogate endpoint. Conversely, the pCR rate in triple negative
ED

tumors is higher, but the paradoxical characteristic of such tumors (excellent

chemotherapeutic response despite poor survival) must be taken into account when using pCR
PT

as a prognostic criterion. Differences in the criteria used for defining pCR and in tumor
E

biomarkers could explain the broad range of pCR rates (2.6% to 54.9% with a median of
CC

16%), as highlighted in a meta-analysis to assess MRI performance for detecting residual

breast cancer after NAC (2050 patients from 44 studies) [62].

A

In the presence of residual tumor, estimation of its size may help to guide the choice between

mastectomy and breast-conserving surgery, and to determine the resection volumes to ensure

negative surgical margins.

16
In conclusion, the two main advantages of post-NAC MRI are: prediction of the presence of

residual disease, and accurate determination of the extent of the surgical resection needed to

completely excise residual tumors.

In patients presenting low grade hormone-receptor-positive breast cancers, low pCR rates

after NAC and the development of highly effective aromatase inhibitors have resulted in a

wider use of pre-operative endocrine therapy [63], although endocrine therapy remains well

T
below chemotherapy in this setting.

IP
6.2 Overview of evidences

R
There are substantial converging evidences confirming the superiority of MRI for assessing

SC
the response to NAC compared to conventional breast imaging. Therefore, the EUSOMA

U
study group recommended breast MRI for evaluating the responses to NAC [1]. Moreover,
N
MRI interpretation should be based on the concomitant evaluation of the pre- and post-NAC
A
MRI data [1]. The usual assessment criterion for the diagnosis of complete response by MRI
M

is the absence of enhancement or a contrast enhancement equal to or less than that of normal

tissue [62].
ED

Published studies have demonstrated MRI high sensitivity in detecting residual disease. A

retrospective multicenter analysis on 746 women from eight NCI-designated MRI

PT

comprehensive cancer centers [64] reported an MRI sensitivity of 83% for residual disease
E

detection (including invasive disease and DCIS). However, the negative predictive value was
CC

only 47% because pCR prevalence was 24% (179/746). Indeed, in more than half of the

patients without any residual enhancement at MRI, residual tumors were detected at the
A

pathological assessment, indicating that the disappearance of any abnormal enhancement is

not sufficient to cancel a surgical procedure. MRI specificity ranges from 47% to 83% [59,

64], depending on the tumor immuno-histochemical profile and the MRI criteria used for

residual tumor diagnosis.

17
Besides the presence or absence of a residual abnormal enhancement, it was recently shown

that the ratio between the mean post-treatment and pre-treatment ADC values is higher in

patients with pCR [65].

As MRI can accurately detect and measure the size of residual tumors, it is useful for guiding

surgeons when planning the resection extent. A meta-analysis comparing MRI and pathologic

tumor measurement showed that MRI tends to slightly overestimate the residual tumor size,

T
with a difference of more than 2 cm in about 15% of tumors and 4 cm in about 5% of tumors

IP
between pathology and MRI measurements [58]. Such overestimation could be due to reactive

R
inflammation or the surrounding sclerosis and necrosis. It may lead the surgeon to perform

SC
wider resections than necessary, and this could results in poorer cosmetic results or

U
unnecessary mastectomies. On the other hand, underestimation of the residual disease might
N
result in positive resection margins that necessitate re-intervention, and in not appropriate
A
breast-conserving approaches. For instance, in a monocentric study on 208 patients, post NAC
M

MRI underestimated the residual tumor size by more than 20 mm in 35 patients (17%),

potentially leading to incorrect breast-conserving surgery indications in 27 patients (13%)

ED

[66]. Several parameters can influence the accuracy of MRI evaluation after NAC: the tumor

MRI pattern, the MRI tumor response appearance, the tumor histology and molecular
PT

characteristics, and the chemotherapy regimen.

E

- Tumor MRI pattern: by differentiating between tumors with unifocal mass

CC

enhancement and with diffuse non-mass enhancement, it was demonstrated that the tumor

response and MRI accuracy in evaluating this response vary according to the tumor MRI
A

pattern. Compared with tumors showing non-mass enhancement, well-defined tumor masses

are more likely to show a NAC response that allows breast-conserving therapy (Fig. 14), with

better agreement between the post-NAC MRI evaluation of the residual tumor and pathology

[67, 68].

18
- The MRI tumor response appearance: this partly depends on the MRI tumor pattern.

Indeed, well-defined mass lesions often shrink in a concentric pattern, whereas lesions with

non-mass enhancement decrease according to a scattered pattern with small foci that can

extend throughout the original area of involvement [67]. However, in some cases mass lesions

may break up into multiple foci (Fig. 15). MRI is accurate for residual tumor evaluation after

concentric shrinkage. Conversely, it is not accurate when lesions become fragmented into

T
small pieces, and residual disease is present as scattered cells or cell clusters [52, 69, 70].

IP
- Tumor histology: MRI accuracy is strongly reduced in patients with widespread DCIS

R
(Fig. 16).

SC
- Cancer molecular subtype: MRI is effective for monitoring the post-NAC response in

U
triple negative or HER2-positive breast cancer, but is substantially weaker in estrogen
N
receptor positive/HER2-negative cancers [71, 72]. Tumors that respond well to chemotherapy
A
tend to have higher pCR rates, with a lower likelihood of a scattered minimal residual disease
M

pattern. - Chemotherapy regimen: it has been shown that MRI sensitivity for pCR prediction

and for residual disease quantification is lower in patients treated with taxane-containing than
ED

with non taxane-containing NAC [73]. Taxane non-specific anti-vascular mode of action

explains the tumor response overestimation observed in two-thirds of patients [73] (Fig. 17).
PT

Similarly, the false-negative rate in residual disease evaluation is higher in patients who
E

received a chemotherapy regimen containing anti-angiogenic agents [74].

CC

Consequently, in the interpretation of post-NAC MRI findings, it is essential to know the

molecular cancer subtype and the chemotherapy regimen. Moreover, particular caution is
A

recommended in patients with HER2-negative and hormone receptor-positive breast cancer

and in patients treated with taxane- or bevacizumab-containing NAC.

In patients treated by neoadjuvant endocrine therapy, MRI is currently used in the evaluation

of tumor response [63, 75]. The same pattern of MRI response than after NAC is described

19
with concentric or scattered responses but little is known about the accuracy of MRI in

predicting pCR.

6.3 Research issues

The relation between MRI accuracy in assessing the residual disease after NAC and the MRI

tumor pattern, MRI tumor response appearance and cancer molecular subtype is well

established. However, it is not clear whether these parameters are confounding factors.

T
Multivariate analyses have shown that the molecular subtype and systemic regimens do not

IP
significantly influence MRI accuracy in predicting the pathologic response [76]. On the other

R
hand, the MRI tumor pattern (mass and non-mass enhancement) could have a more profound

SC
impact on MRI accuracy than the molecular subtype [70]. It would be important to identify

U
the major factors that influence MRI accuracy. It is now recognized that surgery must be
N
performed after NAC and that mastectomy should be recommended in patients with very
A
extensive pre-NAC disease that occupies nearly the whole breast, regardless of the MRI
M

tumor response appearance [48]. It would be interesting to evaluate a less aggressive approach

in women with a mass pattern at MRI presentation. This would require improving MRI
ED

sensitivity and negative predictive value for residual disease diagnosis, including in situ

residual disease. DWI added value for pCR prediction [65] must be confirmed. Moreover, the
PT

use of ultra-high field MRI, which is technically feasible for NAC monitoring [77], should be
E

assessed with the aim of avoiding extensive breast surgery in the absence of residual
CC

enhancement.
A

7. Conclusion and perspectives

MRI has a very valuable role in the diagnosis and management of patients with breast cancers

scheduled for NAC. Conventional MRI is routinely used together with morphologic and

kinetic enhancement analysis. EUSOMA guidelines recommend two MRIs in patients eligible

20
for NAC in order to allow breast conservative therapy: a pretreatment breast MRI at the

condition that performing MRI does not significantly postpone NAC initiation; and a post-

NAC MRI which should preferably be performed two weeks after the last NAC cycle and

within 2 weeks before surgery. By contrast, MRI does not have a role in the assessment of

patients with inoperable breast cancer at presentation.

Functional MRI is currently performed only in the framework of research protocols and DWI

T
is very promising to complement MRI at various diagnosis and management stages (breast

IP
and lymph node staging, post-NAC evaluation). Other MRI tools, including perfusion

R
analysis, spectroscopy and textural analysis, have potential benefits for more targeted

SC
applications (Table 1).

U
FDG PET imaging, a metabolic functional imaging modality, shows changes in tumor
N
metabolism after NAC, with an early decrease in standardized uptake value (SUV). It could
A
potentially be used as an early in vivo test of chemosensitivity, providing that SUV threshold
M

values for metabolic response definition have been defined. By contrast, a recent metaanalysis

[78] has shown that FDG PET was less suitable than MRI for predicting pathologic response
ED

after NAC. The major limitation of FDG PET/CT is its inability to reliably detect sub-

centimetric breast lesions because of its lower spatial resolution resulting in high false-
PT

negative rate [79]. However PET/MRI, a new hybrid imaging modality, could overcome these
E

limitations and combine the strengths of MRI and glucose metabolism imaging.
CC

Conflict of interest
A

The authors declare that they have no conflict of interest.

Role of the funding source

No funding source.

21
References

[1] Sardanelli F, Boetes C, Borisch B et al., Magnetic resonance imaging of the breast:

recommendations from the EUSOMA working group. Eur J Cancer 2010;46:1296-316.

[2] Wasif N, Garreau J, Terando A, Kirsch D, Mund DF, Giuliano AE. MRI versus

ultrasonography and mammography for preoperative assessment of breast cancer. Am

Surg 2009;75:970-5.

T
[3] Gruber IV, Rueckert M, Kagan KO et al., Measurement of tumour size with

IP
mammography, sonography and magnetic resonance imaging as compared to histological

R
tumour size in primary breast cancer. BMC Cancer 2013;13:328.

SC
[4] Mennella S, Garlaschi A, Paparo F et al., Magnetic resonance imaging of breast cancer:

U
factors affecting the accuracy of preoperative lesion sizing. Acta Radiol 2015;56:260-8.
N
[5] Haraldsdottir KH, Jonsson Thorn, Halldorsdottir AB, Tranberg KG, Asgeirsson KS.
A
Tumor Size of Invasive Breast Cancer on Magnetic Resonance Imaging and Conventional
M

Imaging (Mammogram/Ultrasound): Comparison with Pathological Size and Clinical

Implications. Scand J Surg 2017;106:68-73.

ED

[6] Berg WA, Gutierrez L, NessAiver MS, et al. Diagnostic accuracy of mammography,

clinical examination, US, and MR imaging in preoperative assessment of breast cancer.

PT

Radiology 2004;233:830-49.
E

[7] Hata T, Takahashi H, Watanabe K, et al. Magnetic resonance imaging for preoperative
CC

evaluation of breast cancer: a comparative study with mammography and

ultrasonography. J Am Coll Surg 2004;198:190-7.

A

[8] Schnall MD, Blume J, Bluemke DA, et al. MRI detection of distinct incidental cancer in

women with primary breast cancer studied in IBMC 6883. J Surg Oncol 2005;92:32-8.

22
[9] Houssami N, Ciatto S, Macaskill P, et al. Accuracy and surgical impact of magnetic

resonance imaging in breast cancer staging: systematic review and meta-analysis in

detection of multifocal and multicentric cancer. J Clin Oncol 2008;26:3248-58.

[10] Brennan ME, Houssami N, Lord S, et al. Magnetic resonance imaging screening of the

contralateral breast in women with newly diagnosed breast cancer: systematic review and

meta-analysis of incremental cancer detection and impact on surgical management. J Clin

T
Oncol 2009;27:5640-9.

IP
[11] Plana MN, Carreira C, Muriel A, et al. Magnetic resonance imaging in the preoperative

R
assessment of patients with primary breast cancer: systematic review of diagnostic

SC
accuracy and meta-analysis. Eur Radiol 2012;22:26-38.

U
[12] Peters NH, Borel Rinkes IH, Zuithoff NP, Mali WP, Moons KG, Peeters PH.
N
Metaanalysis of MR imaging in the diagnosis of breast lesions. Radiology 2008;246:116-
A
24.
M

[13] Kuehn T, Bauerfeind I, Fehm T, et al. Sentinel-lymph-node biopsy in patients with breast

cancer before and after neoadjuvant chemotherapy (SENTINA): a prospective,

ED

multicentre cohort study. Lancet Oncol 2013;14:609-18.

[14] Curigliano G, Criscitiello C, Esposito A, et al. Best management of locally advanced

PT

inoperable breast cancer. Eur J Cancer 2013;11:289-290.

E

[15]Bazan JG, White J. Imaging of the axilla before preoperative chemotherapy: implications
CC

for postmastectomy radiation. Cancer 2015;121:1187-94.

[16] Boughey JC, Suman VJ, Mittendorf EA, et al. Sentinel lymph node surgery after
A

neoadjuvant chemotherapy in patients with node-positive breast cancer: the ACOSOG

Z1071 (Alliance) Clinical Trial. JAMA 2013;310:1455-1461.

[17]Berg JW. The significance of axillary node levels in the study of breast carcinoma.

Cancer 1955;8:776-8.

23
[18] Rahbar H, Partridge SC, Javid SH, Lehman CD. Imaging axillary lymph nodes in

patients with newly diagnosed breast cancer. Curr Probl Diagn Radiol 2012;41:149-58.

[19] Veronesi U, Arnone P, Veronesi P, et al. The value of radiotherapy on metastatic internal

mammary nodes in breast cancer. Results on a large series. Ann Oncol 2008;19:1553-60.

[20] Veronesi U, Arnone P, Veronesi P, Galimberti V, Luini A, Rotmensz N, et al. The value

of radiotherapy on metastatic internal mammary nodes in breast cancer. Results on a large

T
series. Ann Oncol. 2008;19:1553-60.

IP
[21] Chen RC, Lin NU, Golshan M, Harris JR, Bellon JR. Internal mammary nodes in breast

R
cancer: diagnosis and implications for patient management -- a systematic review. J Clin

SC
Oncol 2008;26:4981-9.

U
[22] Jochelson MS, Lebron L, Jacobs SS, et al. Detection of Internal Mammary Adenopathy
N
in Patients With Breast Cancer by PET/CT and MRI. AJR Am J Roentgenol
A
2015;205:899904.
M

[23] Mortellaro VE, Marshall J, Singer L, et al. Magnetic resonance imaging for axillary

staging in patients with breast cancer. J Magn Reson Imaging 2009;30:309-12.

ED

[24] Javid S, Segara D, Lotfi P, Raza S, Golshan M. Can breast MRI predict axillary lymph

node metastasis in women undergoing neoadjuvant chemotherapy. Ann Surg Oncol

PT

2010;17:1841-6.
E

[25] de Felice C, Cipolla V, Stagnitti A, et al. Diagnostic accuracy of 1.5 Tesla breast
CC

magnetic resonance imaging in the pre-operative assessment of axillary lymph nodes. Eur

J Gynaecol Oncol 2015;36:447-51.

A

[26] You S, Kang DK, Jung YS et al. Evaluation of lymph node status after neoadjuvant

chemotherapy in breast cancer patients: comparison of diagnostic performance of

ultrasound, MRI and 18F-FDG PET/CT. Br J Radiol 2015;88:20150143.

24
[27] Folwer AM, Mankoff DA, Joe B N. Imaging Neoadjuvant Therapy Response in Breast

Cancer. Radiology 2017;285:358-75.

[28] Yamashita T, Takahara T, Kwee TC, et al. Diffusion magnetic resonance imaging with

gadofosveset trisodium as a negative contrast agent for lymph node metastases

assessment. Jpn J Radiol 2011;29:25-32.

[29] Kwee TC, Takahara T, Luijten PR, Nievelstein RA. ADC measurements of lymph nodes:

T
inter- and intra-observer reproducibility study and an overview of the literature. Eur J

IP
Radiol 2010;75:215-20.

R
[30] He N, Xie C, Wei W, et al. A new, preoperative, MRI-based scoring system for

SC
diagnosing malignant axillary lymph nodes in women evaluated for breast cancer. Eur J

Radiol 2012;81:2602-12.
U
N
[31] Memarsadeghi M, Riedl CC, Kaneider A, et al. Axillary lymph node metastases in
A
patients with breast carcinomas: assessment with nonenhanced versus uspio-enhanced
M

MR imaging. Radiology 2006;241:367-77.

[32] Smid M, Wang Y, Zhang Y, et al. Subtypes of breast cancer show preferential site of
ED

relapse. Cancer Res 2008;68:3108-14.

[33] Fernandez-Morales LA, Segui MA, Andreu X, et al. Analysis of the pathologic response
PT

to primary chemotherapy in patients with locally advanced breast cancer grouped

E

according to estrogen receptor, progesterone receptor, and HER2 status. Clin Breast
CC

Cancer 2007;7:55964.

[34] Uematsu T, Kasami M, Yuen S. Neoadjuvant chemotherapy for breast cancer:

A

correlation between the baseline MR imaging findings and responses to therapy. Eur

Radiol 2010;20:2315-22.

25
[35] Prevos R, Smidt ML, Tjan-Heijnen VC, et al. Pre-treatment differences and early

response monitoring of neoadjuvant chemotherapy in breast cancer patients using

magnetic resonance imaging: a systematic review. Eur Radiol 2012;22:2607-16.

[36] Kawashima H, Inokuchi M, Furukawa H, Kitamura S. Triple-negative breast cancer: are

the imaging findings different between responders and nonresponders to neoadjuvant

chemotherapy? Acad Radiol 2011;18:963-9.

T
[37] Pickles MD, Lowry M, Manton DJ, Turnbull LW. Prognostic value of DCE-MRI in

IP
breast cancer patients undergoing neoadjuvant chemotherapy: a comparison with

R
traditional survival indicators. Eur Radiol 2015;25:1097-106.

SC
[38] Heldahl MG, Bathen TF, Rydland J, et al. Prognostic value of pretreatment dynamic

U
contrast-enhanced MR imaging in breast cancer patients receiving neoadjuvant
N
chemotherapy: overall survival predicted from combined time course and volume
A
analysis. Acta Radiol 2010;51:604-12.
M

[39] Park SH, Moon WK, Cho N, et al. Diffusion-weighted MR imaging: pretreatment

prediction of response to neoadjuvant chemotherapy in patients with breast cancer.

ED

Radiology 2010;257:56-63.

[40] Iacconi C, Giannelli M, Marini C, et al. The role of mean diffusivity (MD) as a
PT

predictive index of the response to chemotherapy in locally advanced breast cancer: a

E

preliminary study. Eur Radiol 2010;20:303-8.

CC

[41] Nilsen L, Fangberget A, Geier O, Olsen DR, Seierstad T. Diffusion-weighted magnetic

resonance imaging for pretreatment prediction and monitoring of treatment response of

A

patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy. Acta

Oncol 2010;49:354-60.

[42] Michoux N, Van den Broeck S, Lacoste L, et al. Texture analysis on MR images helps

predicting non-response to NAC in breast cancer. BMC Cancer 2015;15:574.

26
[43] Ahmed A, Gibbs P, Pickles M, Turnbull L. Texture analysis in assessment and prediction

of chemotherapy response in breast cancer. J Magn Reson Imaging 2013;38:89-101.

[44] Houssami N, Macaskill P, Minckwitz GV, Marinovich ML, Mamounas E. Meta-analysis

of the association of the breast cancer subtype and pathologic complete response to

neoadjuvant chemotherapy. Eur J Cancer 2012;48:3342-54.

[45] Li SP, Makris A, Beresford MJ, et al. Use of dynamic contrast-enhanced MR imaging to

T
predict survival in patients with primary breast cancer undergoing neoadjuvant

IP
chemotherapy. Radiology 2011;260:68-78.

R
[46] Hylton NM, Blume JD, Bernreuter WK, et al. Locally advanced breast cancer: MR

SC
imaging for prediction of response to neoadjuvant chemotherapy--results from ACRIN

6657/I-SPY TRIAL. Radiology 2012;263:663-72.

U
N
[47] Hylton NM, Gatsonis CA, Rosen MA, et al. Neoadjuvant Chemotherapy for Breast
A
Cancer: Functional Tumor Volume by MR Imaging Predicts Recurrence-free Survival-
M

Results from the ACRIN 6657/CALGB 150007 I-SPY 1 TRIAL. Radiology

2016;279:44-55.
ED

[48] Tateishi U, Miyake M, Nagaoka T, et al. Neoadjuvant chemotherapy in breast cancer:

prediction of pathologic response with PET/CT and dynamic contrast-enhanced MR

PT

imaging-prospective assessment. Radiology 2012;263:53-63.

E

[49] Marinovich ML, Sardanelli F, Ciatto S, et al. Early prediction of pathologic response to
CC

neoadjuvant therapy in breast cancer: systematic review of the accuracy of MRI. Breast

2012;21:669-77.
A

[50] Cho N, Im SA, Park IA, et al. Breast cancer: early prediction of response to neoadjuvant

chemotherapy using parametric response maps for MR imaging. Radiology

2014;272:385-96.

27
[51]Wu LM, Hu JN, Gu HY, Hua J, Chen J, Xu JR. Can diffusion-weighted MR imaging and

contrast-enhanced MR imaging precisely evaluate and predict pathological response to

neoadjuvant chemotherapy in patients with breast cancer? Breast Cancer Res Treat

2012;135:17-28.

[52] Chen JH, Su MY. Clinical application of magnetic resonance imaging in management of

breast cancer patients receiving neoadjuvant chemotherapy. Biomed Res Int

T
2013;2013:348167.

IP
[53] Pickles MD, Gibbs P, Lowry M, Turnbull LW. Diffusion changes precede size reduction

R
in neoadjuvant treatment of breast cancer. Magn Reson Imaging 2006;24:843-7.

SC
[54] Sharma U, Danishad KK, Seenu V, Jagannathan NR. Longitudinal study of the

U
assessment by MRI and diffusion-weighted imaging of tumor response in patients with
N
locally advanced breast cancer undergoing neoadjuvant chemotherapy. NMR Biomed
A
2009;22:10413.
M

[55] Meisamy S, Bolan PJ, Baker EH, et al. Neoadjuvant chemotherapy of locally advanced

breast cancer: predicting response with in vivo (1)H MR spectroscopy--a pilot study at 4
ED

T. Radiology 2004;233:424-31.

[56] Baek HM, Chen JH, Nie K, et al. Predicting pathologic response to neoadjuvant
PT

chemotherapy in breast cancer by using MR imaging and quantitative 1H MR

E

spectroscopy. Radiology 2009;251:653-62.

CC

[57] Hattangadi J, Park C, Rembert J, et al. Breast stromal enhancement on MRI is associated

with response to neoadjuvant chemotherapy. AJR Am J Roentgenol 2008;190:1630-6.

A

[58] Kong X, Moran MS, Zhang N, Haffty B, Yang Q. Meta-analysis confirms achieving

pathological complete response after neoadjuvant chemotherapy predicts favourable

prognosis for breast cancer patients. Eur J Cancer 2011;47:2084-90.

28
[59] Mazouni C, Peintinger F, Wan-Kau S, et al. Residual ductal carcinoma in situ in patients

with complete eradication of invasive breast cancer after neoadjuvant chemotherapy does

not adversely affect patient outcome. J Clin Oncol 2007;25:2650-5.

[60] Symmans WF, Peintinger F, Hatzis C, et al. Measurement of residual breast cancer

burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol 2007;25:4414-

22.

T
[61] Guarneri V, Broglio K, Kau SW, et al. Prognostic value of pathologic complete response

IP
after primary chemotherapy in relation to hormone receptor status and other factors. J

R
Clin Oncol 2006;24:1037-44.

SC
[62] Marinovich ML, Houssami N, Macaskill P, et al. Meta-analysis of magnetic resonance

U
imaging in detecting residual breast cancer after neoadjuvant therapy. J Natl Cancer Inst
N
2013;105:321-33.
A
[63] Van Dam PA, Van Dam VC, Altintas S, Papadimitriou K, Rolfo C, Trinh XB.
M

Neoadjuvant endocrine treatment in early breast cancer: An overlooked alternative? Eur J

Surg Oncol 2016;42:333-42.

ED

[64] De Los Santos JF, Cantor A, Amos KD, et al. Magnetic resonance imaging as a predictor

of pathologic response in patients treated with neoadjuvant systemic treatment for

PT

operable breast cancer. Translational Breast Cancer Research Consortium trial 017.
E

Cancer 2013;119:1776-83.
CC

[65]Santamaria G, Bargallo X, Fernandez PL, Farrus B, Caparros X, Velasco M.

Neoadjuvant Systemic Therapy in Breast Cancer: Association of Contrast-enhanced MR

A

Imaging Findings, Diffusion-weighted Imaging Findings, and Tumor Subtype with

Tumor Response. Radiology 2017;283:663-672.

[66] Straver ME, Loo CE, Rutgers EJ, et al. MRI-model to guide the surgical treatment in

breast cancer patients after neoadjuvant chemotherapy. Ann Surg 2010;251:701-7.

29
[67] Price ER, Wong J, Mukhtar R, Hylton N, Esserman LJ. How to use magnetic resonance

imaging following neoadjuvant chemotherapy in locally advanced breast cancer. World J

Clin Cases 2015;3:607-13.

[68] Mukhtar RA, Wong J, Mukhtar R, Hylton N, Esserman LJ. Clinically meaningful tumor

reduction rates vary by prechemotherapy MRI phenotype and tumor subtype in the I-SPY

1 TRIAL (CALGB 150007/150012; ACRIN 6657). Ann Surg Oncol 2013;20:3823-30.

T
[69] Partridge SC, Gibbs JE, Lu Y, Esserman LJ, Sudilovsky D, Hylton NM. Accuracy of MR

IP
imaging for revealing residual breast cancer in patients who have undergone neoadjuvant

R
chemotherapy. AJR Am J Roentgenol 2002;179:1193-9.

SC
[70] Ko ES, Han BK, Kim RB, et al. Analysis of factors that influence the accuracy of

U
magnetic resonance imaging for predicting response after neoadjuvant chemotherapy in
N
locally advanced breast cancer. Ann Surg Oncol 2013;20:2562-8.
A
[71] Chen JH, Mehta RS, Carpenter PM, Nalcioglu O, Su MY. Magnetic resonance imaging
M

in predicting pathological response of triple negative breast cancer following neoadjuvant

chemotherapy. J Clin Oncol 2007;25:5667-9.

ED

[72] Loo CE, Straver ME, Rodenhuis S, et al. Magnetic resonance imaging response

monitoring of breast cancer during neoadjuvant chemotherapy: relevance of breast cancer

PT

subtype. J Clin Oncol 2011;29:660-6.

E

[73] Schrading S, Kuhl CK. Breast Cancer: Influence of Taxanes on Response Assessment
CC

with Dynamic Contrast-enhanced MR Imaging. Radiology 2015;277:687-96.

[74] Bahri S, Chen JH, Mehta RS, et al. Residual breast cancer diagnosed by MRI in patients
A

receiving neoadjuvant chemotherapy with and without bevacizumab. Ann Surg Oncol

2009;16:1619-28.

30
[75] Fontein DB, Charehbili A, Nortier JW, et al. Efficacy of six months neoadjuvant

endocrine therapy in postmenopausal, hormone receptor-positive breast cancer patients--a

phase II trial. Eur J Cancer 2014;50:2190-200.

[76] De Los Santos J, Bernreuter W, Keene K, et al. Accuracy of breast magnetic resonance

imaging in predicting pathologic response in patients treated with neoadjuvant

chemotherapy. Clin Breast Cancer 2011;11:312-9.

T
[77] Korteweg MA, Veldhuis WB, Visser F, et al. Feasibility of 7 Tesla breast magnetic

IP
resonance imaging determination of intrinsic sensitivity and high-resolution magnetic

R
resonance imaging, diffusion-weighted imaging, and (1)H-magnetic resonance

SC
spectroscopy of breast cancer patients receiving neoadjuvant therapy. Invest Radiol

2011;46:370-6.
U
N
[78] Li H, Yao L, Jin P, et al. MRI and PET/CT for evaluation of the pathological response to
A
neoadjuvant chemotherapy in breast cancer: A systematic review and meta-analysis.
M

Breast 2018;40:106-115.

[79] Rauch GM, Adrada BE , Kuerer HM, Leung JW, Yang WT. Multimodality Imaging for
ED

Evaluating Response to Neoadjuvant Chemotherapy in Breast Cancer. AJR Am J

Roentgenol 2017;208:290-299.
E PT
CC
A

31
Figure legends

Figure 1: 3D breast volume rendered with skin texture: mediolateral oblique view (a), and

sagittal view (b). Olea Sphere 3.0 sp9 and Olea Sphere research version (Olea Medical SAS,

La Ciotat, France) were used as magnetic resonance image processing software tools. The

inferior and posterior yellow structure corresponds to the heart.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

32
Figure 2: Baseline T1-weighted post-gadolinium subtracted (a) and diffusion weighted (DW)

(b) magnetic resonance images in a 39-year-old woman with left bifocal triple-negative breast

cancer. There is a contralateral well-circumscribed mass lesion without any hypersignal on the

DW image (b). Ultrasound-guided biopsy of this right mass revealed the presence of a

fibroadenoma.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

33
Figure 3: Baseline T1-weighted post-gadolinium subtracted magnetic resonance images (a, b)

in a 63-year-old woman with a bilateral luminal A breast cancer. The right tumor (b) was

discovered by MRI data analysis.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

34
Figure 4: Baseline T1-weighted post-gadolinium subtracted (a) and T2-weighted (b) magnetic

resonance images in a 49-year-old woman with left luminal B breast cancer. The tumor

measured more than 10 cm and invaded the pectoralis major muscle.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

35
Figure 5: Baseline T2-weighted (a) and T1-weighted post-gadolinium subtracted (b) magnetic

resonance images in a 46-year-old woman with axillary lymph node-positive left triple

negative breast cancer used to describe the muscle-based categorization of axillary lymph

nodes by Berg. The dashed lines define the inner and outer borders of the pectoralis minor

muscle. The lymph nodes (arrow) located laterally to the outer border of the pectoralis minor

muscle are level 1 nodes, whereas the lymph nodes (arrowhead) located medially to the inner

T
border of the pectoralis minor muscle are level 3 nodes. Note the posterior voluminous

IP
necrotic primary tumor.

R
SC
U
N
A
M
ED
E PT
CC
A

36
Figure 6: Baseline T2-weighted magnetic resonance image showing the difference between

malignant axillary lymph nodes (arrows on the left side) and benign lymph nodes (arrowhead

on the right side) in a 43-year-old woman with left luminal B breast cancer.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

37
Figure 7: Baseline T1-weighted post-gadolinium subtracted (a), T2-weighted (b), and DWI

magnetic resonance images (c) and ADC map (d) in a 39-year-old woman with left, bifocal,

triple-negative breast cancer (same patient as in Figure 2). The main tumor exhibits a

posterior solid portion (arrow) and an anterior necrotic portion. The necrotic portion is not

contrast-enhanced (a), displays a hypersignal on T2-weighted (b) and DW (c) images and

higher ADC (d) compared with the solid part of the tumor. The tumor appearance did not

T
change after NAC.

R IP
SC
U
N
A
M
ED
E PT
CC
A

38
Figure 8: Quantitative analysis of tumor perfusion with a two-compartment model (vascular

compartment and extra-vascular extra-cellular compartment). The pharmacokinetic

parameters Ktrans, Kep and Ve are calculated from the signal intensity time curve and for this,

the changes in signal intensity need to be converted to the contrast agent concentration in the

tissue by taking into account the arterial input function.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

39
Figure 9: Relationship between the functional tumor volume and the enhancement threshold

considered to be significant. Axial magnetic resonance image (a) showing a mass in the left

outer quadrant of the breast region. The functional tumor volumes were computed using a

semi-automatic segmentation tool based on peak enhancement and the Olea Medical software

(Olea Medical, La Ciotat, France). Functional tumor volumes of 4.65 cm3, 3.55 cm3, and 2.19

cm3 were obtained with a 50% (b), 70% (c), and 100% (d) threshold, respectively.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

40
Figure 10: Functional tumor volume changes after one chemotherapy cycle in a 58-year-old

woman with left luminal B breast cancer (same patient as in Figure 9). Axial T1-weighted

post-gadolinium subtracted magnetic resonance images and functional tumor volume at

baseline (a, c) and after one chemotherapy cycle (b, d). 3D volume segmentation analysis

based on peak enhancement, by using a 70% enhancement threshold for considering a pixel as

a functional tumor pixel, indicated that the functional tumor volume decreased from 3.55 cm3

T
at baseline to 2.68 cm3 after one chemotherapy cycle. After NAC, surgical pathology did not

IP
find any evidence of residual cancer.

R
SC
U
N
A
M
ED
E PT
CC
A

41
Figure 11: Kinetic changes after one chemotherapy cycle in a 58-year-old woman with left

luminal B breast cancer (same patient as in Figures 9 and 10) by dynamic contrast-enhanced

MRI. Color-coded map of the maximum enhancement slope at baseline (a) and after one

chemotherapy cycle (c). Panels 11b and 11d shows the distribution of pixels with slow,

medium, or rapid enhancement during the initial phase of the enhancement kinetics (increase),

the distribution of pixels with persistent, plateau or washout enhancement during the delayed

T
phase of enhancement (course), and the enhancement curves (the percentage of enhancement

IP
is on the Y axis and the time in seconds on the X axis). Comparison of these features at

R
baseline (b) and after one chemotherapy cycle (d) showed that the percentage of pixels with

SC
rapid enhancement decreased from 85.9% to 59.4%, the percentage of pixel with washout

U
decreased from 16.6% to 4.6%, and the enhancement curve pattern changed from a washout
N
pattern to a plateau pattern after one chemotherapy cycle.
A
M
ED
E PT
CC
A

42
 A
CC
EPT
ED
M
A
N
U
SC
RIP

43
T
Figure 12: Changes after one chemotherapy cycle in a 58-year-old woman with right

triplenegative breast cancer. Dynamic contrast-enhanced MRI and diffusion-weighted MRI at

baseline (a, c) and after one chemotherapy cycle (b, d). Contrast-enhanced subtracted

T1weighted images show a decrease in tumor volume between baseline (a) and after one

chemotherapy cycle (b), although the greatest diameter remains unchanged. DWI evaluation

(b=1500s/mm2) before (c) and after one cycle (d) (DWI image in the lower panel and ADC

T
map in the upper panel) shows an increase of the ADC value between baseline and after one

IP
cycle of chemotherapy (0.71x10-3 mm2/s and 1.07x10-3 mm2/s, respectively). Surgical

R
pathology analysis of biopsies did not find any evidence of residual cancer.

SC
U
N
A
M
ED
E PT
CC
A

44
Figure 13: MR spectra acquired before (a) and after one chemotherapy cycle (b) in a 58-

yearold woman with right triple-negative breast cancer (same patient as in Figure 12). An

elevated total tCho peak is visible at 3.23 ppm in the baseline water-fat suppressed spectrum

(a). After one chemotherapy cycle, tCho was no longer detectable, but this could be due to the

small size of the residual lesion.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

45
Figure 14: Baseline (a) and post-NAC (b) axial T1-weighted post-gadolinium subtracted

magnetic resonance images in a 35-year-old woman with luminal B left breast cancer. The

baseline image (a) shows a segmental non-mass enhancement spanning up to 7cm in the

upper outer quadrant of the breast. The post-NAC image (b) shows a slight decrease in the

enhancement intensity without any tumor size change. Surgical pathology confirmed the

presence of an invasive cancer in the upper outer quadrant of the breast.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

46
Figure 15: Baseline (a) and post-NAC (b) axial T1-weighted post-gadolinium subtracted

magnetic resonance images in a 63-year-old woman with luminal A left breast cancer (same

patient as in Figure 3). The baseline image (a) shows a 3cm unifocal mass in the upper inner

quadrant of the breast. The post-NAC image (b) shows tumor fragmentation into several

smaller foci. Surgical pathology found residual invasive disease with several foci of less than

1 cm in size.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

47
Figure 16: Baseline (a, b) and post-NAC (c) axial T1-weighted post-gadolinium subtracted

magnetic resonance images in a 63-year-old woman with hormone receptor-positive left

breast cancer. The baseline images (a, b) show a spiculated mass corresponding to an invasive

tumor and contiguous non-mass enhancement corresponding to ductal carcinoma in situ

(DCIS) that extends superiorly for a total of 6 cm. Despite the lack of enhancement on the

post-NAC image (c), surgical pathology revealed multiple residual foci of DCIS.

T
R IP
SC
U
N
A
M
ED
E PT
CC
A

48
Figure 17: Baseline (a, b) and post-NAC (c, d) axial T1-weighted post-gadolinium subtracted

(a, c) and T2-weighted (b, d) magnetic resonance images in a 50-year-old woman with

luminal B retroareolar right breast cancer who received taxane-based chemotherapy. The

baseline images (a, b) show a 2.1cm retroareolar mass (arrow) with thickening of the

areolarnipple complex. The post-NAC images show a small residual enhancement in the

T1weighted image (c) (arrow), whereas the size of the mass is only slightly decreased in the

T
T2weighted image (d) (arrow). Surgical pathology demonstrated a 1.8-cm residual cancer

IP
lesion the size of which was underestimated in T1-weighted post-gadolinium subtracted

R
sequences.

SC
U
N
A
M
ED
E PT
CC
A

49
 A
CC
EPT
ED
M
A
N
U
SC
RIP

50
T
Table 1: Role of the different MR techniques in function of the clinical questions
Local breast Lymph node Prognosis Early Post-NAC
staging staging response response
assessment assessment
Morphologic Clinical Clinical Clinical Clinical Clinical
pattern practice +++ practice +++ practice + practice + practice +++
Kinetic Clinical Work in Clinical

T
enhancement practice ++ progress practice +
pattern

IP
DWI Work in Work in Work in Work in Work in

R
progress progress progress progress progress

SC
Perfusion Work in Work in
progress progress
Spectroscopy Work in
U progress
N
Textural Work in
A
analysis progress
M
ED
E PT
CC
A

51