JOURNAL OF MAGNETIC RESONANCE IMAGING 30:309 312 (2009)
Original Research
Magnetic Resonance Imaging for Axillary Staging
in Patients With Breast Cancer
Vincent E. Mortellaro, MD,1 Julia Marshall, MD,2 Lauren Singer, MD,1
Steven N. Hochwald, MD,1 Myron Chang, PhD,3 Edward M. Copeland, MD,1 and
Stephen R. Grobmyer, MD1*
Purpose: To determine the utility of MRI for assessing ax-
illary lymph node status in patients with breast cancer.
Materials and Methods: A consecutive series of patients
who underwent breast MR before surgical management of
breast cancer with axillary sampling between 2005 and
2007 were identied. MRs were evaluated for the number of
nodes, contrast kinetics, nodal area, and number of nodes
with no fatty hilum. Data were analyzed in the context of
nal breast pathology, sentinel lymph node status, and
axillary nodal status. Correlations were analyzed using
Kendalls -b test. Reported P values are one-sided.
Results: Fifty-six females (median 58 years) were stud-
ied. Sentinel lymph nodes (SLN) were positive in 15/56
patients (27%). All SLN patients (n 15) had completion
axillary dissection. Extent of nodal disease was 1 (n 3), 2
(n 4), 2 (n 8). Presence of any axillary lymph node with
no fatty hilum and the number of nodes with no fatty hilum
on MR signicantly correlated with pathologic node positiv-
ity (P 0.04); while kinetics, node number, and node size
did not correlate.
Conclusion: Breast MR may be useful in the assessment of
axillary nodes in patients with breast cancer.
Key Words: MRI; breast cancer; lymph nodes; hilum; can-
cer staging
J. Magn. Reson. Imaging 2009;30:309 312.
2009 Wiley-Liss, Inc.
1
Division of Surgical Oncology, Department of Surgery, University of
Florida, Gainesville, Florida.
2
Department of Radiology, University of Florida, Gainesville, Florida.
3
Division of Biostatistics, Department of Epidemiology and Health Pol-
icy Research, University of Florida, Gainesville, Florida.
Portions of this work were presented at the 2008 ASCO Breast Cancer
Symposium, Washington, DC.
*Address reprint requests to: S.R.G., University of Florida, Department
of Surgery, Division of Surgical Oncology, 1600 S.W. Archer Road, PO
Box: 100286, Gainesville, FL. 32610.
E-mail: stephen.grobmyer@surgery.u.edu
Received November 4, 2008; Accepted April 2, 2009.
DOI 10.1002/jmri.21802
Published online May 22, 2009 in Wiley InterScience (www.interscience.
wiley.com).
2009 Wiley-Liss, Inc. 309
BREAST MAGNETIC RESONANCE imaging (MR) is
gaining acceptance for screening in high-risk patients,
detection of occult breast cancer, and the evaluation of
response to neo-adjuvant treatment (1 4). Breast MR is
currently also being evaluated for preoperative staging
(5). Magnetic resonance criteria that are associated
with the presence of breast malignancy include size,
contrast kinetics, and morphology (6).
The presence of metastatic disease is the most impor-
tant prognostic element of breast cancer staging, while
axillary nodal metastasis is the single most important
element for loco-regional staging (7,8). Surgical biopsy
remains the gold-standard for the diagnostic evaluation
of axillary lymph node metastasis (9). Traditional imag-
ing paradigms (computed tomography, positron emis-
sion tomography, ultrasound, mammography) and
physical examination have not proven to be accurate in
the identication of axillary nodal metastasis (10,11).
The development of noninvasive diagnostic modalities
for accurate axillary staging would represent an impor-
tant advance in the management of breast cancer pa-
tients.
In this study, we hypothesize that breast MR may be
useful in the identication of axillary nodal metastasis.
There has been very limited study of this application for
breast MR in this setting, and it remains unclear
whether the MR criteria presently used to identify
breast cancer is also applicable in the evaluation of
axillary nodes on MR.
MATERIALS AND METHODS
A consecutive series of patients who underwent breast
MR using a 1.5 Tesla (T) magnet (Sonata; Siemens)
before surgical management of breast cancer between
2005 and 2007 were identied. No patients had re-
ceived neo-adjuvant chemotherapy. Patients were in-
jected through an IV catheter in the antecubital fossa
with gadodiamide (Omniscan; GEHealthcare) at a rate
of 2 mL/s using a power injector. An in vivo 7-channel
breast coil (Siemens, Inc.) was used for imaging. A se-
ries of ve postcontrast scans were obtained sequen-
tially, with a total acquisition time of 6 min. Imaging
sequences obtained included a precontrast axial three-
dimensional (3D) gradient echo with fat suppression
310 Mortellaro et al.

Table 1
MR Characteristics of patients With Negative LN Metastasis
Number of nodes Median 3 (range, 06)
Predominant lling type 3 (65% of patients)
No type 3 lling 35%
No nodes visualized 15%
Maximum node size (area) Median 80 mm2 (range 18
to 504 mm2)

sive ductal carcinoma 44; invasive lobular carci-

noma 4). Eight patients (15%) had ductal carcinoma
in situ. The median size of the primary invasive breast
Figure 1. Lymph nodes with fatty hilum versus no fatty hi-
cancer was 2cm (range, 0.2 4 cm).
lum. Left: Breast MR, T2 STIR Series. Red arrow highlights an
All patients underwent sentinel lymph node biopsy.
example of a lymph node with a fatty hilum present. Right:
Breast MR, T2 STIR Series. Blue arrow highlights an example Fifteen patients were found to have nodal metastasis on
of a lymph node with no fatty hilum present. sentinel lymph node biopsy. All patients with positive
sentinel lymph node biopsies went on to have axillary
nodal dissections. The median number of nodes re-
moved at axillary dissection was 9 (range, 221). The
(TR, 5;TE, 1), 2-mm-thick slices;T2 STIR (TR, 10,000; distribution of axillary metastasis was as follows: 1
TE, 70), 2-mm-thick slices; and T2 STIR coronal images positive axillary node (n 3), 2 positive axillary nodes
of the chest with the patient in the supine position in a (n 4), and 2 positive axillary nodes (n 8).
body coil. Postcontrast scan parameters are the same There is signicant heterogeneity among measured
as for precontrast T1-weighted series. Images were then MR characteristics of patients with and without axillary
viewed on a dedicated workstation (DynaCAD; In vivo). nodal metastasis (Table 1). Four patients with negative
The average eld of view was 36.8 30.0 cm. Postpro- sentinel lymph nodes were found to have loss of fatty
cessing included image subtraction, maximum inten- hilum on MRI.
sity projection, 3D reconstruction, signal-intensity The presence of any axillary node with no fatty hilum
analysis, and computer-assisted diagnosis. and the number of axillary nodes with no fatty hilum
Flow kinetics were evaluated by placing a volume of seen on MR signicantly correlated with pathologic ax-
interest cursor over the lymph node being evaluated. illary nodal positivity (correlation coefcients 0.26
The volume of interest could be adjusted for size in the and 0.37 and P 0.045 and 0.024, respectively; Fig. 2;
coronal, axial, and sagittal planes to ensure the entire Table 2). The number of lymph nodes, ow kinetics, and
node was included. The computer then calculated the size did not signicantly correlate with pathologic nodal
percentage of each curve present within the node. positivity (Table 2).
Flow kinetics then were plotted on a graph as a curve
representing the tissue enhancement over time. The
kinetic curves were originally developed to characterize DISCUSSION
areas of enhancements in the breast. A type 1 curve has The ability to nonoperatively dene pathologically pos-
continually increasing enhancement over time, a type 2 itive or negative axillary nodes in patients presenting
curve has rapid wash in and then plateaus, and a type
3 curve has rapid wash in and wash out.
MRs then were reviewed by a breast radiologist in a
blinded manner with focus on lymph nodes (number of
nodes, contrast ow kinetics, nodal area, and number
of nodes with no fatty hilum). Nodal area, number of
nodes, and assessment of fatty hilum was performed on
the T2 STIR coronal images. Examples of lymph nodes
with fatty hilum versus no fatty hilum are noted in
Figure 1. Data were analyzed in the context of nal
breast pathology, sentinel lymph node status, and
nodal status. Correlations were analyzed using Ken-
dalls -b test. Reported P values are one-sided. All data
were analyzed using SAS version 9.1.3 software.
Approval for this study was obtained from the facili-
ties Institutional Review Board.

RESULTS
All patients in this study were female (n 56). The Figure 2. Association of no fatty hilum with nodal positivity.
median age was 58 years (range, 30 to 82 years). Forty- Statistically signicant association between no fatty hilum and
eight patients (85%) had invasive breast cancer (inva- nodal positivity.
Axillary Staging of Breast Cancer With MRI 311

Table 2
Correlation Coefcients* between MR Characteristics and Nodal Positivity with P Values** in Parentheses
One or more positive Two or more positive
nodes (yes vs. no) nodes (yes vs. no)
No. of LN 0.11 (0.17) 0.19 (0.060)
Type 1 kinetics 0.029 (0.42) 0.059 (0.33)
Type 2 kinetics 0.042 (0.35) 0.079 (0.22)
Type 3 kinetics 0.035 (0.62) 0.069 (0.72)
Max. area 0.045 (0.36) 0.083 (0.26)
Longest length 0.074 (0.27) 0.10 (0.21)
One or more nodes with no fatty hilum (yes vs. no) 0.21 (0.075) 0.28 (0.035)
No. nodes with no fatty hilum 0.27 (0.037) 0.33 (0.020)
*Correlation coefcients were computed by Kendalls -b.
**All P values were one-sided.

with breast cancer would offer signicant advantages.
For instance, the ability to identify axillary nodal posi-
tivity would obviate the need for sentinel lymph node
biopsy before neoadjuvant chemotherapy. Additionally,
the ability to reliably identify patients without axillary
nodal metastasis could identify a subset of patients in
whom axillary sampling and the attendant morbidity
could be avoided.
The characteristics needed to reliably evaluate axil-
lary lymph node metastasis with MR have not been well
dened. In a study of 65 patients with dynamic contrast
MR, Kvistad et al found that ow kinetics signicantly
correlated with presence of axillary metastasis (12).
This study included patients who had been treated with
and without neoadjuvant chemotherapy. Mumtaz et al
examined the axillae of 78 patients with nondynamic
breast MR (13). These authors demonstrated a correla-
tion between suspicious nodes on MR (arbitrarily de-
ned as size 5 mm and contrast enhancement) and
pathologic nodal positivity. Murray et al reported in a
study of 47 patients, a correlation between nodal en-
hancement and nodal area on MR and pathologic nodal
positivity (14). Several other authors have attempted to
correlate MR features of the primary tumor with nodal
positivity; however, these studies have yielded incon-
sistent results (15,16). An overview of published studies
investigating the use of MR for the evaluation of axillary
nodal metastasis in breast cancer is shown in Table 3.
In the present study, there is a signicant correlation
between the presence of axillary lymph nodes with no
fatty hilum (on T2 STIR images) and pathologic node
positivity (Table 2). This is, to our knowledge, the rst
report of this association. There was signicant overlap
between the MR appearance of axillary nodes in pa-
tients with and without pathologic axillary metastasis
with regard to number of nodes, size, and predominate
ow kinetics. It is important to note that no patients in
this study received neo-adjuvant chemotherapy. The
presently identied MR characteristics may not be ap-
plicable in those patients. This is an area worth further
investigation.
A limitation of MR with current technology may be the
detection of low volume metastasis or micrometastasis
in axillary lymph nodes. This cannot be adequately
evaluated in the present study as the majority of pa-
tients had multiple pathologically positive axillary
nodes. Only 5% of the entire patient cohort presented
with a single pathologically positive axillary node. Val-
idation of the present ndings is warranted in a larger
cohort of patients.
The application of ultra-small iron oxide particles as
contrast agents for lymph node identication with MR
is presently evolving (17). It is likely that further devel-
opments in MR imaging technology, and nanotechnol-
ogy will lead to the expansion of the application of MR to
loco-regional staging and treatment of breast cancer
(18,19).
In conclusion, there are currently no widely accepted
MR criteria for the determination of axillary nodal sta-
tus in breast cancer patients. In this work, we have
identied a new breast MR parameter (axillary lymph
nodes with no fatty hilum), which correlates with nodal
metastasis. The results do not suggest that breast MR
should be ordered solely for axillary staging at present.
However, useful information may be gained regarding
the axilla in those patients in whom breast MR is being

Table 3
Review of Current Literature Regarding the Use of MR in the Identication of Axillary Metastasis
No. of
Authors Year MR correlates of axillary nodal positivity Comments
patients
Kvistad12 2000 65 Increased rate of contrast enhancement Study includes both patients with neo-adjuvant
treatment and without neo-adjuvant
treatment
Mumtaz13 1997 78 Presence of contrast and nodal size Does not examine the use DCE-MRI ow
0.5mm kinetics for evaluation of axillary nodes
Murray14 2002 47 21% contrast enhance of node All patients received axillary lymph node
compared to fatty tissue and nodal dissection at the time of primary resection
area 0.4 mm2
312
performed. Patients with suspicious axillary nodes seen
on breast MR should be considered for image-guided
needle biopsy of the node(s), which if positive, would
obviate the need for sentinel node biopsy. Interpreta-
tion of the axillary ndings on MR should not be based
on criteria used for evaluating breast lesions. Further
investigations of these ndings are clearly warranted
offering hope for modication of current paradigms for
clinical staging and management of the axilla in breast
cancer patients.
REFERENCES
1. Saslow D, Boetes C, Burke W, et al. American Cancer Society
guidelines for breast screening with MRI as an adjunct to mam-
mography. CA Cancer J Clin 2007;57:75 89.
2. Tilanus-Linthorst MM, Obdeijn AI, Bontenbal M, Oudkerk M. MRI
in patients with axillary metastases of occult breast carcinoma.
Breast Cancer Res Treat 1997;44:179 182.
3. Henry-Tillman RS, Harms SE, Westbrook KC, Korourian S, Klim-
berg VS. Role of breast magnetic resonance imaging in determining
breast as a source of unknown metastatic lymphadenopathy. Am J
Surg 1999;178:496 500.
4. Mortellaro VE, Marshall J, Harms SE, Hochwald SN, Copeland EM,
Grobmyer SR. Breast MR for the evaluation of occult nipple dis-
charge. Am Surg 2008;74:739 742.
5. Grobmyer SR, Mortellaro VE, Marshall J, et al. Is there a role for
routine use of MRI in the selection of patients for breast conserving
cancer surgery? J Am Coll Surg 2008;5:10451050.
6. Liberman L, Mason G, Morris EA, Dershaw DD. Does size matter?
Positive predictive value of MRI-detected breast lesions as a func-
tion of lesion size. AJR Am J Roentgenol 2006;186:426 430.
7. Clayton F, Hopkins CL. Pathologic correlates of prognosis in lymph
node-positive breast carcinomas. Cancer 1993;71:1780 1790.
8. Masood S. Prognostic/predictive factors in breast cancer. Clin Lab
Med 2005;25:809 825.
Mortellaro et al.
9. Giuliano AE, Kirgan DM, Guenther JM, Morton DL. Lymphatic
mapping and sentinel lymphadenectomy for breast cancer. Ann
Surg 1994;220:391398; discussion 398 401.
10. Ueda S, Tsuda H, Asakawa H, et al. Utility of 18F-uoro-deoxyglu-
cose emission tomography/computed tomography fusion imaging
(18F-FDG PET/CT) in combination with ultrasonography for axil-
lary staging in primary breast cancer. BMC Cancer 2008;8:165.
11. Fuster D, Duch J, Paredes P, et al. Preoperative staging of large
primary breast cancer with [18F] uorodeoxyglucose positron
emission tomography/computed tomography compared with con-
ventional imaging procedures. J Clin Oncol 2008;26:29.
12. Kvistad KA, Rydland J, Smethurst HB, et al. Axillary lymph node
metastases in breast cancer: preoperative detection with dynamic
contrast-enhanced MRI. Eur Radiol 2000;10:1464 1471.
13. Mumtaz H, Hall-Craggs MA, Davidson T, et al. Staging of symptom-
atic primary breast cancer with MR imaging. AJR Am J Roentgenol
1997;169:417 424.
14. Murray AD, Staff RT, Redpath TW, et al. Dynamic contrast en-
hanced MRI of the axilla in women with breast cancer: comparison
with pathology of excised nodes. Br J Radiol 2002;75:220 228.
15. Hsiang DJ, Yamamoto M, Mehta RS, et al. Predicting nodal status
using dynamic contrast-enhanced magnetic resonance imaging in
patients with locally advanced breast cancer undergoing neoadju-
vant chemotherapy with and without sequential trastuzumab.
Arch Surg 2007;142:855 861; discussion 860 861.
16. Szabo BK, Aspelin P, Kristoffersen Wiberg M, Tot T, Bone B. Inva-
sive breast cancer: correlation of dynamic MR features with prog-
nostic factors. Eur Radiol 2003;13:24252435.
17. Memarsadeghi M, Riedl CC, Kaneider A, et al. Axillary lymph node
metastases in patients with breast carcinomas: assessment with
non-enhanced versus uspio-enhanced MR imaging. Radiology
2006;241:367377.
18. Cuenca AG, Jiang H, Hochwald SN, Delano M, Cance WG, Grob-
myer SR. Emerging implications of nanotechnology on cancer di-
agnostics and therapeutics. Cancer 2006;107:459 466.
19. Sharma P, Brown S, Bengtsson N, et al. Gold speckled multimodal
nanoparticles for non-invasive bioimaging. Chem Mater 2008;20:
6087 6094.