Download Accuracy of Predicting Axillary Lymph Node Positivity by Physical

Ann Surg Oncol
DOI 10.1245/s10434-011-2200-7
ORIGINAL ARTICLE – BREAST ONCOLOGY
Accuracy of Predicting Axillary Lymph Node Positivity
by Physical Examination, Mammography, Ultrasonography,
and Magnetic Resonance Imaging
Stephanie A. Valente, DO1, Gary M. Levine, MD3, Melvin J. Silverstein, MD1,2, Jessica A. Rayhanabad, MD1,
Janie G. Weng-Grumley, MD1, Lingyun Ji, MS4, Dennis R. Holmes, MD1, Richard Sposto, PhD4,
and Stephen F. Sener, MD1
1
Division of Breast and Soft Tissue Surgery, Department of Surgery, Keck School of Medicine, University of Southern
California, Los Angeles, CA; 2Division of Breast Services, Department of Surgery, Hoag Memorial Hospital Presbyterian,
Newport Beach, CA; 3Division of Breast Services, Department of Radiology, Hoag Memorial Hospital Presbyterian,
Newport Beach, CA; 4Department of Preventive Medicine, University of Southern California, Los Angeles, CA
ABSTRACT
Background. Axillary lymph node status continues to be
among the most important prognostic variables regarding
breast cancer survival. We were interested in our ability to
accurately predict axillary nodal involvement by using
physical examination and standard breast imaging studies
in combination.
Methods. A retrospective review was performed of 244
consecutive patients diagnosed with invasive breast carcinoma between May 2008 and December 2010 who
underwent physical examination of the axilla, digital
mammography, axillary ultrasonography, and contrastenhanced breast magnetic resonance imaging and who had
subsequent histopathologic evaluation of one or more
axillary lymph nodes.
Results. A total of 62 (25%) of 244 women were found to
have positive axillary lymph nodes on final histopathologic
examination, 42% of whom were able to be identified
preoperatively. The sensitivity for predicting axillary
metastasis if any one or more examination modalities were
suspicious was 56.5%. The specificity for predicting axillary metastasis if any three or more modalities were
suspicious was 100%. Of the patients who had all four
modalities negative, 14% were ultimately found to have
histologically positive nodes at the time of surgery.
Ó Society of Surgical Oncology 2012
First Received: 18 April 2011
S. A. Valente, DO
e-mail: [email protected]
Conclusions. Physical examination and multimodal
imaging in combination are useful for preoperative axillary
staging and treatment planning. However, they remain
inadequate definitive predictors of axillary lymph node
involvement.
Axillary lymph node status remains among the most
important breast cancer prognostic factors and is essential
for establishing treatment decisions. The standard for
determining axillary involvement is via a sentinel lymph
node biopsy (SLNB). SLNB uses radiotracer or blue dye to
identify the first node or nodes that drain the breast, and
thus the initial nodes to encounter metastatic disease. The
SLNB is typically performed at the time of surgical
resection of the primary tumor and has an accuracy of
93.5–97.5%.1–4 This invasive surgical staging procedure
carries associated morbidity including longer surgical time,
an additional surgical scar, painful preoperative injections,
lymphedema, seroma, and possible sensory parasthesias.5,6
Knowledge of axillary lymph node involvement before
surgical intervention has allowed for improvements in
individualized multidisciplinary treatment options. These
include offering neoadjuvant chemotherapy, planning for
immediate reconstruction, or considering the use of intraoperative accelerated partial breast radiotherapy.
Additionally, if a patient is diagnosed with metastatic
axillary disease preoperatively, the surgeon has the
opportunity to discuss with the patient the specific indications for performing an axillary lymph node dissection.
Currently, there is no noninvasive diagnostic alternative
as accurate as the sentinel lymph node technique for
S. A. Valente et al.
axillary lymph node staging. Multiple nonsurgical methods
have been used with variable success to predict lymph node
involvement including physical examination (PE), digital
mammography (MMG), ultrasonography (US), computed
tomographic scan, positron emission tomographic imaging,
and magnetic resonance imaging (MRI).7–16 If a suspicious
lymph node is found on imaging, patients may undergo
US-guided fine-needle aspiration or core needle biopsy to
obtain a cytologic or histologic diagnosis. Preoperative
axillary lymph node needle biopsy has increased in frequency and has a high positive predictive value.17
However, accurately screening patients to determine whether axillary lymph nodes are suspicious and require
subsequent needle biopsy remains a challenge.
At our breast center, women who present with the
diagnosis of breast cancer are routinely investigated with
MMG, focused breast and axillary US and bilateral breast
contrast-enhanced MRI. We hypothesized that combining
the results of PE and these imaging modalities would
increase the sensitivity, specificity and accuracy of
detecting metastatic involvement of axillary nodes before
definitive surgical intervention.
METHODS
We reviewed a prospectively collected database after
obtaining institutional review board approval. We identified 264 consecutive patients who sought care from May
2008 to December 2010 for newly diagnosed invasive
breast carcinoma. Per routine assessment, patients had their
breast mass and axillary lymph nodes evaluated by PE and
imaged by MMG, US, and MRI. A total of 244 women
were identified who had all four evaluation examinations
completed and had subsequent surgery performed at our
institution. All images were reviewed by dedicated breast
radiologists who were not blinded to the fact that patients
had an invasive primary breast cancer. PE and all imaging
modalities were sequentially and prospectively interpreted
and results were recorded in our breast cancer database.
Physical Examination
An axillary lymph node was identified as positive if it
could be palpated by at least one physician on PE. It was
recorded as negative if a node was not palpable on PE,
even if it appeared suspicious on imaging.
were noted as suspicious. Lymph nodes considered
abnormal on mammogram included size [2 cm, rounded
or irregular shape, spiculated margins, absence of a fatty
hilum, or increased density.8
US Imaging
All patients received an US of their axilla on the side of the
involved breast. The axilla was scanned with a high-frequency transducer (17 Hz) and any abnormal lymph nodes
were noted as suspicious. Lymph nodes considered abnormal
on US included rounded shape, a long-to-short axis ratio of
\2, hypoechoic, compression or disappearance of the fatty
hilum, or cortical thickening or asymmetry.10,11,17
Magnetic Resonance Imaging
Per our standard breast cancer protocol, all patients
received dynamic contrast-enhanced MRI of bilateral
breasts and axilla. The axillary lymph nodes were evaluated with pre- and postcontrast images using the breast
MRI contrast protocol according to Kvistad et al.15 Any
abnormal lymph nodes were recorded as suspicious.
Lymph nodes considered abnormal on MRI included size
[10 mm, rounded shape, eccentric cortical hypertrophy or
abnormal signal intensity enhancement on T1-weighted
images.15,16
Surgical Management of Axillary Lymph Nodes
A patient with a negative axilla by PE and all three
imaging modalities underwent a SLNB at the time of surgery. SLNB was performed with radioactive colloid and
blue dye.2 If the sentinel lymph node was positive, a
completion level I and level II axillary lymph node dissection was performed.
If a patient was found to have a suspicious axillary
lymph node on PE or any imaging modality and the node
could be visualized by US, an US-guided core needle
biopsy was performed. If there was a histologically positive
lymph node on needle biopsy, the patient received a level I
and level II axillary lymph node dissection at the time of
her breast operation. If the core biopsy was negative for
metastatic disease, the patient received a SLNB at the time
of surgery.
Digital Mammography
Histopathologic Evaluation
MMG consisted of baseline mediolateral and craniocaudal views and any additional assessment images.
Mammograms were evaluated and in addition to the suspicious breast mass, any abnormal appearing lymph nodes
US-guided lymph node core needle biopsies were
evaluated by pathologists using permanent sectioning.
SLNB samples were evaluated by immediate frozen
section and hematoxylin and eosin staining. The nodes
Predicting Axillary Lymph Node Positivity
were subsequently submitted for permanent sectioning and
immunohistochemical assay. As per the American Joint
Committee on Cancer breast cancer stage classification
published in 2010, any patients with isolated tumor cells
(N0i?) were considered to be node negative and did not
undergo any additional lymph node surgery.18
Statistical Analysis
The preoperative predictive and postoperative final
results were confirmed. The overall sensitivity, specificity,
positive predictive value, negative predictive value, and
accuracy of each test and combination of tests were
calculated.
The sensitivity and specificity for each modality individually and in combination were then plotted. The
McNemar’s test was used to compare the sensitivity or
specificity between two modalities or combinations of
modalities.
RESULTS
We identified and evaluated a total of 244 women with
invasive breast carcinoma who had all four modalities
completed and had surgery performed at our institution.
Mean patient age was 58 years (range 31–85 years).
Invasive ductal carcinoma was diagnosed in 216 patients
(88.5%) and invasive lobular carcinoma in 28 patients
(11.5%). At final surgical pathology, a total of 62 of the
244 patients (25%) with breast carcinoma had one or more
positive lymph nodes. Of note, 13 patients had lymph
nodes with isolated tumor cells (N0i?) found at final
pathologic evaluation. These patients were included in the
node negative group and no additional lymph nodes were
removed.
Table 1 summarizes the distribution of tumor characteristics including tumor stage and lymph node involvement.
TABLE 1 Breast cancer type and tumor stage with associated lymph
node positivity
Cancer type
Number (N = 244)
Positive lymph node (%)
Invasive lobular
28
4 (14)
Invasive ductal
216
58 (27)
Cancer T stage
244
With increasing tumor size, the incidence of lymph node
involvement increased. Patients with T1mic disease had no
lymph node involvement, while 50% of patients with T3
disease had one or more positive lymph nodes.
PE, MMG, US, and MRI modalities were evaluated
individually and then in various increasing combinations
for their ability to detect suspicious axillary lymph nodes.
There were four combination groups in total: any one or
more modalities, any two or more modalities, any three or
more modalities, and all four modalities.
Tables 2 and 3 illustrate the diagnostic performance of
PE, MMG, US, and MRI individually and in combination.
All individual examinations had similar accuracy rates in
predicting the presence or absence of lymph node
involvement (accuracy 79.5–83.2%).
Further assessment of the sensitivity and specificity of
each modality and combination of modalities showed that
higher sensitivity resulted in lower specificity (Table 3;
Fig. 1). As shown in Fig. 1, when evaluating sensitivity,
the different modalities or combinations of increasing
numbers of modalities seemed to cluster into four groups.
The approach of assessing a patient as having suspicious
lymph nodes when one or more testing modalities was
suspicious showed the highest sensitivity at 56.5%. The US
examination alone had the second highest sensitivity for
identifying suspicious lymph nodes at 43.5%. Declaring
one or more modality suspicious was significantly more
sensitive than relying on US alone (P = 0.005). However,
despite its high sensitivity, one or more modalities also had
a significantly worse specificity (91.8%) compared to any
of the other approaches evaluated (e.g., P = 0.005 for the
comparison in specificity between more than one modality
vs. US).
The next group on the graph was MRI, two or more
modalities, and PE. Although it appeared that US had a
similar specificity but better sensitivity than MRI (Fig. 1),
the difference in sensitivity between the two examinations
was not statistically significant (P = 0.25). Interestingly,
the sensitivity for the last group of three approaches
(MMG, three or more modalities, and all four modalities)
was significantly lower than the other tests (P \ 0.01,
Fig. 1). This low sensitivity was offset by the fact that
specificity for predicting axillary metastases was 100% if
0 (0)
TABLE 2 Diagnostic performance results of the four modalities in
detecting lymph node involvement in breast cancer patients
20
38
1 (0.5)
6 (16)
Finding
T1c
89
17 (19)
Truly positive
22
13
27
23
T2
76
30 (39)
Truly negative
179
181
175
176
T3
16
8 (50)
Falsely positive
3
1
7
6
T4
0
0 (0)
Falsely negative
40
49
35
39
T1mic
T1a
T1b
5
PE
MMG
US
MRI
S. A. Valente et al.
TABLE 3 Sensitivity, specificity, accuracy, positive predictive value, and negative predictive value for each modality and combination of
modalities in detecting lymph node involvement in breast cancer patients: Final pathology results
Lymph node positive,
N1–3 = 62
Lymph node negative,
N0 = 182
Total (n = 244)
Sensitivity (%)
Specificity (%)
Accuracy (%)
PE
22 (35.5)
179 (98.4)
201 (82.4)
88
81.7
MMG
13 (21.0)
181 (99.5)
194 (79.5)
92.9
78.7
US
27 (43.5)
175 (96.2)
202 (82.8)
79.4
83.3
MRI
23 (37.1)
176 (96.7)
199 (81.6)
79.3
81.9
C1
35 (56.5)
167 (91.8)
202 (82.8)
70
86.1
C2
23 (37.1)
180 (98.9)
203 (83.2)
92
82.2
C3
15 (24.2)
182 (100)
197 (80.7)
100
79.5
All 4
12 (19.4)
182 (100)
194 (79.5)
100
78.4
Positive predictive
value (%)
Negative predictive
value (%)
Suspicious by:
Modalities used
core needle biopsy before surgery and proceeded straight to
axillary lymph node dissection.
Sensitivity
1.0
Clinical exam
MMG
0.9
US
MRI
Any one
Any two
Any three
All four
Lymph Node Involvement
0.8
0.7
In our lymph node positive group, 38 (61%) of 62
women had more than one lymph node positive on axillary
lymph node dissection. In the 14% of patients where all
four modalities were negative with a subsequent positive
SLNB, 15 (56%) of 27 women had more than one lymph
node involved.
0.6
0.5
0.4
0.3
0.2
0.1
DISCUSSION
0.0
0 .1
0.2
0.3
0.4
0.5
0.6
0.7
0 .8
0 .9
1.0
1-Specificity
FIG. 1 Sensitivity and 1-specificity for each modality and combinations of modalities for detecting lymph node involvement in breast
cancer patients
there were three or more modalities or all four modalities
positive.
Core Biopsy Evaluation
Figure 2 shows that either by PE, MMG, US, or MRI, a
total of 50 patients were found preoperatively to have
suspicious lymph nodes. Of the 50 patients, 35 (70%) were
able to have the suspicious lymph node identified by subsequent US and therefore received an US-guided core
needle biopsy. Biopsy correctly diagnosed 26 (90%) of 29
patients with metastatic disease. There were no falsepositive results. Ultimately, 26 (42%) of 62 patients with
positive lymph nodes had axillary metastases confirmed via
In breast cancer patients, preoperative clinical staging
and planning is of paramount importance because positive
axillary lymph node metastasis changes many of the
treatment and surgical options offered to patients. The
ability to assemble this information before definitive surgical intervention has improved greatly with the advent of
new imaging and minimally invasive biopsy techniques.
The objective of this study was to evaluate the four
modalities (PE, MMG, US, and MRI), most commonly
used to evaluate axillary nodal status to assess whether the
sensitivity, specificity and accuracy for identifying pathologic lymph nodes could be improved if the results of such
examinations were evaluated in combination.
PE is the oldest and most rudimentary method used to
evaluate lymph node status. The sensitivity of our PE in
detecting disease (35.5%) is similar to what others have
reported at 25–32.3%.9,15 At PE, the physician cannot
differentiate between an enlarged lymph node that is cancerous versus one that is inflamed or reactive; which may
explain why specificity is low with this method.
Predicting Axillary Lymph Node Positivity
FIG. 2 Schematic of lymph node
evaluation
Women with breast cancer
N = 264
All 4 modalities not
completed or surgery not
performed at our institution
N = 20
Total patients
N = 244
Any modality
Suspicious LN
N = 50
Core biopsy
not done
N = 15
SLNB
positive
N=6
SLNB
negative
N=9
MMG is the standard imaging modality used in
screening for breast disease. Although axillary lymph
nodes can be visualized on some of the imaging projections, it is not consistent. This is because with MMG
positioning, most of the axilla is pushed out of the image
fields; thus, only the lower part of the axilla can be visualized.8 This makes MMG a less sensitive method for
imaging the axilla; however, if there are suspicious nodes
identified on MMG, the results from our data show that it is
highly likely that they are pathologic (specificity 99.5%).
For breast cancer evaluation, MRI is currently the best
study to show anatomy in relation to pathology. In addition
to level I and level II axillary lymph nodes, it allows
evaluation of internal mammary and level III lymph nodes.
The downside to this examination is accessibility, cost, and
a patient’s physical restrictions (claustrophobia, kidney
function, implanted metal objects, etc). Nonetheless, MRI
has increasingly been used for breast cancer evaluation
with a reported broad sensitivity range of for detecting
axillary metastasis reported as 36–78% and specificity
93–100%.14–16
Axillary US is a portable and simple test that is routinely
used preoperatively to evaluate lymph node involvement.
Similar to our results, axillary US sensitivity for abnormal
lymph nodes is reported as 45.2–86.2% and specificity is
40.5–86.6%.9–11,17 Because of its ease, it is used to guide
biopsies of axillary lymph nodes when indicated.
With the addition of US-guided fine-needle aspiration or
core needle biopsy of suspicious lymph nodes, specificity for
detecting metastatic lymph nodes can be increased.11,14,17
Core biopsy
positive
N = 26
SLNB
positive
N = 26
SLNB
negative
N=0
Any modality
No suspicious LN
N = 194
Core biopsy
negative
N=9
SLNB
positive
N=3
SLNB
negative
N=6
SLNB
positive
N = 27
SLNB
negative
N = 167
Recent studies have reported successfully diagnosing
7.8–16.2% of patients with axillary involvement preoperatively via US guided FNA.17,19 In our study, using all
modalities as screening tools and confirming with US
guided core biopsy, 11% of our total patients were diagnosed with axillary metastases before surgery and were
spared a SLNB procedure. In other words, of the 25% of
our cancer patients found to have metastases to the axillary
lymph nodes, 42% of them were able to be detected
preoperatively.
PE and current imaging technologies are not able to
detect all suspicious lymph nodes because a macroscopic
amount of tumor burden is required for an abnormal lymph
node to be detected on palpation or imaging.8 Microscopic
disease cannot be expected to be detected, but remains
clinically important.
If a lymph node appears suspicious on any imaging
modality, it is generally biopsied under US guidance.
However, it is difficult to ensure exact correlation of specific axillary lymph nodes detected by different imaging
modalities. This creates a problem with targeting when the
modality that initially detected the suspicious lymph node
cannot be used to guide the needle biopsy. Additionally,
percutaneous lymph node needle biopsy only samples a
portion of a lymph node and therefore may miss a cluster of
tumor cells present elsewhere in the lymph node. Moreover, it has been found that only 64–78.3% of the nodes
targeted as suspicious on preoperative imaging truly correlate with the sentinel lymph node or nodes found by
SLNB.20,21 For all these reasons, a negative needle biopsy
S. A. Valente et al.
result does not exclude metastatic disease or preclude the
need for SLNB.
In this study, the combination of PE, MMG, US, and
MRI demonstrated a trade off in sensitivity and specificity
for prediction of lymph node involvement in breast cancer
patients.
Our data suggest that if any one or more of these axillary
examination modalities revealed a suspicious node, the
patient has at least a 56.5% chance of harboring metastatic
disease within the lymph node, and US-guided needle
biopsy of the suspicious node should be recommended. If
three or four modalities are suspicious, the chance that the
lymph node contains disease is 100%, though US core
needle biopsy should still be performed to pathologically
confirm the presence of metastatic carcinoma. The results
also indicate that if all four evaluation modalities are
negative, the patient has an 86% chance of having negative
lymph nodes on final surgical SLNB pathology. However,
this still leaves approximately 14% of patients with positive sentinel nodes despite a negative preoperative
assessment.
Our results confirm that currently, there is no imaging
modality or combination of modalities that have an accuracy rate as high as the gold standard of surgical SLNB, or
that reliably identifies a subset of patients with a negative
assessment in whom sentinel node biopsy may be safely
omitted. However, preoperative PE and imaging of the
axilla with subsequent percutaneous biopsy when indicated
did allow 42% of our patients with axillary metastasis to
have knowledge of their lymph node involvement before
surgical intervention.
This study shows that although PE and imaging studies,
when negative, cannot be as dependable as SLNB, collectively, their information is better than individual
imaging assessment. Additionally, when these results are
positive for axillary metastasis, they offer patients more
extensive preoperative treatment knowledge and options,
and for this reason, they should be included as part of
preoperative breast cancer assessment.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
REFERENCES
1. Giuliano AE, Jones RC, Brennan M, Statman R. Sentinel lymphadenectomy in breast cancer. J Clin Oncol. 1997;15:2345–50.
2. McMasters KM, Tuttle TM, Carlson D J, et al. Sentinel lymph
node biopsy for breast cancer: a suitable alternative to routine
axillary dissection in multi-institutional practice when optimal
technique is used. J Clin Oncol. 2000;18:2560–6.
3. Veronesi U, Paganelli G, Viale G, et al. A randomized comparison of sentinel-node biopsy with routine axillary dissection in
breast cancer. N Engl J Med. 2003;349:546–53.
4. Krag DN, Anderson SJ, Julian TB, et al. Technical outcomes of
sentinel-lymph-node resection and conventional axillary-lymph-
19.
20.
21.
node dissection in patients with clinically node-negative breast
cancer: results from the NSABP B-32 randomized phase III trial.
Lancet Oncol. 2007;8:881–8.
Crane-Okada R, Wascher RA, Elashoff D, Giuliano AE. Longterm morbidity of sentinel node biopsy versus complete axillary
dissection for unilateral breast cancer. Ann Surg Oncol. 2008;15:
1996–2005.
Purushotham AD, Upponi S, Klevesath MB, et al. Morbidity after
sentinel lymph node biopsy in primary breast cancer: results from
a randomized controlled trial. J Clin Oncol. 2005;23:4312–21.
Specht MC, Fey JV, Borgen PI, Cody HS. Is the clinically
positive axilla in breast cancer really a contraindication to sentinel lymph node biopsy? J Am Coll Surg. 2005;200:10–4.
Shetty MK, Carpenter WS. Sonographic evaluation of isolated
abnormal axillary lymph nodes identified on mammograms. J
Ultrasound Med. 2004;23:63–71.
Pamilo M, Soiva M, Lavast EM. Real-time ultrasound, axillary
mammography, and clinical examination in the detection of
axillary lymph node metastases in breast cancer patients. J
Ultrasound Med. 1989;8:115–20.
Nori J, Vanzi E, Bazzocchi M, et al. Role of axillary ultrasound
examination in the selection of breast cancer patients for sentinel
node biopsy. Am J Surg. 2007;193:16–20.
Lee MC, Eatrides J, Chau A, et al. Consequences of axillary
ultrasound in patients with T2 or greater invasive breast cancers.
Ann Surg Oncol. 2010;18:72–7.
March DE, Wechsler RJ, Kurtz AB, Rosenberg AL, Needleman
L. CT-pathologic correlation of axillary lymph nodes in breast
carcinoma. J Comput Assist Tomogr. 1991;15:440–4.
Peare R, Staff RT, Heys SD. The use of FDG-PET in assessing
axillary lymph node status in breast cancer: a systematic review
and meta-analysis of the literature. Breast Cancer Res Treat.
2010;123:281–90.
Fernández AG, Fraile M, Giménez N, et al. Use of axillary
ultrasound, ultrasound-fine needle aspiration biopsy and magnetic
resonance imaging in the preoperative triage of breast cancer
patients considered for sentinel node biopsy. Ultrasound Med.
Biol. 2011;37:16–22.
Kvistad KA, Rydland J, Smethurst HB, Lundgren S, Fjosne HE,
Haraldseth O. Axillary lymph node metastases in breast cancer:
preoperative detection with dynamic contrast-enhanced MRI. Eur
Radiol. 2000;10:1464–71.
Yoshimura G, Sakurai T, Oura S, et al. Evaluation of axillary
lymph node status in breast cancer with MRI. Breast Cancer.
1999;6:249–58.
Park SH, Kim MJ, Park BW, Moon HJ, Kwak JY, Kim EK.
Impact of preoperative ultrasonography and fine-needle aspiration of axillary lymph nodes on surgical management of primary
breast cancer. Ann Surg Oncol. 2011;18:738–44.
Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A.
American Joint Commision on Cancer. Breast cancer: prespectives on anatomic staging—based on the AJCC staging manual,
7th ed. New York: Springer, 2010. p. 56.
Baruah BP, Goyal A, Young P, Douglas-Jones AG, Mansel RE.
Axillary node staging by ultrasonography and fine-needle aspiration cytology in patients with breast cancer. Br J Surg. 2010;
97:680–3.
Britton PD, Provenzano E, Barter S, et al. Ultrasound guided
percutaneous axillary lymph node core biopsy: how often is the
sentinel lymph node being biopsied? Breast. 2009;18:13–6.
Nathanson SD, Burke M, Slater R, Kapke A. Preoperative identification of the sentinel lymph node in breast cancer. Ann Surg
Oncol. 2007;14:3102–10.