Application of MR Mammography Beyond Local Staging: Is There a Potential to Accurately Assess Axillary Lymph Nodes? Evaluation of an Extended Protocol in an Initial Prospective Study
Abstract
OBJECTIVE. The purpose of our study was to clinically test an extended MR mammography (MRM) protocol for combined local staging (T-staging) and locoregional staging (N-staging) of breast cancer within one single examination using a dedicated whole-body scanner.
SUBJECTS AND METHODS. Fifty-six consecutive primary breast cancer patients without prior treatment underwent MRM and surgicopathological N-staging. The MRM protocol (10 minutes; axial T1-weighted gradient-recalled echo; dynamic contrast-enhanced; T2-weighted; turbo spin-echo) was extended to evaluate axillary lymph nodes (90 seconds; coronal T2-weighted HASTE; T1-weighted volumetric breath-hold examination; field of view, both axillae, supraclavicular nodes, and cervical nodes). A dedicated whole-body scanner was used. First, two experienced radiologists independently rated the presence of lymph node metastasis (present or absent, weighted kappa). Second, predefined descriptors were applied by both readers to differentiate lymph node status. These were statistically analyzed using univariate chi-square statistics, sensitivity and specificity, positive likelihood ratio, diagnostic odds ratio (OR), and multivariate statistics (binary logistic-regression, receiver operating characteristics, and chi-squared automatic interaction detection [CHAID] tree).
RESULTS. Most significant predictors (p < 0.001) of present metastasis were “irregular margin” (diagnostic OR, 14.0), “inhomogeneous cortex” (diagnostic OR, 8.4), “perifocal edema” (positive likelihood ratio, 100) and “asymmetry” (diagnostic OR, 19.5). CHAID tree identified “asymmetry” and “irregular margin” as significant predictors (adjusted-p < 0.05) for present metastasis (PPV: 100%), whereas absence of “asymmetry” and “homogeneous internal structure” were highly predictive of absent metastasis (negative predictive value, 94.3%). Combination of significant descriptors using binary logistic regression revealed an area under the receiver operating characteristic curve of 0.93 (p < 0.001). Interrater agreement was “almost-perfect” (κ = 0.95).
CONCLUSION. Combined T-staging and locoregional staging (N-staging) was possible within one imaging session using the proposed protocol. Despite a minimal increase in examination time, high diagnostic accuracy and excellent interrater reliability were achieved.
Introduction
Breast cancer is the most common cancer in women in the western world [1]. Before therapy is initiated, numerous staging procedures have to be performed, including local staging (T-staging) and locoregional staging (N-staging). These two steps are important because T-staging and presence of axillary lymph node metastases have significant impact on the overall prognosis as well as the therapeutic approach [2, 3].
Currently, accuracy of diagnostic imaging in the pretherapeutic assessment of nodal status in primary breast cancer is limited [4]. Accordingly, surgical axillary lymph node sampling or clearance remains the standard of care. However, it can be associated with long-term morbidity and side effects even if less invasive procedures are performed, such as sentinel biopsy. Many primary cancers are now treated with neoadjuvant chemotherapy [5]. Because surgical staging is usually performed in combination with definite breast surgery, pretherapeutic surgical staging of axillary lymph nodes is difficult in this setting [5, 6]. These are but a few reasons why further refinement of noninvasive pretherapeutic procedures would significantly improve management of patients with primary breast cancer.
In its third decade of clinical investigation MR mammography (MRM) has become a standard imaging tool for pretherapeutic staging of breast cancer. Although research still is necessary, this modality has shown high evidence for identification of multifocal, multicentric, and contralateral disease as well as the assessment of tumor size [7–12]. However, use of MRI for the evaluation of axillary lymph nodes remains controversial. This is because complete bilateral assessment of axillary lymph nodes level I to III and the supraclavicular region is limited if only dedicated breast coils are used [13–15]. A dedicated examination of the axilla performed in the same or an additional session can be accurate but requires scanning times per axilla that are not clinically feasible [16, 17].
Using modern whole body MRI systems with multichannel capability, additional axillary staging sequences with increased field of view can be added to a dedicated MRM examination by combination of several surface coils at the same time. This would allow fast bilateral scanning of both breasts, both axillae, and the supraclavicular region. Accordingly, this prospective study was performed to design and clinically test an extended MRM protocol for fast bilateral scanning of both breasts, both axillae and the supraclavicular region with special respect to diagnostic accuracy of this approach to differentiate nodal positive from nodal negative breast cancers.
Subjects and Methods
Setting and Patients
This prospective single-center cross-sectional investigation was approved by the local ethical review board. Study design was planned according to appropriate guidelines [18]. Our breast MRI department is part of an academic multidisciplinary breast cancer center, which is certified according to DIN EN ISO 9001 standards. Every patient gave written consent before undergoing MRM. Data acquisition covered a consecutive series of patients imaged at our institution during 15 months. The inclusion criterion for examinations was patients referred for preoperative local staging with known breast cancer. Patients with cancer recurrence and secondary breast cancer were not eligible. This was necessary because posttherapeutic changes after lymph node dissection, e.g., scarring, adhesions, etc., would potentially bias evaluation. Similarly, pure in situ cancers were not eligible, because pretest probability of axillary metastasis is extremely low and thus potentially might bias statistical analysis [19].
Test Methods
Reference standard—Surgical verification was defined as the reference standard for N-staging. According to surgical routine at our affiliated clinic of gynecology, patients initially underwent sentinel lymph node sampling using periareolar injection of radiolabeled colloids (99Tc) and blue dye. In case of positive lymph nodes, axillary dissection of level I and II was performed. Both surgeons and pathologists were specialized in breast diseases and worked at our affiliated certified breast cancer center. Information on the number of metastatic and excised lymph nodes was collected. Lymph node classification was based on the TNM (6th release) classification system. Accordingly, lymph nodes were classified as negative, showing micrometastasis only (metastatic foci ≤ 2 mm), or showing macrometastases. For calculation of diagnostic parameters, lymph node status was dichotomized as “nodal status positive (micro- or macrometastases)” or “negative,” respectively.
Technical specifications of MRI (index test)— A dedicated whole-body 1.5-T MRI unit (Magnetom Avanto, Siemens Healthcare) providing multichannel capability was used for image acquisition. For signal detection, a dedicated four-channel receive-only bilateral breast coil was used in combination with multichannel phased-array surface body coils. Coils were provided by the vendor of the MRI unit and were combined for image data acquisition. Patient positioning was prone with both breasts hanging into the bilateral surface coil. Phased-array body coils were put on the back of the patient. Using this approach, repositioning of the patient during the MRI examination could be avoided.
The first part of the imaging procedure consisted of our conventional MRM protocol, which is in accordance to international guidelines [20]. All scans were acquired in the axial plane, beginning with repetitive (n = 8) spoiled dynamic T1-weighted 2D FLASH 2D gradient-echo sequences at 1-minute intervals (no fat saturation). After one unenhanced scan was obtained, gadodiamide (Omniscan, GE Healthcare) was administered IV as a rapid bolus (flow rate, 3 mL/s) by a power injector (Spectris, Medrad). After a 35-second delay from the start of contrast agent injection, the remaining (n = 7) measurements were performed. Technical parameters for T1-weighted scanning were as follows: TR/TE, 106/4.56; flip angle, 80°; in-plane resolution, 1.1 × 0.9 × 3 mm; basic matrix, 384 pixels; and time of acquisition, 1 minute 2 seconds. Additionally, one T2-weighted turbo spin-echo sequence without fat saturation was performed in the same orientation and slice position. Technical parameters for T2-weighted scanning were TR/TE effective, 8900/193; flip angle, 90°; in-plane resolution, 0.8 × 0.7 × 3 mm; basic matrix, 512 pixels; and time of acquisition, 2 minutes 15 seconds.
To further assess nodal status, two additional sequences were measured. Both were acquired in coronal orientation. Initially, T2-weighted HASTE was performed without breath-hold. Technical parameters were as follows: TR/TE effective, 1100/118; flip angle, 150°; spatial resolution, 1.4 × 1.1 × 5 mm; basic matrix, 448 pixels; field of view, 500 mm; 46 slices, distance factor, 10%; and time of acquisition, 51 seconds. Finally, a T1-weighted, contrast-enhanced (11 minutes after gadodiamide), fat-saturated, spoiled, 3D volumetric-interpolated gradient-echo sequence was performed as a breath-hold examination (VIBE). Technical parameters were as follows: TR/TE, 2.82/0.83; flip angle, 10°; spatial resolution, 1.8 × 1.6 × 2.5 mm; basic matrix, 320 pixels; field of view, 500 mm; 104 slices; and time of acquisition, 25 seconds. All imaging sequences were obtained using parallel imaging, factor 2. The field of view of additional sequences was optimized to cover all locoregional lymph nodes as well as adjacent compartments of lymphatic drainage. In particular axillae (levels I–III), supraclavicular nodes, and inferior cervical nodes (levels IV–VII) were imaged.
Interpretation of index test—All scans were independently and prospectively rated by two radiologists. Both had significant experience in MRI (one with greater than 3000 examinations and the other with greater than 1000 examinations). Besides knowledge of the diagnosis of “invasive breast cancer,” no clinical details were known. In particular, the readers were blinded to results of surgical N-staging. The latter was defined on an axilla and breast basis and not on single lymph nodes because this was a level III analysis according to Obuchowski et al. [21].
First, ipsilateral lymph nodes were independently assessed using predefined quantitative and qualita tive descriptors. The latter were derived from pre vious studies on this issue and on BI-RADS de scriptors [13, 22–24]. Definitions of qualitative descriptors are summarized in Table 1. Quantitative descriptors included measurement of short and long axes to calculate the Solbiati index [25]. On the basis of these criteria, both readers independently scored nodal status of the ipsilateral axilla. For this purpose, a 5-level confidence score was defined as follows: I (no lymph nodes), II (benign lymph nodes), III (probably benign lymph nodes), IV (probably metastatic lesion lymph nodes), and V (metastatic lesion lymph nodes). Results of this scoring were used for analysis of interrater agreement. In a second step, the two readers identified the most suspect ipsilateral lymph node in consensus and applied all descriptors as defined. Results of the most suspect lymph node and the global descriptor “asymmetry” were compared with the overall lymph node status (i.e., nodal status positive or nodal status negative) as defined earlier (level III analysis). According to the prospective study design, accuracy of our protocol was not known in advance. Consequently, interpretation results did not interfere with clinical patient management.
| Descriptor | Definition | Rating Scheme |
|---|---|---|
| Margin | Contour of the lymph node was classified as “smooth” vs “irregular.” |  |
| Cortex | Cortical structures could be rated as “homogeneous” if a normal external thin and C-shaped cortex was present. In absence of these characteristics, it was classified as “inhomogeneous.” If focal increase of cortical diameter by a factor of ≤ 2 was present, “nodular thickening” was diagnosed. |  |
| Hilus sign | Presence of fatty hilum was analyzed and could be classified as “present” vs “absent.” |  |
| Perifocal edema | “Perifocal edema” assesses presence of markedly prolonged T2 times of soft tissue surrounding the node. It could be classified as “present” vs “absent.” |  |
| Rim enhancement | “Rim enhancement” describes higher signal intensity of lymph node periphery in relation to its center on coronal T1-weighted volume-interpolated breath-hold scans 11 minutes after contrast agent injection and could be “present” or “absent.” |  |
| Asymmetrya | If axilla of cancer bearing breast revealed asymmetric lymph nodes in terms of number or size compared with contralateral side, “asymmetry” was diagnosed. |  |
a
Asymmetry was rated in comparison with the contralateral side. All other descriptors were rated in the most suspect lymph node identified by the readers.
Statistical Analysis
Ratings of both readers were compared using weighted kappa statistics. Kappa values of 0.4–0.6 were interpreted as “moderate,” > 0.6–0.8 as “substantial,” and > 0.8–1 as “almost perfect” [19]. The Mann-Whitney U test was used to test whether the two independent samples of benign and malignant lymph nodes came from the same distribution.
Basic analysis of descriptors was based on contin gency tables. Analysis included chi-square statistics, standard parameters of diagnostic accuracy (sensitivity, specificity, positive and negative likelihood ratios, diagnostic odds ratio). In case of quantitative parameters, receiver operating characteristic (ROC) curve analysis followed by quantification of the area under the ROC curve (AUC) was performed.
Significant association of all parameters with nodal status was further evaluated by binary logistic regression and chi-square automatic interaction detection tree (CHAID-tree) analysis. Binary logistic regression was used to identify overall accuracy for assessment of nodal status in our patient group. For this purpose, all significant parameters on basic analysis were entered into the model. Predictive values were saved and subsequently underwent ROC analysis with nodal status as an independent variable to quantify overall accuracy of the model. Then, a CHAID-tree analysis was calculated. The latter is able to detect and visualize significant descriptor combinations. Corresponding significance values were adjusted using the Bonferroni method. Statistical software packages, including PASW 18 (SPSS), MedCalc 11, (MedCalc), and Answer Trees (SPSS), were used for data analysis and generation of graphs. Multivariate statistics in this article have exploratory character.
Results
Patients
Fifty-six women fulfilled the inclusion criteria. The mean age was 60.3 years (SD, 10.4 years; range, 30–82 years). In three patients, bilateral breast cancer was present. Accordingly, 59 axillae were evaluated. The most common cancer typing was invasive ductal carcinoma (n = 40, 67.8%) and invasive lobular carcinoma (n = 6, 10.2%). Mixed invasive ductal and lobular carcinoma was present in seven patients (11.9%). Grading according to Elston and Ellis [26] was reported as “intermediate” in the majority of cases (grade 2, 62.7%; n = 37), with the remaining being grade 3 (28.8%; n = 17) or grade 1 cancers (5.1%; n = 1). In four cases, grading was not reported by pathology for technical reasons. Four patients (6.8%) showed micrometastatic (≤ 2 mm) foci, and 13 patients (22.0%) had lymph nodes with macrometastases on surgicopathologic verification (total, 28.8%). The mean number of resected lymph nodes was 24.9 (range, 11–45), and the mean number of positive lymph nodes was 7.7 (range, 1–44). The mean positive-to-resected lymph node ratio in cases with lymph node metastasis was 27.3% (SD, 32.2%). T-stage was higher for lymph node positive compared with lymph node negative cases (p = 0.006). The exact distribution of T-stages clustered by nodal status is given in Figure 1.
Estimates of Diagnostic Accuracy
Bilateral examination of axillae and the breast was technically successful in all patients. Ratings of both readers revealed “almost perfect” interrater agreement for nodal staging (weighted κ = 0.95).
Basic analysis—The mean longest diameter of lymph nodes was 13.0 mm (SD, 5.8 mm) versus 16.5 mm (SD, 9.6 mm) in nodal positive versus nodal negative cases (p = 0.18). The shortest lymph node diameter differed significantly between positive (10.8 mm; SD, 6.8 mm) and negative (6.3 mm; SD, 1.6 mm) lymph nodes (p = 0.016). The Solbiati index reached an AUC of 0.694 (95% CI, 0.558–0.830; p < 0.005). At a cutoff value of ≤ 1.875, sensitivity (88.2%; 95% CI, 63.6–98.5%), specificity (52.4%; 95% CI, 36.4–68.0%), and positive and negative likelihood ratios, were moderate (1.9 and 0.2). Diagnostic accuracy of categoric parameters is summarized in Table 2.
| Descriptor | p | Sensitivity (%) | Specificity (%) | PPV (%) | LR+ | LR– | DOR |
|---|---|---|---|---|---|---|---|
| Margin (irregular) | 0.001 | 41.2 | 95.2 | 77.8 | 8.6 | 0.6 | 14.0 |
| Cortex | |||||||
| Homogeneous | 0.0001 | 29.4 | 16.7 | 12.5 | 0.4 | 4.2 | 0.1 |
| Inhomogeneous | 47.1 | 90.5 | 66.7 | 4.9 | 0.6 | 8.4 | |
| Nodular thickening | 23.5 | 92.9 | 57.1 | 3.3 | 0.8 | 4.0 | |
| Hilus sign | 0.015 | 52.9 | 14.3 | 20.0 | 0.6 | 3.3 | 0.2 |
| Perifocal edema | 0.001 | 29.4 | 100.0 | 100.0 | ND | 0.7 | ND |
| Rim sign | 0.005 | 23.5 | 100.0 | 100.0 | ND | 0.8 | ND |
| Asymmetry | 0.0001 | 76.5 | 85.7 | 68.4 | 5.4 | 0.3 | 19.5 |
Note—p is given for nodal positive versus nodal negative breast cancers. PPV = positive predictive value, LR+ = positive likelihood ratio, LR– = negative likelihood ratio, DOR = diagnostic odds ratio. ND = not defined.
Multivariate analysis—All parameters significantly associated with nodal status were entered into binary logistic regression. The given model revealed significant potential to differentiate nodal status positive versus nodal status negative cancers (p < 0.001, Nagelkerke R2 = 0.663). Corresponding ROC analysis of predicted probabilities revealed an AUC of 0.93 (95% CI: 0.823–0.978). At a cutoff level of a predicted probability of > 0.4223, sensitivity of 87.5% (95% CI, 61.7–98.4%), specificity of 92.7% (95% CI, 80.1–98.5%), and positive and negative likelihood ratios of 11.96 and 0.13 were identified (Fig. 2 and Table 3). The determined CHAID-tree is given in Figure 3. Accordingly, the decision tree analysis identified combination of “asymmetry” and “irregular margin” as most accurate predictors for nodal status positive (positive predictive value [PPV], 100%), whereas absence of “asymmetry” and presence of “homogeneous cortex” were highly predictive of nodal status negative (negative predictive value [NPV], 94.3%). Figures 4 and 5 illustrate clinical examples of lymph node characteristics.
| Finding | LN– | LN+mic | LN+mac | Total |
|---|---|---|---|---|
| MRI negative | 39 | 1 | 1 | 41 |
| MRI positive | 3 | 3 | 12 | 18 |
| Total | 42 | 4 | 13 | 59 |
Note—LN– = nodal negative, LN+mic = nodal positive with micrometastases, LN+mac = nodal positive with macrometastases.
Discussion
Our study results show that axillary staging can be combined with a conventional MRM examination using a dedicated whole-body scanner. Such a protocol can be performed at one time, avoiding a significant increase in scanning time. Our initial results in a consecutive series of patients show high accuracy and reliability of our approach. The data are of clinical importance because they could help to solve a diagnostic dilemma, i.e., to accurately perform noninvasive N-staging without significantly increasing scanning time.
Unlike previous studies, we used a dedicated whole-body scanner [13–17]. This allowed both coverage of an appropriate field of view and fast image acquisition. Previous investigations applied conventional MRI, using surface coils for axillary imaging with or without additional breast coils [16, 17]. Accordingly, either scanning time or field of view had to be sacrificed. On the other hand, if breast coils are used for axillary evaluation, the field of view is always limited [13–15]. Therefore, evaluation of the whole anatomic region of interest, including bilateral examination to check for asymmetry and assessment of adjacent lymphatic compartments, was not possible [2]. As shown in Figure 4, our approach enables accurate assessment of the whole extent of locoregional disease at a glance, even if metastatic involvement is advanced.
Another difference from previous studies is the interpretation of the index test. As recommended by MRI BI-RADS, we used a mix of quantitative and qualitative descriptors [22]. These, were based on common interpretation criteria as they are frequently used in lymph node and breast imaging [13–17, 22–24]. Our catalog included evaluation of “asymmetry,” which is a basic diagnostic criterion in radiology and should be checked whenever possible and appropriate. Surprisingly, previous MRI studies on this issue did not use this descriptor for assessment of nodal status [13–17]. Applying our criteria in combination, an accurate differentiation between nodal positive versus negative cases was achieved and interrater agreement was excellent. However, our approach of rating most descriptors in the most suspect lymph node identified by the observers for diagnosis of the overall axilla lymph node status does have limitations. Classified as level III according to Obuchowski et al. [21], a dedicated lesion-by-lesion comparison of histopathologic features with the rated lymph node was not possible. Sensitivity and specificity calculations were thus performed for general N-stage. According to Obuchowski and coworkers, an overestimation of sensitivity as well as an underestimation of the false-positive rate is likely compared with a lesion-by-lesion approach in the case of multiple lesions per patient [21]. In the current study, three of four micrometastases were correctly predicted by MRI (Table 3). Because of technical considerations, these lesions ≤ 2 mm may not be visualized directly. Therefore, other sensitivity and specificity values for single-lesion descriptors as well as overall diagnostic accuracy might be found in such an analysis. However, the purpose of our approach is screening to determine whether the patient has any lymph node metastases or not. In the case of positive findings, further procedures are scheduled. Consequently, we consider our approach appropriate and applicable on the clinical setting described in this study.
Further limitations of this study have to be discussed. First, we had to define a cutoff between image quality and length of protocol. This is why, unlike previous investigators, we chose rather basic sequences. The T2-weighted HASTE sequence used in this study had a high slice thickness of 5 mm. This was necessary to achieve sufficient signal-to-noise ratio in short imaging time but could be compensated by the smaller slices of the T1-weighted contrast-enhanced VIBE sequence. Of course, current MRI scanners allow scanning with higher image quality [27]. Yet, every protocol intended for clinical use has to be optimized in terms of what is technically possible and what is clinically necessary. In the given setting, identification of locoregional tumor spread has the highest oncologic priority [2, 3] because axillary lymph node status impacts both therapy and outcome of patients [3]. If a patient shows suspect finding in such an initial examination, further procedures can be scheduled, including axillary surgery (with or without imaging guidance), minimal invasive biopsy, or dedicated axillary MRI [6, 16].
Fig. 5 —Examples of pathologic lymph nodes.
A, T2-weighted HASTE image in 82-year-old woman reveals lymph node (invasive lobular cancer T3, nodal positive, grade 3) with cortical thickening (arrowheads);
B, T1-weighted volume-interpolated breath-hold image shows enlarged lymph node with positive Solbiati index (invasive ductal cancer, T1c, nodal positive, grade 3). Qualitative analysis revealed “inhomogeneous cortex,” “irregular margin,” “absence of hilus sign,” and “rim enhancement” (arrow).
Second, we did not assess mediastinal lymph nodes. Although a significant proportion of lymphatic drainage uses this pathway, its oncologic impact remains controversial [28]. Currently, surgical staging is not routinely performed in mediastinal lymph nodes because therapeutic and prognostic value is not well validated [28]. Accordingly, the majority of oncologic studies focus on axillary lymph nodes. Nevertheless, previous investigations report feasibility of MRI to detect infiltration of mediastinal lymph nodes [29]. Basically, our protocol also covers this compartment and the presented evaluation criteria would be applicable. However, because of the absence of an appropriate standard of reference we did not assess mediastinal lymph nodes.
Future steps in our research include validation of results in a larger study population. Such data should be used to correlate axillary tumor load with imaging findings. Because the percentage of involved lymph nodes is an important prognostic factor, results are likely to add significant prognostic value to MRM investigations [30]. Furthermore, the impact of MRM on the axillary surgical approach should be systematically investigated. This is particularly important because advanced nodal involvement is an important cause of false-negative sentinel node biopsy [31]. Extensive nodal involvement can be easily depicted using our approach, even in levels II and III, which are difficult to image on an ultrasound examination.
In conclusion, using a dedicated whole-body MRI system with multichannel capability, we were able to show clinical feasibility of a fast protocol combining local and locoregional staging of breast cancer using MRM. Cancers could be accurately and reliably classified as nodal positive versus nodal negative. Because the presence of lymph node metastases impacts overall prognosis as well as the therapeutic approach, our results have the potential to optimize management of breast cancer patients. Future studies should evaluate the incremental value of our findings with respect to patient management.
Footnotes
P. A. T. Baltzer and M. Dietzel contributed equally to this study.
Address correspondence to P. A. T. Baltzer ([email protected]).
WEB
This is a Web exclusive article.
References
1.
Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, 2009. CA Cancer J Clin 2009; 59:225–249
2.
Edge S, Byrd D, Carducci M, Wittekind C, eds. TNM classification of malignant tumours, 7th ed. New York, NY: Springer-Verlag, 2009
3.
Michaelson JS, Silverstein M, Sgroi D, et al. The effect of tumor size and lymph node status on breast carcinoma lethality. Cancer 2003; 98:2133 –2143
4.
Fuster D, Duch J, Paredes P, et al. Preoperative staging of large primary breast cancer with [18F] fluorodeoxyglucose positron emission tomography/computed tomography compared with conventional imaging procedures. J Clin Oncol 2008; 26:4746 –4751
5.
Kilbride KE, Lee MC, Nees AV, et al. Axillary staging prior to neoadjuvant chemotherapy for breast cancer: predictors of recurrence. Ann Surg Oncol 2008; 15:3252 –3258
6.
Abe H, Schmidt RA, Kulkarni K, Sennett CA, Mueller JS, Newstead GM. Axillary lymph nodes suspicious for breast cancer metastasis: sampling with US-guided 14-gauge core-needle biopsy—clinical experience in 100 patients. Radiology 2009; 250:41–49
7.
DeMartini W, Lehman C. A review of current evidence-based clinical applications for breast magnetic resonance imaging. Top Magn Reson Imaging 2008; 19:143 –150
8.
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 –3258
9.
Schnall MD, Blume J, Bluemke DA, et al. Diagnostic architectural and dynamic features at breast MR imaging: multicenter study. Radiology 2006; 238:42–53
10.
Lehman CD, Gatsonis C, Kuhl CK, et al. MRI evaluation of the contralateral breast in women with recently diagnosed breast cancer. N Engl J Med 2007; 356:1295 –1303
11.
Sardanelli F. Overview of the role of pre-operative breast MRI in the absence of evidence on patient outcomes. Breast 2010; 19:3 –6
12.
Sardanelli F, Hunink MG, Gilbert FJ, Di Leo G, Krestin GP. Evidence-based radiology: why and how? Eur Radiol 2010; 20:1 –15
13.
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 –1471
14.
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 –312
15.
Mumtaz H, Hall-Craggs MA, Davidson T, et al. Staging of symptomatic primary breast cancer with MR imaging. AJR 1997; 169:417 –424
16.
Murray AD, Staff RT, Redpath TW, et al. Dynamic contrast enhanced MRI of the axilla in women with breast cancer: comparison with pathology of excised nodes. Br J Radiol 2002; 75:220–228
17.
Luciani A, Dao TH, Lapeyre M, et al. Simultaneous bilateral breast and high-resolution axillary MRI of patients with breast cancer: preliminary results. AJR 2004; 182:1059 –1067
18.
Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: The STARD Initiative. Radiology 2003; 226:24–28
19.
Sardanelli F, Di Leo G. Biostatistics for radiologists. New York, NY: Springer-Verlag,2009
20.
Mann RM, Kuhl CK, Kinkel K, Boetes C. Breast MRI: guidelines from the European Society of Breast Imaging. Eur Radiol 2008; 18:1307 –1318
21.
Obuchowski NA, Mazzone PJ, Dachman AH. Bias, underestimation of risk, and loss of statistical power in patient-level analyses of lesion detection. Eur Radiol 2010; 20:584–594
22.
American College of Radiology. Breast imaging reporting and data system (BI-RADS), 4th ed. Reston, VA: American College of Radiology, 2003
23.
Esen G. Ultrasound of superficial lymph nodes. Eur J Radiol 2006; 58:345 –359
24.
Baltzer PA, Yang F, Dietzel M, et al. Sensitivity and specificity of unilateral edema on T2w-TSE sequences in MR-mammography considering 974 histologically verified lesions. Breast J 2010; 16:233 –239
25.
Solbiati L, Cioffi V, Ballarati E. Ultrasonography of the neck. Radiol Clin North Am 1992; 30:941–954
26.
Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. Part 1. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991; 19:403–410
27.
Schmidt GP, Baur-Melnyk A, Haug A, et al. Comprehensive imaging of tumor recurrence in breast cancer patients using whole-body MRI at 1.5 and 3 T compared to FDG-PET-CT. Eur J Radiol 2008; 65:47 –58
28.
Pugliese MS, Stempel MM, Cody HS 3rd, Morrow M, Gemignani ML. Surgical management of the axilla: do intramammary nodes matter? Am J Surg 2009; 198:532 –537
29.
Kinoshita T, Odagiri K, Andoh K, et al. Evaluation of small internal mammary lymph node metastases in breast cancer by MRI. Radiat Med 1999; 17:189–193
30.
Voordeckers M, Vinh-Hung V, Van de Steene J, Lamote J, Storme G. The lymph node ratio as prognostic factor in node-positive breast cancer. Radiother Oncol 2004; 70:225–230
31.
de Kanter AY, Menke-Pluijmers MB, Henzen-Logmans SC, et al. Reasons for failure to identify positive sentinel nodes in breast cancer patients with significant nodal involvement. Eur J Surg Oncol 2006; 32:498 –501
Information & Authors
Information
Published In
Copyright
© American Roentgen Ray Society.
History
Submitted: April 29, 2010
Accepted: August 24, 2010
First published: November 23, 2012
Keywords
Authors
Metrics & Citations
Metrics
Citations
Export Citations
To download the citation to this article, select your reference manager software.