- Email: [email protected]
Diagnostic accuracy of apparent diffusion coefficient value in differentiating metastatic form benign axillary lymph nodes in cancer breast
The Egyptian Journal of Radiology and Nuclear Medicine (2014) 45, 1011–1016
Egyptian Society of Radiology and Nuclear Medicine
The Egyptian Journal of Radiology and Nuclear Medicine www.elsevier.com/locate/ejrnm www.sciencedirect.com
ORIGINAL ARTICLE
Diagnostic accuracy of apparent diffusion coefficient value in differentiating metastatic form benign axillary lymph nodes in cancer breast Ahmed A. Azeem Ismail a b
a,*
, Doaa Ibrahim Hasan a, Hasan Abd-Alshakor
b
Diagnostic Radiology Department, Zagazig University, Egypt General Surgery Department, Zagazig University, Egypt
Received 17 May 2014; accepted 9 June 2014 Available online 4 July 2014
KEYWORDS Axillary lymph node; Diffusion-weighted imaging; Apparent diffusion coefficient (ADC)
Abstract Purpose: To detect axillary lymph node metastasis based on diffusion Weighted MR and apparent diffusion coefficient (ADC) in the known breast cancer cases. Patient and methods: Forty-four patients were included in this study for preoperative MRI staging of the breast cancer and axillary lymph node assessment. The lymph node criteria (long/short-axis ratio, T2WI, DWI and ADC value) were included in the analysis. Images were obtained with diffusion sensitizing gradients of 0 and 750 mm2/s. The ADC was calculated. Results: Thirty-two patients had metastatic axillary lymph nodes and 12 cases had no malignant LN involvement. There was no significant difference between both in S/L ratio, T2WISI (p < 0.140 & p = 0.079, respectively), while statistically significant difference between benign and malignant lymph nodes in both DWI and ADC mean values (p < 0.0001 & p < 0.007, respectively). The optimal ADC cut off value was 6.8 · 10 3 mm2/s for differentiation between benign and malignant lymph nodes with accuracy 96.7%, sensitivity 100%, specificity 87%, PPV 95.4% and NPV 100%. Conclusion: Compared with lymph node size or routine magnetic resonance sequences, DWI and ADC are promising techniques for differentiating metastatic and non metastatic axillary lymph nodes in known breast cancer patients. 2014 Production and hosting by Elsevier B.V. on behalf of Egyptian Society of Radiology and Nuclear Medicine.
1. Introduction
* Corresponding author. Tel.: +20 1114077774. E-mail addresses: [email protected] (A.A. Azeem Ismail), [email protected] (D.I. Hasan), hasanabdalskakor@yahoo. mail.com (H. Abd-Alshakor). Peer review under responsibility of Egyptian Society of Radiology and Nuclear Medicine.
Axillary lymph node involvement is the most common route of breast cancer metastases, and is an essential prognostic factor and an important determinant in the treatment of patients with breast cancer (1–4). Axillary lymph node dissection with histopathological examination of the surgical specimen is regarded as the gold standard for detecting involvement of the lymph nodes (1,2).
0378-603X 2014 Production and hosting by Elsevier B.V. on behalf of Egyptian Society of Radiology and Nuclear Medicine. http://dx.doi.org/10.1016/j.ejrnm.2014.06.003
1012 Axillary dissection is a separate procedure requiring hospitalization, general anesthesia, and a 1- to 2-week period of postoperative drain care (5,6). The effect of axillary lymph node dissection on long-term treatment results in invasive breast cancer is controversial; it is in fact frequently questioned because it is associated with high morbidity rates and neoadjuvant chemotherapy is already administered in aggressive local cancers (7). With magnetic resonance imaging (MRI), computed tomography, and ultrasonography techniques, the preoperative assessment of axillary lymph node status is mainly based on the measurement of nodal dimensions, such as maximum long-axis diameter, short-axis diameter or long/short-axis ratio, but these techniques are somewhat limited. In addition, morphologic criteria (e.g., shape, thickened lobular cortex, displacement and/or absence of fatty hilum), enhancement patterns, and grouping of lymph nodes are further important parameters (8–11). All these criteria remain controversial, and recommendations for differentiation between metastatic and non-metastatic lymph nodes vary widely. Diffusion-weighted imaging (DWI) facilitates the noninvasive characterization of various tissues on the basis of their water diffusion properties. Any architectural changes in the proportion of extracellular to intracellular water protons will alter the diffusion coefficient of the tissue (12–14), thus providing information about the biophysical properties of tissues in vivo. Quantitative measurement of apparent diffusion coefficient (ADC) could reflect the degree of restriction of diffusion of different tissues (15,16). In the present study, we aimed to determine metastatic axillary lymph node involvement by a simple and noninvasive method with focus on the feasibility of DWI technique to detect metastatic lymph nodes in pathologically proven breast cancer patients. 2. Patients and methods 2.1. Patients From July 2011 to April 2013, 48 consecutive patients (age range, 24–70 years) were included with histologically proven breast cancer and pre-operative MRI for staging and detection of axillary LN involvement. This study was approved by ethics committee of the University, and all patients gave written informed consent before being included. Four of these 48 patients were secondarily excluded for the following reasons: pulmonary metastases detected on MRI (= n1), general health unsuitable for surgery (= n1), cancelation of surgery by the patient (= n1), and poor magnetic resonance (MR) image quality related to motion-artifacts (= n1). Some cases had previous US examinations; we did not take these data and other morphological criteria into account or include them in statistical analysis. 2.2. Magnetic resonance imaging protocol and related dynamic data processing Magnetic resonance imaging was performed on a 1.5 Tesla system (Philips, Achivea –Class B-USA) with the patient in the prone position and the breast suspended in 4 channel dedicated breast coil. After obtaining an axial localizer sequence,
A.A. Azeem Ismail et al. axial T1-weighted images [turbo-spin echo (TSE); repetition time (TR), 613 ms; echo time (TE), 9.1 ms; section thickness (ST), 3.5 mm], axial short T1 inversion recovery (STIR) sequences (TE: 9370 ms, TE: 70 ms, ST: 3.5 mm) were acquired [field of view (FOV): 400 · 350 mm]. For dynamic study a standard dose (0.1 mmol/kg) of contrast agent with gadolinium was injected through the antecubital vein in each patient immediately following the acquisition of precontrast series. Following the injection of contrast agent, 15 mL physiological saline flush was injected in each patient. After the contrast injection, DCE-MRI was performed with gradient-recalled sequences (TR, 4.4 ms; TE, 1.6 ms; flip angle, 12; FOV, 350 · 350 mm, 256 · 128; section thickness 1.5 mm with no section gap). Including those acquired during the precontrast period, a total of 7 series each consisting of 80 consecutive images were acquired. In every 51 s six subtracted images (image series achieved by subtracting the precontrast image from each series) were transferred to the mean curve analysis in the work station. After the dynamic study, sagittal late postcontrast fat-saturated T1-weighted TSE images (TR, 532 ms; TE, 4.01 ms; ST, 3.0 mm; FOV: 400 · 400 mm) were acquired in order to recheck and correlate the lesions and axillary lymph nodes. DWI was performed using a diffusion-weighted echo planar imaging (EPI) sequence with spectral spatial fat suppression and parallel imaging (reduction factor = 2); TR/TE = 7000/ 71.5 ms, 3 NEX, matrix = 192 · 192, bandwidth = 1953 Hz/ pixel, FOV = 36 cm, slice thickness = 5 mm, gap = 0. Diffusion gradients were applied in six directions with b = 0 and 750 s/mm2, and the scan time was 160 s. 2.2.1. Measurements on MRI and image analysis The axillary lymph nodes with the short-axis diameter equal to or greater than 5 mm were included in our study. The long-axis (L) and short-axis (S) diameters were measured on the corresponding maximum section of each lymph node on axial T2WI, respectively, and the long/short-axis ratio (L/S) was calculated. The signal intensity on T2WI and DWI (ST2WI, SDWI) of each selected lymph node was reported. The ADC maps were calculated on a pixel-by-pixel basis. We used the corresponding central slice of each lymph node on the ADC map, a round or elliptical region of interest (ROI) covering about 3/4 of the entire area of the selected lymph node attempting to avoid inclusion of the margins; while T2W imaging was used as the anatomical reference. The apparent diffusion coefficient (ADC) value was calculated for lymph nodes. The area of the ROI in the selected axillary lymph nodes was 25–150 mm2, according to their different sizes. For ADC measurement, ROIs were placed in the LNs; all the values were averaged from three-time measurements and expressed as the mean ± standard deviation. FNAB or/and true cut Biopsy was/were performed in dominant lymph node or nodes visualized in images in all cases. The dominant lymph nodes were identified according to the following criteria: asymmetric cortical thickening, cortical irregularity, cortical heterogeneity, hilar deformation and loss of fatty sinus, and loss of ovality (rounding). 2.2.2. Statistical analysis Statistical analyses were performed by using Statistical Product and Service Solutions (SPSS, Version 15.0.1, Inc., Chicago,
Diagnostic accuracy in differentiating metastatic form benign axillary lymph nodes
1013
IL). Differences of mean L/S ratio and the mean ADC values between the axillary malignant and non malignant lymph nodes were compared using independent t test. Fisher’s exact test was also used to assess the relationship between both groups’ signal intensity appearance on T2WI (SIT2WI) and signal intensity on DWI (SIDWI). A p value <0.05 was considered statistically significant. 3. Results In the present study the patients’ age ranged from 24 to 70 years (mean 54 years). Among the 44 cases, dominant lymph nodes were evaluated consisting of 32 (72.7%) metastatic and 12 (27.2%) non malignant lymph nodes. Of thirtytwo metastatic cases, 8 (25%) of the cases were bilateral and unilateral in 24 (75%) of the cases, while non malignant lymph nodes were bilateral in 3 cases and unilateral in 9 cases. The pathological results for 32 malignant breast cases were as follows: 21 invasive ductal carcinoma, 10 diagnosed as lobular carcinoma, and 1 medullary carcinoma. Regarding the 12 benign cases 5 of them were inflammatory and the other 7 cases were non specific enlargement. The mean values of L/S ratio for the axillary benign (Fig. 1) and metastatic (Figs. 2 and 3) lymph nodes were 2.1 ± 0.12 mm, 1.6 ± 0.13 mm, respectively (table 1). Depending on the S/L ratio between the malignant and benign LNs there was no statistically significant difference. Depending on the signal intensity criteria on T2WI and signal-intensity of the LNs in DWI, there was no significant difference between the signal intensity of benign and metastatic L.Ns on T2 WI (p = 0.079), however, there was a statistically significant difference between non malignant and malignant LNs (p < 0.001) in DWI (Figs. 1 and 2, respectively). The mean value of the ADC for the axillary metastatic LNs (Figs. 2 and 3) is lower (0.79 ± 0.23 · 10 3 mm2/s) than mean AD values for benign lymph nodes (1. 42 ± 0.57 · 10 3 mm2/ s) (fig. 1). The difference between the two groups was statistically significant (Table 1). In this study, statistical analysis of our results showed that optimal ADC cut off value for differentiation between benign and malignant lymph nodes is 6.8 · 10 3 mm2/s with accuracy 96.7%, sensitivity 100%, specificity 87%, PPV 95.4% and NPV 100%. 4. Discussion Preoperative detection of axillary lymph node metastases is desirable because in many settings it may allow a surgeon to bypass sentinel lymph node biopsy (SLNB) and proceed directly to complete axillary lymph node dissection (ALND), sparing patients an additional operative procedure and expediting their progression through their cancer treatment course (17). DCE-MRI is now performed routinely in patients newly diagnosed with breast cancer at many institutions in preoperative evaluation as it had advantages over sonographic evaluation of the axilla which include the ability to compare directly axillary lymph nodes in question with the contralateral axilla and decreased dependence on operator experience. Unlike the evaluation of primary tumors on MRI, enhancement kinetics features have not proven useful for discrimination of
Fig. 1 A 25 year old female patient with malignant left breast mass and fibroadenosis of the both breasts. (A) sagittal T2FS image demonstrates the enlarged left axillary LNs (15 · 20, S/L diameter), displaying rounded shape, thickened cortex and no central fatty signal. (B and C) DWI & color coded ADC map showing hypointense not restricted diffusion, ADC value of prominent left axillary lymph node is measured 1.63 · 10 3 mm2/s (arrow).
abnormal nodes because many normal nodes exhibit kinetics features typical of breast malignancy (e.g., delayed washout) (18). As a simple, fast, non-invasive and non-contrast tool, diffusion weighted imaging (DWI) has been the most potential representative of fMRI in LN mapping in recent years (13,15). We found that no statistically significant difference between malignant and benign LNs depending on the S/L ratio,
1014
Fig. 2 A 42 year old female patient with malignant RT breast mass. (A) T2FS image demonstrates the enlarged RT axillary LNs (16 · 24 mm, S/L diameter), displaying oval shape, irregular thickened cortex and no fatty hilum. (B and C) DWI & color coded ADC map showing heterogeneous hyperintense signal and restricted diffusion, ADC value of prominent RT axillary lymph node is measured 0.78 · 10 3 mm2/s (arrow).
similarly Stets et al. (3) and Michel et al. (4), stated that inaccurate differentiation between the metastatic and benign LNs depending on the LN size and this criteria gave is low sensitivity in the diagnosis of metastatic nodes (53–62%). Moreover, Obwegeser et al. (19) reported that, in histological analysis
A.A. Azeem Ismail et al.
Fig. 3 A 48 year old female patient with malignant RT breast mass and breast edema with skin thickening. (A) Axial T2FS image demonstrates the enlarged RT axillary LNs (12 · 16 mm, S/ L diameter), displaying oval shape, irregular thickened cortex and effaced faint fatty hilum. (B and C) DWI & color coded ADC map showing heterogeneous hyperintense signal and restricted diffusion, ADC value of the RT axillary lymph node is measured 0.81 · 10 3 mm2/s (arrow).
of 1249 axillary lymph nodes in 71 patients with breast cancer, 13.7% of metastatic nodes were 5–9 mm in short-axis diameter. Also we found, there was no statistical difference between axillary non metastatic and metastatic lymph nodes, regarding their signal intensity on T2WI, in agreement with the results obtained by Som (20). In the present study the high signal intensity nodes on DWI were significantly higher than those on T2WI, the malignant
Diagnostic accuracy in differentiating metastatic form benign axillary lymph nodes
1015
Table 1 Comparison of the results of the mean value for the lymph node S/L ratio, T2WI, DWI signal intensity and mean ADCs of the metastatic and non malignant. MRI criteria
Malignant (n = 32, %)
Non malignant (n = 12, %)
p
The mean S/L ratio SIT2WI Hypointense Hyperintense Heterogeneous SI-DWI (B = 750) Hyperintense Hypointense Mean ADC (x10 3 mm2/s)
1.6 ± 0.13
2.1 ± 0.12
p < 0.140b
2/32, 6.2 28/32, 87.5 2/32, 8.6
6/12, 50 5/12, 41.6 1/12, 8.3
p = 0.079a
31/32, 96.8 1/32, 3.1 0.79 ± 0.23
1/12, 8.3 11/12, 91.6 1.42 ± 0.57
p < 0.0001a
p < 0.007
b
L/S = long/short-axis ratio, SIT2WI = signal Intensity on T2WI, SI-DWI = signal intensity diffusion-weighted imaging ADC = mean apparent diffusion coefficient p value <0.05 was considered statistically significant. a Fisher’s exact test. b Independent-t test.
L.Ns were shown with marked high signal intensities on DWI that made it easy to identify. Similar to Wang et al. (21) findings in their study on rabbit models, stated that statistical significant difference comparing the signal intensity on T2 WI and DWI between inflammatory and malignant L.Ns groups. On the contrary Vandecaveye et al. (22), found that the visual assessment of DWI to differentiate the axillary inflammatory from metastatic lymph nodes was rather a difficult task, as in some cases both groups’ lymph nodes showed high signal intensities on DWI. The ADC values calculated from DWI can provide quantitative analysis of microscopic water diffusivity in the target tissues and can completely exclude the T2 shine-through effect. Among our patients lower ADC values were always observed in metastatic lymph nodes and there was a statistically significant difference in the ADC mean values between metastatic and non- metastatic axillary L.Ns. We use the optimal 6.8 · 10 3 mm2/s ADC cut off value for differentiation between benign and malignant lymph nodes and this optimal cut off recorded high accuracy 96.7%, sensitivity 100%, specificity 87%, PPV 95.4% and NPV 100%. Near to our findings Kim et al. (23) reported that ADC value can help to differentiate malignant from benign LNs in predicting nodal metastasis in head and neck cancer. Using threshold value of 0.85 · 10 3/mm2 s 1, the accuracy, sensitivity and specificity of ADC value for differential diagnosis are 91.0%, 91.3% and 91.1%, respectively. They stated the reasons for a more powerful diagnostic ability of DWI than other size-dependent imaging methods because, normal size LNs with micrometastasis are evaluated correctly and another is hyper plastic benign nodes are appropriately excluded. Also, Isßıl et al. (24) found that, the ADC values of malignant lymph nodes were significantly lower than the benign ones (p = 0.001), but they used 1.22 · 10 3 mm2/s as the cut-off ADC value, and reported a sensitivity of 75.6% and a specificity of 71.1%. On the other hand Wang et al., 17 had reported sensitivity, specificity and accuracy, which were 86.2%, 79.3%, 81.2%, when the rADC value (ADC lesion/ ADC reference site) was chosen as 0.80. Further studies with a large number of subjects will increase confidence in the results of DWI and guide physicians involved in diagnosis, treatment and follow up of breast cancer patients.
In conclusion: DWI and ADC values had high accuracy for differentiating axillary non metastatic (benign) from metastatic lymph nodes. Conflict of interest None declared. References (1) Orel SG, Schnall MD. MR imaging of the breast for the detection, diagnosis, and staging of breast cancer. Radiology 2001;220:13–30. (2) Morrow M. Role of axillary dissection in breast cancer management. Ann Surg Oncol 1996;3:233–4. (3) Stets C, Brandt S, Wallis F, Buchmann J, Gilbert FJ, HeywangKo¨brunner SH. Axillary lymph node metastases: a statistical analysis of various parameters in MRI with USPIO. J Magn Reson Imaging 2002;16:60–8. (4) Michel S, Keller T, Fro¨hlich J, Fink D, Caduff R, Seifert B, et al. Preoperative breast cancer staging: MR imaging of the axilla with ultrasmall superparamagnetic iron oxide enhancement. Radiology 2002;225:527–36. (5) Cady B. The need to reexamine axillary lymph node dissection in invasive breast cancer. Cancer 1994;73:505–8. (6) Ruffin W, Stacey C, Younger J, Hoover H. Rationale for routine axillary dissection in carcinoma of the breast. J Am Coll Surg 1995;180:245–51. (7) Recht A, Houlihan M. Axillary lymph nodes and breast cancer: a review. Cancer 1995;76(9):1491–512. (8) Shetty M, Carpenter W. Sonographic evaluation of isolated abnormal axillary lymph nodes identified on mammograms. J Ultrasound Med 2004;23:63–71. (9) Krishnamurthy S, Sneige N, Bedi D, Edieken BS, Fornage B, Kuerer H, et al. Role of ultrasound-guided fine-needle aspiration of indeterminate and suspicious axillary lymph nodes in the initial staging of breast carcinoma. Cancer 2002;95:982–8. (10) Steinkamp HJ, Cornehl M, Hosten N, Pegios W, Vogl T, Felix R. Cervical lymphadenopathy: ratio of long- to short-axis diameter as a predictor of malignancy. Br J Radiol 1995;68:266–70. (11) Roth Y, Tichler T, Kostenich G, Ruiz-Cabello J, Maier SE, Cohen JS, et al. High-b-value diffusion-weighted MR imaging for pretreatment prediction and early monitoring of tumor response to therapy in mice. Radiology 2004;232:685–92.
1016 (12) Squillaci E, Manenti G, Cova M, Di Roma M, Miano R, Palmieri G, et al. Correlation of diffusion-weighted MR imaging with cellularity of renal tumours. Anticancer Res 2004;24:4175–9. (13) de Bondt R, Hoeberigs M, Nelemans P, Deserno W, Peutz K, Kremer B. Diagnostic accuracy and additional value of diffusionweighted imaging for discrimination of malignant cervical lymph nodes in head and neck squamous cell carcinoma. Neuroradiology 2009;51:183–92. (14) Ono K, Ochiai R, Yoshida T, Kitagawa M, Omagari J, Kobayashi H, et al. Comparison of diffusion-weighted MRI and 2-[fluorine-18]-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) for detecting primary colorectal cancer and regional lymph node metastases. J Magn Reson Imaging 2009;29:336–40. (15) Kim JK, Kim KA, Park BW, Kim N, Cho KS. Feasibility of diffusion-weighted imaging in the differentiation of metastatic from nonmetastatic lymph nodes: early experience. J Magn Reson Imaging 2008;28:714–9. (16) Park SO, Kim JK, Kim KA, Park BW, Kim N, Cho G, et al. Relative apparent diffusion coefficient: determination of reference site and validation of benefit for detecting metastatic lymph nodes in uterine cervical cancer. J Magn Reson Imaging 2009;29:383–90. (17) Habib R, Savannah C, Sara H, Constance D. Imaging axillary lymph nodes in patients with newly diagnosed breast cancer. Curr Probl Diagn Radiol 2012;41:149–58. (18) Mortellaro VE, Marshall J, Singer L, Hochwald SN, Chang M, Copeland EM, et al. Magnetic resonance imaging for axillary
A.A. Azeem Ismail et al.
(19)
(20)
(21)
(22)
(23) (24)
staging in patients with breast cancer. J Magn Reson Imaging 2009;30:309–12. Obwegeser R, Lorenz K, Hohlagschwandtner M, Czerwenka K, Schneider B, Kubista E. Axillary lymph nodes in breast cancer: is size related to metastatic involvement? World J Surg 2000;24:546–50. Som PM. Detection of metastasis in cervical lymph nodes: CT and MR criteria and differential diagnosis. AJR Am J Roentgenol 1992;158:961–9. Wang J, Liao Q, Zhang Y, Yu C, Bai R, Sun H. Differential diagnosis of axillary inflammatory and metastatic lymph nodes in rabbit models by using diffusion-weighted imaging: compared with conventional magnetic resonance imaging. Korean J Radiol 2012;13:458–66. Vandecaveye V, De Keyzer F, Vander Poorten V, Dirix P, Verbeken E, Nuyts S. Head and neck squamous cell carcinoma: value of diffusion-weighted MR imaging for nodal staging. Radiology 2009;251:134–46. Kim J, Jang Y, Cho G. Multidisciplinary functional MR imaging for prostate cancer. Korean J Radiol 2009;10:535–51. Isßıl B, S ß ebnem O, Teoman C. The role of diffusion weighted imaging in the assessment of axillary lymph nodes. J Breast Health 2013;9:76–81.
Egyptian Society of Radiology and Nuclear Medicine
The Egyptian Journal of Radiology and Nuclear Medicine www.elsevier.com/locate/ejrnm www.sciencedirect.com
ORIGINAL ARTICLE
Diagnostic accuracy of apparent diffusion coefficient value in differentiating metastatic form benign axillary lymph nodes in cancer breast Ahmed A. Azeem Ismail a b
a,*
, Doaa Ibrahim Hasan a, Hasan Abd-Alshakor
b
Diagnostic Radiology Department, Zagazig University, Egypt General Surgery Department, Zagazig University, Egypt
Received 17 May 2014; accepted 9 June 2014 Available online 4 July 2014
KEYWORDS Axillary lymph node; Diffusion-weighted imaging; Apparent diffusion coefficient (ADC)
Abstract Purpose: To detect axillary lymph node metastasis based on diffusion Weighted MR and apparent diffusion coefficient (ADC) in the known breast cancer cases. Patient and methods: Forty-four patients were included in this study for preoperative MRI staging of the breast cancer and axillary lymph node assessment. The lymph node criteria (long/short-axis ratio, T2WI, DWI and ADC value) were included in the analysis. Images were obtained with diffusion sensitizing gradients of 0 and 750 mm2/s. The ADC was calculated. Results: Thirty-two patients had metastatic axillary lymph nodes and 12 cases had no malignant LN involvement. There was no significant difference between both in S/L ratio, T2WISI (p < 0.140 & p = 0.079, respectively), while statistically significant difference between benign and malignant lymph nodes in both DWI and ADC mean values (p < 0.0001 & p < 0.007, respectively). The optimal ADC cut off value was 6.8 · 10 3 mm2/s for differentiation between benign and malignant lymph nodes with accuracy 96.7%, sensitivity 100%, specificity 87%, PPV 95.4% and NPV 100%. Conclusion: Compared with lymph node size or routine magnetic resonance sequences, DWI and ADC are promising techniques for differentiating metastatic and non metastatic axillary lymph nodes in known breast cancer patients. 2014 Production and hosting by Elsevier B.V. on behalf of Egyptian Society of Radiology and Nuclear Medicine.
1. Introduction
* Corresponding author. Tel.: +20 1114077774. E-mail addresses: [email protected] (A.A. Azeem Ismail), [email protected] (D.I. Hasan), hasanabdalskakor@yahoo. mail.com (H. Abd-Alshakor). Peer review under responsibility of Egyptian Society of Radiology and Nuclear Medicine.
Axillary lymph node involvement is the most common route of breast cancer metastases, and is an essential prognostic factor and an important determinant in the treatment of patients with breast cancer (1–4). Axillary lymph node dissection with histopathological examination of the surgical specimen is regarded as the gold standard for detecting involvement of the lymph nodes (1,2).
0378-603X 2014 Production and hosting by Elsevier B.V. on behalf of Egyptian Society of Radiology and Nuclear Medicine. http://dx.doi.org/10.1016/j.ejrnm.2014.06.003
1012 Axillary dissection is a separate procedure requiring hospitalization, general anesthesia, and a 1- to 2-week period of postoperative drain care (5,6). The effect of axillary lymph node dissection on long-term treatment results in invasive breast cancer is controversial; it is in fact frequently questioned because it is associated with high morbidity rates and neoadjuvant chemotherapy is already administered in aggressive local cancers (7). With magnetic resonance imaging (MRI), computed tomography, and ultrasonography techniques, the preoperative assessment of axillary lymph node status is mainly based on the measurement of nodal dimensions, such as maximum long-axis diameter, short-axis diameter or long/short-axis ratio, but these techniques are somewhat limited. In addition, morphologic criteria (e.g., shape, thickened lobular cortex, displacement and/or absence of fatty hilum), enhancement patterns, and grouping of lymph nodes are further important parameters (8–11). All these criteria remain controversial, and recommendations for differentiation between metastatic and non-metastatic lymph nodes vary widely. Diffusion-weighted imaging (DWI) facilitates the noninvasive characterization of various tissues on the basis of their water diffusion properties. Any architectural changes in the proportion of extracellular to intracellular water protons will alter the diffusion coefficient of the tissue (12–14), thus providing information about the biophysical properties of tissues in vivo. Quantitative measurement of apparent diffusion coefficient (ADC) could reflect the degree of restriction of diffusion of different tissues (15,16). In the present study, we aimed to determine metastatic axillary lymph node involvement by a simple and noninvasive method with focus on the feasibility of DWI technique to detect metastatic lymph nodes in pathologically proven breast cancer patients. 2. Patients and methods 2.1. Patients From July 2011 to April 2013, 48 consecutive patients (age range, 24–70 years) were included with histologically proven breast cancer and pre-operative MRI for staging and detection of axillary LN involvement. This study was approved by ethics committee of the University, and all patients gave written informed consent before being included. Four of these 48 patients were secondarily excluded for the following reasons: pulmonary metastases detected on MRI (= n1), general health unsuitable for surgery (= n1), cancelation of surgery by the patient (= n1), and poor magnetic resonance (MR) image quality related to motion-artifacts (= n1). Some cases had previous US examinations; we did not take these data and other morphological criteria into account or include them in statistical analysis. 2.2. Magnetic resonance imaging protocol and related dynamic data processing Magnetic resonance imaging was performed on a 1.5 Tesla system (Philips, Achivea –Class B-USA) with the patient in the prone position and the breast suspended in 4 channel dedicated breast coil. After obtaining an axial localizer sequence,
A.A. Azeem Ismail et al. axial T1-weighted images [turbo-spin echo (TSE); repetition time (TR), 613 ms; echo time (TE), 9.1 ms; section thickness (ST), 3.5 mm], axial short T1 inversion recovery (STIR) sequences (TE: 9370 ms, TE: 70 ms, ST: 3.5 mm) were acquired [field of view (FOV): 400 · 350 mm]. For dynamic study a standard dose (0.1 mmol/kg) of contrast agent with gadolinium was injected through the antecubital vein in each patient immediately following the acquisition of precontrast series. Following the injection of contrast agent, 15 mL physiological saline flush was injected in each patient. After the contrast injection, DCE-MRI was performed with gradient-recalled sequences (TR, 4.4 ms; TE, 1.6 ms; flip angle, 12; FOV, 350 · 350 mm, 256 · 128; section thickness 1.5 mm with no section gap). Including those acquired during the precontrast period, a total of 7 series each consisting of 80 consecutive images were acquired. In every 51 s six subtracted images (image series achieved by subtracting the precontrast image from each series) were transferred to the mean curve analysis in the work station. After the dynamic study, sagittal late postcontrast fat-saturated T1-weighted TSE images (TR, 532 ms; TE, 4.01 ms; ST, 3.0 mm; FOV: 400 · 400 mm) were acquired in order to recheck and correlate the lesions and axillary lymph nodes. DWI was performed using a diffusion-weighted echo planar imaging (EPI) sequence with spectral spatial fat suppression and parallel imaging (reduction factor = 2); TR/TE = 7000/ 71.5 ms, 3 NEX, matrix = 192 · 192, bandwidth = 1953 Hz/ pixel, FOV = 36 cm, slice thickness = 5 mm, gap = 0. Diffusion gradients were applied in six directions with b = 0 and 750 s/mm2, and the scan time was 160 s. 2.2.1. Measurements on MRI and image analysis The axillary lymph nodes with the short-axis diameter equal to or greater than 5 mm were included in our study. The long-axis (L) and short-axis (S) diameters were measured on the corresponding maximum section of each lymph node on axial T2WI, respectively, and the long/short-axis ratio (L/S) was calculated. The signal intensity on T2WI and DWI (ST2WI, SDWI) of each selected lymph node was reported. The ADC maps were calculated on a pixel-by-pixel basis. We used the corresponding central slice of each lymph node on the ADC map, a round or elliptical region of interest (ROI) covering about 3/4 of the entire area of the selected lymph node attempting to avoid inclusion of the margins; while T2W imaging was used as the anatomical reference. The apparent diffusion coefficient (ADC) value was calculated for lymph nodes. The area of the ROI in the selected axillary lymph nodes was 25–150 mm2, according to their different sizes. For ADC measurement, ROIs were placed in the LNs; all the values were averaged from three-time measurements and expressed as the mean ± standard deviation. FNAB or/and true cut Biopsy was/were performed in dominant lymph node or nodes visualized in images in all cases. The dominant lymph nodes were identified according to the following criteria: asymmetric cortical thickening, cortical irregularity, cortical heterogeneity, hilar deformation and loss of fatty sinus, and loss of ovality (rounding). 2.2.2. Statistical analysis Statistical analyses were performed by using Statistical Product and Service Solutions (SPSS, Version 15.0.1, Inc., Chicago,
Diagnostic accuracy in differentiating metastatic form benign axillary lymph nodes
1013
IL). Differences of mean L/S ratio and the mean ADC values between the axillary malignant and non malignant lymph nodes were compared using independent t test. Fisher’s exact test was also used to assess the relationship between both groups’ signal intensity appearance on T2WI (SIT2WI) and signal intensity on DWI (SIDWI). A p value <0.05 was considered statistically significant. 3. Results In the present study the patients’ age ranged from 24 to 70 years (mean 54 years). Among the 44 cases, dominant lymph nodes were evaluated consisting of 32 (72.7%) metastatic and 12 (27.2%) non malignant lymph nodes. Of thirtytwo metastatic cases, 8 (25%) of the cases were bilateral and unilateral in 24 (75%) of the cases, while non malignant lymph nodes were bilateral in 3 cases and unilateral in 9 cases. The pathological results for 32 malignant breast cases were as follows: 21 invasive ductal carcinoma, 10 diagnosed as lobular carcinoma, and 1 medullary carcinoma. Regarding the 12 benign cases 5 of them were inflammatory and the other 7 cases were non specific enlargement. The mean values of L/S ratio for the axillary benign (Fig. 1) and metastatic (Figs. 2 and 3) lymph nodes were 2.1 ± 0.12 mm, 1.6 ± 0.13 mm, respectively (table 1). Depending on the S/L ratio between the malignant and benign LNs there was no statistically significant difference. Depending on the signal intensity criteria on T2WI and signal-intensity of the LNs in DWI, there was no significant difference between the signal intensity of benign and metastatic L.Ns on T2 WI (p = 0.079), however, there was a statistically significant difference between non malignant and malignant LNs (p < 0.001) in DWI (Figs. 1 and 2, respectively). The mean value of the ADC for the axillary metastatic LNs (Figs. 2 and 3) is lower (0.79 ± 0.23 · 10 3 mm2/s) than mean AD values for benign lymph nodes (1. 42 ± 0.57 · 10 3 mm2/ s) (fig. 1). The difference between the two groups was statistically significant (Table 1). In this study, statistical analysis of our results showed that optimal ADC cut off value for differentiation between benign and malignant lymph nodes is 6.8 · 10 3 mm2/s with accuracy 96.7%, sensitivity 100%, specificity 87%, PPV 95.4% and NPV 100%. 4. Discussion Preoperative detection of axillary lymph node metastases is desirable because in many settings it may allow a surgeon to bypass sentinel lymph node biopsy (SLNB) and proceed directly to complete axillary lymph node dissection (ALND), sparing patients an additional operative procedure and expediting their progression through their cancer treatment course (17). DCE-MRI is now performed routinely in patients newly diagnosed with breast cancer at many institutions in preoperative evaluation as it had advantages over sonographic evaluation of the axilla which include the ability to compare directly axillary lymph nodes in question with the contralateral axilla and decreased dependence on operator experience. Unlike the evaluation of primary tumors on MRI, enhancement kinetics features have not proven useful for discrimination of
Fig. 1 A 25 year old female patient with malignant left breast mass and fibroadenosis of the both breasts. (A) sagittal T2FS image demonstrates the enlarged left axillary LNs (15 · 20, S/L diameter), displaying rounded shape, thickened cortex and no central fatty signal. (B and C) DWI & color coded ADC map showing hypointense not restricted diffusion, ADC value of prominent left axillary lymph node is measured 1.63 · 10 3 mm2/s (arrow).
abnormal nodes because many normal nodes exhibit kinetics features typical of breast malignancy (e.g., delayed washout) (18). As a simple, fast, non-invasive and non-contrast tool, diffusion weighted imaging (DWI) has been the most potential representative of fMRI in LN mapping in recent years (13,15). We found that no statistically significant difference between malignant and benign LNs depending on the S/L ratio,
1014
Fig. 2 A 42 year old female patient with malignant RT breast mass. (A) T2FS image demonstrates the enlarged RT axillary LNs (16 · 24 mm, S/L diameter), displaying oval shape, irregular thickened cortex and no fatty hilum. (B and C) DWI & color coded ADC map showing heterogeneous hyperintense signal and restricted diffusion, ADC value of prominent RT axillary lymph node is measured 0.78 · 10 3 mm2/s (arrow).
similarly Stets et al. (3) and Michel et al. (4), stated that inaccurate differentiation between the metastatic and benign LNs depending on the LN size and this criteria gave is low sensitivity in the diagnosis of metastatic nodes (53–62%). Moreover, Obwegeser et al. (19) reported that, in histological analysis
A.A. Azeem Ismail et al.
Fig. 3 A 48 year old female patient with malignant RT breast mass and breast edema with skin thickening. (A) Axial T2FS image demonstrates the enlarged RT axillary LNs (12 · 16 mm, S/ L diameter), displaying oval shape, irregular thickened cortex and effaced faint fatty hilum. (B and C) DWI & color coded ADC map showing heterogeneous hyperintense signal and restricted diffusion, ADC value of the RT axillary lymph node is measured 0.81 · 10 3 mm2/s (arrow).
of 1249 axillary lymph nodes in 71 patients with breast cancer, 13.7% of metastatic nodes were 5–9 mm in short-axis diameter. Also we found, there was no statistical difference between axillary non metastatic and metastatic lymph nodes, regarding their signal intensity on T2WI, in agreement with the results obtained by Som (20). In the present study the high signal intensity nodes on DWI were significantly higher than those on T2WI, the malignant
Diagnostic accuracy in differentiating metastatic form benign axillary lymph nodes
1015
Table 1 Comparison of the results of the mean value for the lymph node S/L ratio, T2WI, DWI signal intensity and mean ADCs of the metastatic and non malignant. MRI criteria
Malignant (n = 32, %)
Non malignant (n = 12, %)
p
The mean S/L ratio SIT2WI Hypointense Hyperintense Heterogeneous SI-DWI (B = 750) Hyperintense Hypointense Mean ADC (x10 3 mm2/s)
1.6 ± 0.13
2.1 ± 0.12
p < 0.140b
2/32, 6.2 28/32, 87.5 2/32, 8.6
6/12, 50 5/12, 41.6 1/12, 8.3
p = 0.079a
31/32, 96.8 1/32, 3.1 0.79 ± 0.23
1/12, 8.3 11/12, 91.6 1.42 ± 0.57
p < 0.0001a
p < 0.007
b
L/S = long/short-axis ratio, SIT2WI = signal Intensity on T2WI, SI-DWI = signal intensity diffusion-weighted imaging ADC = mean apparent diffusion coefficient p value <0.05 was considered statistically significant. a Fisher’s exact test. b Independent-t test.
L.Ns were shown with marked high signal intensities on DWI that made it easy to identify. Similar to Wang et al. (21) findings in their study on rabbit models, stated that statistical significant difference comparing the signal intensity on T2 WI and DWI between inflammatory and malignant L.Ns groups. On the contrary Vandecaveye et al. (22), found that the visual assessment of DWI to differentiate the axillary inflammatory from metastatic lymph nodes was rather a difficult task, as in some cases both groups’ lymph nodes showed high signal intensities on DWI. The ADC values calculated from DWI can provide quantitative analysis of microscopic water diffusivity in the target tissues and can completely exclude the T2 shine-through effect. Among our patients lower ADC values were always observed in metastatic lymph nodes and there was a statistically significant difference in the ADC mean values between metastatic and non- metastatic axillary L.Ns. We use the optimal 6.8 · 10 3 mm2/s ADC cut off value for differentiation between benign and malignant lymph nodes and this optimal cut off recorded high accuracy 96.7%, sensitivity 100%, specificity 87%, PPV 95.4% and NPV 100%. Near to our findings Kim et al. (23) reported that ADC value can help to differentiate malignant from benign LNs in predicting nodal metastasis in head and neck cancer. Using threshold value of 0.85 · 10 3/mm2 s 1, the accuracy, sensitivity and specificity of ADC value for differential diagnosis are 91.0%, 91.3% and 91.1%, respectively. They stated the reasons for a more powerful diagnostic ability of DWI than other size-dependent imaging methods because, normal size LNs with micrometastasis are evaluated correctly and another is hyper plastic benign nodes are appropriately excluded. Also, Isßıl et al. (24) found that, the ADC values of malignant lymph nodes were significantly lower than the benign ones (p = 0.001), but they used 1.22 · 10 3 mm2/s as the cut-off ADC value, and reported a sensitivity of 75.6% and a specificity of 71.1%. On the other hand Wang et al., 17 had reported sensitivity, specificity and accuracy, which were 86.2%, 79.3%, 81.2%, when the rADC value (ADC lesion/ ADC reference site) was chosen as 0.80. Further studies with a large number of subjects will increase confidence in the results of DWI and guide physicians involved in diagnosis, treatment and follow up of breast cancer patients.
In conclusion: DWI and ADC values had high accuracy for differentiating axillary non metastatic (benign) from metastatic lymph nodes. Conflict of interest None declared. References (1) Orel SG, Schnall MD. MR imaging of the breast for the detection, diagnosis, and staging of breast cancer. Radiology 2001;220:13–30. (2) Morrow M. Role of axillary dissection in breast cancer management. Ann Surg Oncol 1996;3:233–4. (3) Stets C, Brandt S, Wallis F, Buchmann J, Gilbert FJ, HeywangKo¨brunner SH. Axillary lymph node metastases: a statistical analysis of various parameters in MRI with USPIO. J Magn Reson Imaging 2002;16:60–8. (4) Michel S, Keller T, Fro¨hlich J, Fink D, Caduff R, Seifert B, et al. Preoperative breast cancer staging: MR imaging of the axilla with ultrasmall superparamagnetic iron oxide enhancement. Radiology 2002;225:527–36. (5) Cady B. The need to reexamine axillary lymph node dissection in invasive breast cancer. Cancer 1994;73:505–8. (6) Ruffin W, Stacey C, Younger J, Hoover H. Rationale for routine axillary dissection in carcinoma of the breast. J Am Coll Surg 1995;180:245–51. (7) Recht A, Houlihan M. Axillary lymph nodes and breast cancer: a review. Cancer 1995;76(9):1491–512. (8) Shetty M, Carpenter W. Sonographic evaluation of isolated abnormal axillary lymph nodes identified on mammograms. J Ultrasound Med 2004;23:63–71. (9) Krishnamurthy S, Sneige N, Bedi D, Edieken BS, Fornage B, Kuerer H, et al. Role of ultrasound-guided fine-needle aspiration of indeterminate and suspicious axillary lymph nodes in the initial staging of breast carcinoma. Cancer 2002;95:982–8. (10) Steinkamp HJ, Cornehl M, Hosten N, Pegios W, Vogl T, Felix R. Cervical lymphadenopathy: ratio of long- to short-axis diameter as a predictor of malignancy. Br J Radiol 1995;68:266–70. (11) Roth Y, Tichler T, Kostenich G, Ruiz-Cabello J, Maier SE, Cohen JS, et al. High-b-value diffusion-weighted MR imaging for pretreatment prediction and early monitoring of tumor response to therapy in mice. Radiology 2004;232:685–92.
1016 (12) Squillaci E, Manenti G, Cova M, Di Roma M, Miano R, Palmieri G, et al. Correlation of diffusion-weighted MR imaging with cellularity of renal tumours. Anticancer Res 2004;24:4175–9. (13) de Bondt R, Hoeberigs M, Nelemans P, Deserno W, Peutz K, Kremer B. Diagnostic accuracy and additional value of diffusionweighted imaging for discrimination of malignant cervical lymph nodes in head and neck squamous cell carcinoma. Neuroradiology 2009;51:183–92. (14) Ono K, Ochiai R, Yoshida T, Kitagawa M, Omagari J, Kobayashi H, et al. Comparison of diffusion-weighted MRI and 2-[fluorine-18]-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) for detecting primary colorectal cancer and regional lymph node metastases. J Magn Reson Imaging 2009;29:336–40. (15) Kim JK, Kim KA, Park BW, Kim N, Cho KS. Feasibility of diffusion-weighted imaging in the differentiation of metastatic from nonmetastatic lymph nodes: early experience. J Magn Reson Imaging 2008;28:714–9. (16) Park SO, Kim JK, Kim KA, Park BW, Kim N, Cho G, et al. Relative apparent diffusion coefficient: determination of reference site and validation of benefit for detecting metastatic lymph nodes in uterine cervical cancer. J Magn Reson Imaging 2009;29:383–90. (17) Habib R, Savannah C, Sara H, Constance D. Imaging axillary lymph nodes in patients with newly diagnosed breast cancer. Curr Probl Diagn Radiol 2012;41:149–58. (18) Mortellaro VE, Marshall J, Singer L, Hochwald SN, Chang M, Copeland EM, et al. Magnetic resonance imaging for axillary
A.A. Azeem Ismail et al.
(19)
(20)
(21)
(22)
(23) (24)
staging in patients with breast cancer. J Magn Reson Imaging 2009;30:309–12. Obwegeser R, Lorenz K, Hohlagschwandtner M, Czerwenka K, Schneider B, Kubista E. Axillary lymph nodes in breast cancer: is size related to metastatic involvement? World J Surg 2000;24:546–50. Som PM. Detection of metastasis in cervical lymph nodes: CT and MR criteria and differential diagnosis. AJR Am J Roentgenol 1992;158:961–9. Wang J, Liao Q, Zhang Y, Yu C, Bai R, Sun H. Differential diagnosis of axillary inflammatory and metastatic lymph nodes in rabbit models by using diffusion-weighted imaging: compared with conventional magnetic resonance imaging. Korean J Radiol 2012;13:458–66. Vandecaveye V, De Keyzer F, Vander Poorten V, Dirix P, Verbeken E, Nuyts S. Head and neck squamous cell carcinoma: value of diffusion-weighted MR imaging for nodal staging. Radiology 2009;251:134–46. Kim J, Jang Y, Cho G. Multidisciplinary functional MR imaging for prostate cancer. Korean J Radiol 2009;10:535–51. Isßıl B, S ß ebnem O, Teoman C. The role of diffusion weighted imaging in the assessment of axillary lymph nodes. J Breast Health 2013;9:76–81.