Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print October 2, 2007, 10.1148/radiol.2451061804
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Supplemental tables
Right arrow All Versions of this Article:
2451061804v1
245/3/806    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by King, A. D.
Right arrow Articles by Tse, G. M. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by King, A. D.
Right arrow Articles by Tse, G. M. K.
(Radiology 2007;245:806-813.)
© RSNA, 2007

Head and Neck Imaging

Malignant Cervical Lymphadenopathy: Diagnostic Accuracy of Diffusion-weighted MR Imaging1

Ann D. King, FRCR, Anil T. Ahuja, FRCR, David K. W. Yeung, PhD, Devin K. Y. Fong, BSc (Hons), Yolanda Y. P. Lee, FRCR, Kenny I. K. Lei, FRCR, and Gary M. K. Tse, FRCP

1 From the Department of Diagnostic Radiology & Organ Imaging (A.D.K., A.T.A., D.K.Y.F., Y.Y.P.L.), Department of Clinical Oncology (D.K.W.Y., K.I.K.L.), and Department of Anatomical and Cellular Pathology (G.M.K.T.), Faculty of Medicine, the Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, New Territories, Hong Kong Special Administrative Region, China. Received October 19, 2006; revision requested December 21; revision received February 12, 2007; final version accepted April 2. Supported in part by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (project no. CUHK4300/04M). Address correspondence to A.D.K. (e-mail: king2015{at}cuhk.edu.hk).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Purpose: To prospectively determine the diagnostic accuracy of diffusion-weighted magnetic resonance (MR) imaging for discrimination of malignant neck nodes due to lymphoma, squamous cell carcinoma (SCC), and undifferentiated nasopharyngeal carcinoma (NPC), with histologic findings and imaging criteria as reference standards.

Materials and Methods: Ethics committee approval and informed consent were obtained. Patients with malignant lymphadenopathy underwent 1.5-T diffusion-weighted MR imaging. A region of interest was drawn around the malignant node on apparent diffusion coefficient (ADC) maps; ADC values were compared (Kruskal-Wallis test). Receiver operating characteristic analysis was employed to investigate whether ADC values could aid in discrimination among malignancies.

Results: Forty-three patients (34 men, nine women; mean age, 54 years) with 43 nodes underwent imaging. Mean ADC values for lymphoma (n = 8), NPC (n = 17), and SCC (n = 18) were (0.664 ± 0.071 [standard deviation]) x 10–3 mm2/sec, (0.802 ± 0.128) x 10–3 mm2/sec, and (1.057 ± 0.169) x 10–3 mm2/sec, respectively, with significant differences between SCC and lymphoma or NPC (P < .001) and between NPC and lymphoma (P = .04). To optimize sensitivity and specificity with equal weighting, ADC threshold values for distinguishing between SCC and NPC, between SCC and lymphoma, and between NPC and lymphoma were 0.894 x 10–3 mm2/sec, 0.824 x 10–3 mm2/sec, and 0.694 x 10–3 mm2/sec, respectively. To produce a 100% specificity while sensitivity is maximized, the following ADC threshold values were obtained for prediction of differentiation between malignancies: (a) SCC versus lymphoma, greater than 0.824 x 10–3 mm2/sec (sensitivity, 94%), and lymphoma versus SCC, less than 0.767 x 10–3 mm2/sec (sensitivity 88%); (b) NPC versus SCC, less than 0.764 x 10–3 mm2/sec (sensitivity, 47%), and SCC versus NPC, greater than 1.093 x 10–3 mm2/sec (sensitivity, 39%); (c) NPC versus lymphoma, greater than 0.788 x 10–3 mm2/sec (sensitivity, 53%), and lymphoma versus NPC, no suitable threshold value.

Conclusion: Diffusion-weighted MR imaging shows significant differences among malignant nodes of SCC, lymphoma, and NPC. ADC threshold values can help distinguish SCC from lymphoma.

Supplemental material: http://radiology.rsnajnls.org/cgi/content/full/2451061804/DC1

© RSNA, 2007


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Diffusion-weighted imaging is a magnetic resonance (MR) technique that shows potential in the characterization of lesions. The use of diffusion-weighted MR imaging in the head and neck region is challenging, especially because of susceptibility artifacts. However, early reports show that it can be successfully used in this region (1–6). The potential for diffusion-weighted MR imaging to help in distinguishing between benign and malignant cervical nodes has been reported in some studies (3,6), and results of these studies, together with those of studies of predominantly primary tumors (1,2), also suggest that there are differences in the apparent diffusion coefficient (ADC) of squamous cell carcinoma (SCC) and lymphoma. Thus, the purpose of our study was to prospectively determine the diagnostic accuracy of diffusion-weighted MR imaging in the discrimination of malignant neck nodes due to lymphoma, SCC, and undifferentiated nasopharyngeal carcinoma (NPC), by using histologic findings and imaging criteria as the reference standards.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Patients and Reference Standards
The local ethics committee granted ethical approval for the study, and informed consent was obtained. Consecutive patients with cervical lymphadenopathy who were undergoing MR imaging for staging of a histologically proved SCC, lymphoma, or NPC were enrolled in this prospective study. Histologic findings at biopsy were the reference standard for the primary tumor, whereas the reference standard for a node to be considered malignant due to those histologic findings was specific imaging criteria for a malignant node. In each patient, the largest abnormal node was selected for evaluation, provided it met the criteria for classification as a malignant node (defined as a node with the shortest axial diameter of 11 mm in the jugulodigastric region, that of 5 mm in the retropharyngeal region, and that of 10 mm in all other regions of the neck or as any node with necrosis or extracapsular neoplastic spread irrespective of size) (7–9). The maximum and minimum diameters of the node were measured, and the criteria for selection of a malignant node were recorded.

MR Technique
All MR examinations were performed with a 1.5-T whole-body system (Intera NT; Philips Medical Systems, Best, the Netherlands) with a 30 mT/m maximum gradient capability. A standard receive-only head and neck coil was used for both conventional imaging and diffusion-weighted MR imaging to include nodes from the base of the skull to the suprasternal notch. In all patients, the protocol included transverse fat-suppressed T2-weighted turbo spin-echo MR imaging (repetition time msec/echo time msec, 2500/100; section thickness, 4 mm, with no intersection gap; number of signals acquired, two), transverse T1-weighted spin-echo MR imaging (477/12; section thickness, 4 mm, with no intersection gap; number of signals acquired, two), and contrast material–enhanced transverse T1-weighted spin-echo MR imaging performed after administration of a bolus injection of 0.1 mmol/kg of gadoterate meglumine (Dotarem; Guerbet, Aulnay-sous-Bois, France) with a 249 x 512 matrix.

Diffusion-weighted MR imaging was performed before the contrast-enhanced T1-weighted MR imaging sequence, and 11 fat-suppressed diffusion-weighted MR images in the head and neck were acquired in the transverse plane by using a spin-echo single-shot echo-planar imaging sequence (2000/75; section thickness, 4 mm, with no gap; field of view, 230 mm; acquisition matrix, 112 x 112; reconstruction matrix, 256 x 256; number of signals acquired, four) sensitized to incoherent motion by a pair of gradient pulses. Six diffusion-weighted MR images were acquired with b values of 0, 100, 200, 300, 400, and 500 sec/mm2. A relatively low maximum b value was selected to avoid image distortion (10), which is usually observed at higher b values, and to prevent loss of signal.

Three diffusion-weighted MR images were acquired for each b value with the diffusion-sensitization gradient along the readout, phase-encoding, and section-selection directions. By using this imaging sequence, the image coverage was 4.4 cm and the imaging time was 2 minutes 12 seconds. The isotropic diffusion-weighted MR image was calculated on a pixel-by-pixel basis (11), according to the following equation: Formula where Ib is the resultant isotropic signal intensity of the diffusion-weighted image acquired with a b value and Ix, Iy, and Iz are the original signal intensity values of the diffusion-weighted MR images obtained with gradient sensitization along the readout, phase-encoding, and section-selection directions, respectively.

The ADC value was calculated by a medical physicist (D.K.W.Y., with 10 years of working experience with diffusion-weighted MR imaging) with a six-point regression method at b values of 0, 100, 200, 300, 400, and 500 sec/mm2 by using the following equation: Ib = I0·exp(–b·ADC), where Ib and I0 are the mean signal intensity values in the region of interest at a b value of b sec/mm2 and at a b value of 0 sec/mm2, respectively (12). The use of multiple data points, instead of only two data points as commonly used in diffusion-weighted MR imaging studies (1,3), might improve the accuracy of our ADC values, as linear regression performed with multiple experimental data points might reduce systematic errors when few data points are used in the calculation of ADC.

Image Analysis
A region of interest was drawn on the ADC map by a head and neck radiologist (A.D.K., with more than 10 years of working experience with MR imaging of the head and neck). For each node, two measurements were made: (a) ADC of the whole node, for which the ADC value at every section was measured and the mean ADC value was calculated, and (b) ADC of a single section, which was selected because it contained the largest solid component, with the exclusion of any necrotic areas as judged from the T1- and T2-weighted contrast-enhanced MR images. To verify the robustness of the images acquired during the study, a region of interest was drawn to evaluate the ADC value of the upper spinal cord in all three groups of patients. The spinal cord was chosen rather than the cerebrospinal fluid because it has been shown to be more reliable (1). The size of the region of interest for the whole node, a single section through the node, and the spinal cord were recorded.

Statistical Analysis
ADC data were analyzed by using the median and interquartile range, unless otherwise stated. Comparison of the ADC values among lymphoma, SCC, and NPC was performed by using the Kruskal-Wallis test. If there was a significant difference among the three groups, post hoc pairwise comparisons were then made by using the Mann-Whitney test, with Bonferroni correction. Receiver operating characteristic (ROC) analysis was employed to investigate the discriminatory capability of the ADC value for distinguishing between (a) lymphoma and SCC, (b) SCC and NPC, and (c) lymphoma and NPC. The area under the ROC curve was used to give a measure of the global performance of using the ADC values as effective indicators for discrimination.

The value that corresponded to the nearest point of the ROC curve to the top left-hand corner was chosen as the optimal threshold value in that it optimizes both sensitivity and specificity (given equal weighting). The value that corresponded to a specificity of 100%, while the sensitivity was maximized, also was evaluated. Within the group of SCC nodes, the significance of any difference in the ADC value between poorly differentiated and well- or moderately differentiated carcinomas was determined by using the Mann-Whitney test. To avoid dependency issues that may have been caused by the influence of multiple nodes per patient on the results, only one node was examined in each patient. Statistical analyses were performed by using software (SPSS 11.0; SPSS, Chicago, Ill). All statistical tests were two sided, and a difference with a P value of less than .05 was considered statistically significant.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Patients and Nodes
Forty-seven patients were enrolled in the study between March 2004 and March 2006 (Figs 1–3, Table 1). Four patients were excluded because images obtained in them were too distorted by artifact for analysis. The study included 34 men and nine women, with an age range of 35–80 years and a mean age of 54 years, who had lymphadenopathy from non-Hodgkin lymphoma (n = 8), SCC (n = 18), and NPC (n = 17). Data about the size of the region of interest on diffusion-weighted MR images were as follows: For the whole node, the range was 0.30–158.7 cm3 and the mean was 17.9 cm3; for the single section through the node, the range was 0.04–3.06 cm3 and the mean was 0.39 cm3; and for the spinal cord, the range was 0.02–0.08 cm3 and the mean was 0.04 cm3. All nodes included in the results were larger than the minimum size listed in the criteria and, therefore, no node was selected on the basis of only necrosis or extracapsular neoplastic spread. Thirty-five nodes were 15 mm or larger in minimum diameter (for lymphoma, n = 8; for SCC, n = 16; for NPC, n = 11). Two nodes were smaller than 10 mm in minimum diameter and were 6 and 9 mm in minimum diameter and 11 and 14 mm in maximum diameter, but these nodes with lymphadenopathy from NPC were in the retropharyngeal region and therefore met the malignancy criteria for a node at this site. In those nodes with necrosis identified on the conventional MR image, the necrosis could be identified also on the diffusion-weighted MR image and on the ADC map. Details about histologic findings, site of primary tumor, nodal size, and mean ADC values for the whole node, single section through the node, and spinal cord are shown (Table 1).


Figure 1
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1: Flowchart of patients with malignant lymphadenopathy who underwent diffusion-weighted MR imaging (DWI) at our institution during a 2-year period.

 

Figure 2A
View larger version (171K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a: Nonnecrotic left-sided SCC node in a 63-year-old man. (a) Transverse contrast-enhanced T1-weighted MR image (477/12) shows an enlarged node (arrows). (b) Transverse diffusion-weighted MR image (2000/75) acquired by using a b value of 500 sec/mm2 shows the same SCC node (arrows) seen as a hyperintense lesion. (c) ADC map shows the manually draw region of interest (white line) on the SCC node (arrows) that appears as a hypointense lesion. The ADC value was 0.997 x 10–3 mm2/sec.

 

Figure 2B
View larger version (114K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b: Nonnecrotic left-sided SCC node in a 63-year-old man. (a) Transverse contrast-enhanced T1-weighted MR image (477/12) shows an enlarged node (arrows). (b) Transverse diffusion-weighted MR image (2000/75) acquired by using a b value of 500 sec/mm2 shows the same SCC node (arrows) seen as a hyperintense lesion. (c) ADC map shows the manually draw region of interest (white line) on the SCC node (arrows) that appears as a hypointense lesion. The ADC value was 0.997 x 10–3 mm2/sec.

 

Figure 2C
View larger version (139K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2c: Nonnecrotic left-sided SCC node in a 63-year-old man. (a) Transverse contrast-enhanced T1-weighted MR image (477/12) shows an enlarged node (arrows). (b) Transverse diffusion-weighted MR image (2000/75) acquired by using a b value of 500 sec/mm2 shows the same SCC node (arrows) seen as a hyperintense lesion. (c) ADC map shows the manually draw region of interest (white line) on the SCC node (arrows) that appears as a hypointense lesion. The ADC value was 0.997 x 10–3 mm2/sec.

 

Figure 3A
View larger version (147K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a: Necrotic right-sided lymphoma node in a 49-year-old man. (a) Transverse contrast-enhanced T1-weighted MR image (477/12) shows the lymphoma node (arrows) with a central necrotic area (arrowheads). (b) Transverse diffusion-weighted MR image (2000/75) acquired by using a b value of 500 sec/mm2 shows the node (arrows) with a hypointense central necrotic area (arrowheads) surrounded by a hyperintense peripheral rim. (c) ADC map shows the region of interest drawn around the solid component of the tumor (arrows), excluding the central necrotic area (arrowheads). The ADC value was 0.783 x 10–3 mm2/sec.

 

Figure 3B
View larger version (143K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b: Necrotic right-sided lymphoma node in a 49-year-old man. (a) Transverse contrast-enhanced T1-weighted MR image (477/12) shows the lymphoma node (arrows) with a central necrotic area (arrowheads). (b) Transverse diffusion-weighted MR image (2000/75) acquired by using a b value of 500 sec/mm2 shows the node (arrows) with a hypointense central necrotic area (arrowheads) surrounded by a hyperintense peripheral rim. (c) ADC map shows the region of interest drawn around the solid component of the tumor (arrows), excluding the central necrotic area (arrowheads). The ADC value was 0.783 x 10–3 mm2/sec.

 

Figure 3C
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3c: Necrotic right-sided lymphoma node in a 49-year-old man. (a) Transverse contrast-enhanced T1-weighted MR image (477/12) shows the lymphoma node (arrows) with a central necrotic area (arrowheads). (b) Transverse diffusion-weighted MR image (2000/75) acquired by using a b value of 500 sec/mm2 shows the node (arrows) with a hypointense central necrotic area (arrowheads) surrounded by a hyperintense peripheral rim. (c) ADC map shows the region of interest drawn around the solid component of the tumor (arrows), excluding the central necrotic area (arrowheads). The ADC value was 0.783 x 10–3 mm2/sec.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Histologic Findings, Site of Primary Tumor, Nodal Size, and Mean ADC Values

 
ADC Values
Results of statistical analysis of the median ADC value of the whole node and the single section through the node indicate that the ADC value of SCC was significantly higher (ie, less restricted) than the ADC value of NPC and lymphoma (Fig 4, Table 2). The ADC value of NPC was higher than that of lymphoma, but this difference was only statistically significant when ADC values were compared by using the single-section technique. The mean ADC of poorly differentiated SCC was (1.121 ± 0.192 [standard deviation]) x 10–3 mm2/sec for the whole node and (1.033 ± 0.203) x 10–3 mm2/sec for the single-section measurement. For well- and moderately differentiated SCC, the mean ADC was (1.231 ± 0.275) x 10–3 mm2/sec for the whole node and (1.212 ± 0.127) x 10–3 mm2/sec for the single section. There was no significant difference in the ADC values between poorly differentiated SCC and well- and moderately differentiated SCC by using the whole-node (P = .76) or the single-section-throughthe-node (P = .28) technique. There was no significant difference in the ADC value of the spinal cord among the three groups of nodal cancer (Table 2).


Figure 4
View larger version (6K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4: ADC values of the affected nodes obtained from the single-section measurement. Horizontal bar represents mean value for each group.

 

View this table:
[in this window]
[in a new window]

  		Table 2. Median ADC Value for Each Nodal Cancer Group

 
ROC Analysis
The ROC curves and analysis of the ADC values among the three groups of nodal cancer are shown (Fig 5, Tables E1, E2 [http://radiology.rsnajnls.org/cgi/content/full/2451061804/DC1]). By using the single-section technique, the optimal ADC threshold value for distinguishing between SCC and NPC, between lymphoma and SCC, and between lymphoma and NPC was 0.894 x 10–3 mm2/sec, 0.824 x 10–3 mm2/sec, and 0.694 x 10–3 mm2/sec, respectively, when we optimized both sensitivity and specificity with equal weighting. To produce a specificity of 100% while we maximized sensitivity, the optimal ADC threshold values for the prediction of differentiation of (a) SCC from lymphoma and SCC from NPC were greater than 0.824 x 10–3 mm2/sec and greater than 1.093 x 10–3 mm2/sec, respectively; (b) NPC from SCC and NPC from lymphoma were less than 0.764 x 10–3 mm2/sec and greater than 0.788 x 10–3 mm2/sec, respectively; and (c) lymphoma from SCC was less than 0.767 x 10–3 mm2/sec. There was no suitable ADC threshold value to distinguish lymphoma from NPC.


Figure 5A
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5a: ROC curves for the ADC value for discrimination (a) between lymphoma and SCC, (b) between SCC and NPC, and (c) between lymphoma and NPC.

 

Figure 5B
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5b: ROC curves for the ADC value for discrimination (a) between lymphoma and SCC, (b) between SCC and NPC, and (c) between lymphoma and NPC.

 

Figure 5C
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5c: ROC curves for the ADC value for discrimination (a) between lymphoma and SCC, (b) between SCC and NPC, and (c) between lymphoma and NPC.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Diffusion-weighted MR imaging shows significant differences in the ADC values among the three groups of nodal cancers, with restriction in diffusion increasing from SCC to NPC to lymphoma. Higher ADC values for SCC compared with those for lymphoma have been shown also in previous studies about the evaluation of predominantly primary tumors (1,2) and nodes (3). The differences in restriction of diffusion may be attributed to differences in cellularity, necrosis, and perfusion. Cellularity may play a major role, and greater cellularity and less extracellular space (13) may lead to restriction of diffusion in lymphoma. In addition, lymphoma has intracellular differences that may restrict water diffusion still further, although it is debatable whether the slow-diffusing intracellular water component is detectable at the relatively low b values used in this clinical study (14). Necrosis is a further factor that influences diffusion (1,15): As the amount of necrosis increases, the ADC value increases (16). It is interesting that even after necrotic areas were excluded with the single-section technique, SCC nodes continued to show less restriction to diffusion; it has been speculated that areas of micronecrosis also contribute to the ADC of the node (1). However, the size of these necrotic foci is smaller than the size of the MR voxel, and the necrotic foci are probably too small to contribute to the differences in ADC values between lymphoma and SCC (16), whereas pathologic analysis has shown that micronecrosis is common in both lymphoma and SCC (2). The final possible contributing factor to the difference between the ADC value of lymphoma and that of SCC is perfusion. At low b values, perfusion has an influence on the ADC value, and it has been postulated that hypervascular areas in SCC may contribute to the higher ADC value (1–3). However, it has been shown that, at b values greater than 100 sec/mm2, capillary perfusion does not appear to substantially influence the results of quantitative diffusion measurements (17), and, therefore, perfusion may not be an important factor in most clinical studies.

The ADC value of nodes from NPC was found to lie between that of SCC and lymphoma and was significantly lower than that of SCC (P < .001) with both the whole-node and the single-section techniques. When compared with the ADC value of lymphoma, the ADC value of NPC was higher but this result was only barely significant (P = .04) when we used the single-section technique. This result may be explained by the fact that NPC has more histologic similarities with lymphoma than it does with SCC in terms of cellularity and contains numerous lymphocytes as an integral part of the tumor.

In order for diffusion-weighted MR imaging to be used in routine practice for MR characterization of neck nodes, ideally there should be a clearly defined ADC threshold value, the results should be reproducible across centers, and the technique should be robust and widely available. In previous studies (1,3), an overlap has been found in the ADC value for non-Hodgkin lymphoma and poorly differentiated SCC (1,3). In this study, there was only one patient with such an overlap. Therefore, an ADC value of greater than 0.824 x 10–3 mm2/sec could be used to identify SCC and an ADC value of less than 0.767 x 10–3 mm2/sec could be used to identify lymphoma, with a high specificity of 100%, while a sensitivity value of 94% and 88%, respectively, is maintained. In many centers, these ADC value limits are potentially useful because the main differential diagnosis of malignant lymphadenopathy lies between these two cancers. However, the employment of cutoff values is also influenced by the incidence of other cancers in the population. In this study population, NPC is common; therefore, to maintain a high specificity of 100%, the threshold ADC value that could be used to identify SCC would have to be increased to 1.093 x 10–3 mm2/sec at the expense of reducing the sensitivity to only 39%, whereas no lower threshold ADC value could be determined to distinguish lymphoma from NPC.

The ADC threshold values in this study may not be directly reproducible in other centers because of differences in technique, differences that include the selection of the b values and region of interest. A maximum b value of 500 sec/mm2 was used in this study to try to limit the possible effects of distortion and reduced signal-to-noise ratio on the ADC value; such factors are problems at higher b values. Unlike in previous studies, where only two b values were used to obtain data points in the calculation of the ADC value, we used six b values to obtain more data points to allow more accurate calculation of the ADC value by using a curve-fitting procedure. In addition, the results may be influenced also by the selection of the region of interest, especially with respect to whether necrotic portions of the node are included or excluded in the calculation of the mean ADC value. From the data in the limited number of studies in small numbers of patients (1–3), there appears to be a variation in the reproducibility of diffusion-weighted MR imaging results; reported mean ADC values ranged from (0.41 ± 0.105) x 10–3 mm2/sec to (1.13 ± 0.43) x 10–3 mm2/sec for SCC (1,2) and from (0.23 ± 0.056) x 10–3 mm2/sec to (0.66 ± 0.17) x 10–3 mm2/sec (1,3) for non-Hodgkin lymphoma. The mean results of this study, with the single-section technique used to measure only the solid component, appear to be similar to those of Wang et al (1) and Maeda et al (2) ([0.96 ± 0.11] x 10–3 mm2/sec and [0.65 ± 0.09] x 10–3 mm2/sec for SCC and non-Hodgkin lymphoma, respectively), studies also in which only the solid portion of the tumor was measured. It should also be stressed that these results may only be applicable to non-Hodgkin lymphoma. In the study by Razek et al (6), the mean ADC value of Hodgkin lymphoma was higher than that of non-Hodgkin lymphoma, which may have contributed to the greater overlap in the ADC values of lymphoma and SCC in this study.

Finally, in order for diffusion-weighted MR imaging to be used routinely, the technique should be robust and widely available. One of the major problems of performing diffusion-weighted MR imaging in the head and neck is image distortion. Unfortunately, increasing the b value, in an attempt to improve the sensitivity to diffusion, leads to an increase in the susceptibility artifact. To reduce susceptibility artifact, a relatively large transmitter bandwidth (1.833 kHz) and low b values (maximum of 500 sec/mm2) were employed in this study. Other methods to decrease susceptibility artifact include the use of an antisusceptiblity device around the neck (1) or the use of sequences that are less sensitive to this artifact, such as the line scan diffusion-weighted technique (2). These techniques are not widely available and were not used in this study. Despite this fact, diffusion-weighted MR imaging was successfully performed in most patients, and susceptibility artifact was only of sufficient severity to prevent analysis of nodes in 8% of patients. However, artifact may be more of a problem in the assessment of primary pharyngeal tumors that lie directly at the interface with air. In addition, the ADC values of the spinal cord were consistent across the three groups of cancer in this study. As a result, the diffusion-weighted MR imaging technique employed in this study is one that is robust and, by using the single-section technique to analyze the ADC map, takes only a few minutes to perform.

A limitation of the study was that the histologic findings were based on biopsy of the primary tumor, whereas identification of the malignant node was based on imaging criteria. This was because patients with lymphoma and NPC are not treated surgically, and even within the group of patients with SCC, a large number now are treated with chemotherapy and radiation therapy rather than surgery. However, the imaging diagnosis of a malignant node was based on recognized criteria, and only the largest rather than all abnormal nodes were evaluated; as a result, most of the nodes were well above the minimum size criterion for a malignant node. A further limitation was the relatively small number of lymphomatous and poorly differentiated SCC nodes. The number of these nodes could have been increased with evaluation of more than one node in each patient, but this change would have produced dependency issues about the influence of multiple nodes per patient on the results; therefore, this method was not used.

In summary, we found a significant difference in diffusion-weighted MR imaging results for the three nodal cancers (the ADC value for lymphoma was less than that for NPC and less than that for SCC) by using the single-section technique. In a clinical setting, where the main differential diagnosis of malignant lymphadenopathy lies between SCC and lymphoma, an ADC value of greater than 0.824 x 10–3 mm2/sec could be used to distinguish SCC and an ADC value of less than 0.767 x 10–3 mm2/sec could be used to distinguish lymphoma, with a specificity of 100% while a high sensitivity is maintained. However, when a third tumor, namely NPC, is introduced, the overlap in the ADC values results in difficulty in identifying threshold values that maintain high specificity and sensitivity values. It is unclear whether these results are reproducible across different centers, and, at present, each center may need to establish its own reference values until further data are available.


   ADVANCES IN KNOWLEDGE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
IMPLICATION FOR PATIENT CARE
 References
 


   IMPLICATION FOR PATIENT CARE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
References
 


   ACKNOWLEDGMENTS
 
We acknowledge Kai C. Choi, PhD, for his support with the statistical analysis in this study.


   FOOTNOTES
 

Abbreviations: ADC = apparent diffusion coefficient • NPC = nasopharyngeal carcinoma • ROC = receiver operating characteristic • SCC = squamous cell carcinoma

Author contributions: Guarantor of integrity of entire study, A.D.K.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, A.D.K., D.K.W.Y.; clinical studies, all authors; statistical analysis, D.K.W.Y., D.K.Y.F.; and manuscript editing, A.D.K., D.K.W.Y.

Authors stated no financial relationship to disclose.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 

  1. Wang J, Takashima S, Takayama F, et al. Head and neck lesions: characterization with diffusion-weighted echo-planar MR imaging. Radiology 2001;220:621–630. [Abstract/Free Full Text]
  2. Maeda M, Kato H, Sakuma H, Maier SE, Takeda K. Usefulness of the apparent diffusion coefficient in line scan diffusion-weighted imaging for distinguishing between squamous cell carcinomas and malignant lymphomas of the head and neck. AJNR Am J Neuroradiol 2005;26:1186–1192. [Abstract/Free Full Text]
  3. Sumi M, Sakihama N, Sumi T, et al. Discrimination of metastatic cervical lymph nodes with diffusion-weighted MR imaging in patients with head and neck cancer. AJNR Am J Neuroradiol 2003;24:1627–1634. [Abstract/Free Full Text]
  4. Motoori K, Yamamoto S, Ueda T, et al. Inter- and intratumoral variability in magnetic resonance imaging of pleomorphic adenoma. J Comput Assist Tomogr 2004;28:233–246. [CrossRef][Medline]
  5. Habermann CR, Gossrau P, Graessner J, et al. Diffusion-weighted echo-planar MRI: a valuable tool for differentiating primary parotid gland tumors? Rofo 2005;177:940–945. [Published correction appears in Rofo 2005;177:1312.]
  6. Razek AA, Soliman NY, Elkhamary S, Alsharaway MK, Tawfik A. Role of diffusion-weighted MR imaging in cervical lymphadenopathy. Eur Radiol 2006;16:1468–1477. [CrossRef][Medline]
  7. Som PM. Lymph nodes of the neck. Radiology 1987;165:593–600. [Free Full Text]
  8. van den Brekel MW, Stel HV, Castelijns JA, et al. Cervical lymph node metastasis: assessment of radiologic criteria. Radiology 1990;177:379–384.
  9. Lam WW, Chan YL, Leung SF, Metreweli C. Retropharyngeal lymphadenopathy in nasopharyngeal carcinoma. Head Neck 1997;19:176–181. [CrossRef][Medline]
  10. Jezzard P, Barnett AS, Pierpaoli C. Characterization of and correction for eddy current artifacts in echo planar diffusion imaging. Magn Reson Med 1998;39(5):801–812.
  11. Ward R, Caruthers S, Yablon C, Blake M, DiMasi M, Eustace S. Analysis of diffusion changes in posttraumatic bone marrow using navigator-corrected diffusion gradients. AJR Am J Roentgenol 2000;174:731–734. [Abstract/Free Full Text]
  12. Szafer A, Zhong J, Anderson AW, Gore JC. Diffusion-weighted imaging in tissues: theoretical models. NMR Biomed 1995;8:289–296. [Medline]
  13. Guo AC, Cummings TJ, Dash RC, Provenzale JM. Lymphomas and high-grade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 2002;224:177–183. [Abstract/Free Full Text]
  14. Mulkern RV, Gudbjartsson H, Westin CF, et al. Multi-component apparent diffusion coefficients in human brain. NMR Biomed 1999;12:51–62. [CrossRef][Medline]
  15. Kim YJ, Chang KY, Song IC, et al. Brain abscess and necrotic or cystic brain tumor: discrimination with signal intensity on diffusion weighted imaging. AJR Am J Roentgenol 1998;171:1487–1490. [Abstract/Free Full Text]
  16. Lang P, Wendland MF, Saeed M, et al. Osteogenic sarcoma: noninvasive in vivo assessment of tumor necrosis with diffusion-weighted MR imaging. Radiology 1998;206:227–235. [Abstract/Free Full Text]
  17. Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, Laval-Jeantet M. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging, Radiology 1988;168:497–505.



This article has been cited by other articles:


Home page
 RadiologyHome page
V. Vandecaveye, F. De Keyzer, V. Vander Poorten, P. Dirix, E. Verbeken, S. Nuyts, and R. Hermans
Head and Neck Squamous Cell Carcinoma: Value of Diffusion-weighted MR Imaging for Nodal Staging
Radiology, April 1, 2009; 251(1): 134 - 146.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Neuroradiol.Home page
R. Kapur, A.R. Sepahdari, M.F. Mafee, A.M. Putterman, V. Aakalu, L.J.A. Wendel, and P. Setabutr
MR Imaging of Orbital Inflammatory Syndrome, Orbital Cellulitis, and Orbital Lymphoid Lesions: The Role of Diffusion-Weighted Imaging
AJNR Am. J. Neuroradiol., January 1, 2009; 30(1): 64 - 70.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Supplemental tables
Right arrow All Versions of this Article:
2451061804v1
245/3/806    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by King, A. D.
Right arrow Articles by Tse, G. M. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by King, A. D.
Right arrow Articles by Tse, G. M. K.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE