| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology and Cancer Biology, Nagasaki University School of
Dentistry, 1-7-1 Sakamoto, Nagasaki 852-8588, Japan. Address correspondence to
T. Nakamura
2 Philips Medical Systems, Toyko, Japan.
Received November 30, 2004;
accepted after revision February 1, 2005.
Address correspondence to T. Nakamura
(taku{at}net.nagasaki-u.ac.jp).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. We performed MR microimaging on 24 histologically proven metastatic nodes, 14 histologically proven lymphomas, and 35 histologically or clinically proven benign nodes in the necks of 26 patients. The lymph nodes were imaged with T1-weighted spin-echo, fat-suppressed T2-weighted turbo spin-echo, and spin-echo diffusion-weighted echo-planar sequences using a 47-mm microscopy coil.
RESULTS. MR microimaging provided high-resolution images of the nodes. Hilar fat was lost in 92%, 79%, and 46% of the metastatic nodes, lymphomas, and benign nodes, respectively. Smooth nodal margins were lost in 58%, 23%, and 9% of metastatic nodes, lymphomas, and benign nodes, respectively. Heterogeneous nodal parenchyma on T1- or fat-suppressed T2-weighted images, or both, was observed in 88%, 29%, and 23% of metastatic nodes, lymphomas, and benign nodes, respectively. The apparent diffusion coefficients were significantly different among these three node groups (p < 0.001), with metastatic nodes being the highest, followed by benign nodes. Logistic regression analyses showed that heterogeneous nodal parenchyma and apparent diffusion coefficient levels were significant in discriminating metastatic nodes, and apparent diffusion coefficient levels in discriminating lymphomas. Combined use of these MR microscopic criteria on nodal architecture and apparent diffusion coefficients yielded 90% accuracy (86% sensitivity, 94% specificity) and 93% accuracy (85% sensitivity, 95% specificity) for discriminating metastatic nodes and lymphomas, respectively.
CONCLUSION. The nodal architecture and apparent diffusion coefficient levels on MR microimaging may provide useful information in diagnosing benign and malignant nodes in the neck.
Keywords: cancer • head and neck imaging • lymph nodes • MR microimaging
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
On CT and MRI, the internal architecture of the node was poorly imaged, and additional evaluation of the parenchymal signals from the node did not significantly improve the diagnostic ability of the size criteria for metastatic nodes [7]. The nodal size alone was also reported to have limited value in the diagnosis of metastatic nodes in the neck [2]. A separate trial using diffusion-weighted MRI was performed for the detection of metastatic nodes to show that the apparent diffusion coefficient (ADC) levels were useful determinants in discriminating metastatic nodes from benign nodes in patients with head and neck cancer [8]. The ADCs also effectively differentiate the nodal lymphomas from metastatic nodes.
Lymph node metastases, which are associated with proliferation and necrosis of cancer cells, angiogenesis, and obliteration of hilar fat, cause drastic changes in the components and structures of the nodes. Recently, an application of the small surface coil has been reported to show detailed architecture of the superficial organs [9]. In this study, we sought to determine whether the assessment of the architecture and ADCs of the nodes on high-resolution MR microimaging using a microscopy coil might be effective in differentiating metastatic nodes, lymphomas, and benign nodes in the neck.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
All metastatic nodes were surgically removed or biopsied and were histopathologically proven. Of the 10 patients with squamous cell carcinoma, the primary cancers arose in the oropharynx (n = 3), gingiva (n = 3), larynx (n = 3), and tongue (n = 1).
The lymphoma nodes were biopsied or excised and histologic types were determined. The histologic types of the lymphomas were four B-cell neoplasms, including one mantle cell lymphoma, two diffuse large B-cell lymphomas, and one follicular lymphoma; one T-cell neoplasm (peripheral unspecified lymphoma); and one Hodgkin's lymphoma (nodular lymphocyte predominant).
Some of the benign nodes were imaged and biopsied because they were clinically suggestive of malignancy, mainly as a result of their large size; and the remaining benign nodes were clinically evaluated for their responsiveness to treatment with antibiotics after long (6 months) periods of follow-up.
MR Microimaging
MR microimaging was performed using a 1.5-T MR imager (Gyroscan Intera 1.5
T Master, Philips Medical Systems) and a 47-mm microscopy coil. The coil was
positioned so that the palpable nodes were under the coil and was secured
using adhesive tape. Axial T1-weighted images of the cervical lymph nodes were
obtained using a conventional spin-echo sequence (TR/TE, 550/10; number of
signal acquisitions, 3). Axial, fat-suppressed SPIR (spectral presaturation
with inversion recovery) T2-weighted images of the lymph nodes were obtained
using a turbo spin-echo sequence (3,000/90; number of signal acquisitions, 6)
and a turbo factor of 11. For all of these sequences, the section thickness
was 2 mm. MRI was performed with a matrix of 160 x 128, a field of view
of 7 cm, and an interslice gap of 0.2 mm. The acquisition time for each
sequence was less than 4 min. To compensate for the image intensity
inhomogeneity inherent to the use of small microscopy coils, we used the CLEAR
(constant level appearance) postprocessing technique (Philips Medical
Systems). CLEAR uses the premeasured sensitivity profile of the coil to
calculate the compensation needed to apply to the pixel intensities to achieve
even image intensity. The 26 patients also underwent conventional T1- and
fat-suppressed T2-weighted imaging for primary cancer lesions and for deeper
nodal lesions.
Diffusion-Weighted MR Microimaging
Axial diffusion-weighted images of the nodes were obtained using a
single-shot, spin-echo type echo-planar imaging sequence (2,973/121; number of
signal acquisitions, 6) using the same 47-mm microscopy coil. The sequence was
repeated for two values of motion-probing gradients (b = 500 and 1,000
sec/mm2). The section thickness was 2 mm. Diffusion-weighted MRI
was performed with a matrix of 80 x 56, a field of view of 70 mm, and an
interslice gap of 0.2 mm. The acquisition time was less than 3 min.
The ADCs were determined using the two b factors, by the following formula:
 |
Assessments for Diagnostic Ability
We calculated the sensitivity (number of nodes positive at both imaging and
histology / number of nodes positive at histology) and specificity (number of
nodes negative at both imaging and histology / number of nodes negative at
histology) after varying combinations of MR microimaging criteria determined
on T1- and fat-suppressed T2-weighted images plus varying thresholds of ADC
value, for discrimination of metastatic nodes, nodal lymphomas, and benign
nodes. The accuracy was calculated by the following formula: (number of nodes
positive at both imaging and histology + number of nodes negative at both
imaging and histology) / total number of nodes.
Positive and negative predictive values were also used to assess the performance of the MRI criteria in the detection of metastatic nodes. The positive and negative predictive values are the percentages of nodes interpreted using the MR microscopic criteria as positive or negative for the particular diagnoses that were positive histopathologically.
Statistical Analyses
Logistic analysis was performed to identify MRI characteristics that could
be used as predictive indicators for differentiating metastatic nodes,
lymphomas, and benign nodes in the neck. MR images that were found to be
important at univariate analysis were entered into multivariate models to
determine their independent predictive value. Analysis was performed with the
statistical software package SPSS version 6.1 for Windows (SPSS).
Significance in the architectural data and ADC data between the different patient groups was determined using the Fisher's exact probability test, Mann-Whitney U test, or Student's t test, as indicated in the table footnotes, using commercially available statistical software (StatView 4.51, Abacus Concepts).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
Margins of the Nodes
We found that T1-weighted imaging was good for the depiction of nodal
margins. Nodal margins blending into the surrounding tissue were found in 58%
of metastatic nodes (Figs. 2A,
2B, and
2C,
Table 1). Irregular margins
were also found in some lymphomas (23%) but were rarely observed in benign
nodes (9%) (Table 1).
|
|
|
Parenchymal Architecture of the Nodes
Metastatic nodes frequently (88%) exhibited heterogeneous architecture of
the parenchyma on T1- or fat-suppressed T2-weighted MR microimages, or both
(Table 1). MR microimaging
showed that metastatic nodes contained hypointense to intermediately intense
areas indicative of cancer cell nests and interstitial fibrous tissue with
(Figs. 3A,
3B,
3C,
4A,
4B,
4C,
4D,
4E, and
4F) or without (Figs.
4A,
4B,
4C,
4D,
4E, and
4F) central hyperintense areas
indicative of liquefaction necrosis on fat-suppressed T2-weighted images.
These changes in nodal architecture were depicted well on fat-suppressed
T2-weighted images compared with T1-weighted images
(Table 2).
|
|
|
|
|
|
|
|
|
|
Heterogeneous architecture of the nodal parenchyma was also evident in some lymphomas (29%) and benign nodes (23%), but the incidence was significantly low compared with metastatic nodes (Table 1). Basically, lymphomas exhibited homogeneous architecture. However, lymphomas may be associated with narrowed hilum and blood vessels, which were depicted as the so-called small-vessel sign [10] on fat-suppressed T2-weighted images (Figs. 5A, 5B, and 5C). In two cases of lymphomas (diffuse large B-cell type), unenhanced foci were noted, suggestive of necrosis. Unlike metastatic nodes, however, these foci were hypointense on T2-weighted images.
|
|
|
|
|
|
|
|
|
On multivariate analysis, however, the heterogeneous parenchyma and the ADCs were statistically significant in differentiating metastatic nodes from benign nodes or nodal lymphomas (Table 3). On the other hand, only the ADCs were statistically significant in differentiating lymphomas from the other two entities (Table 4).
Finally, we assessed the diagnostic ability of MRI findings using combined criteria of nodal architecture and ADCs. The best results we obtained were as follows: When a node was diagnosed as metastatic if it exhibited heterogeneous parenchyma on T1- or fat-suppressed T2-weighted images and an ADC equal to or greater than 0.73 x 10-3 mm2/sec, we obtained 83% sensitivity, 94% specificity, 87% positive and 92% negative predictive values, and 90% accuracy. When a node was diagnosed as lymphoma if it had an ADC equal to or less than 0.51 x 10-3 mm2/sec, we obtained 86% sensitivity, 95% specificity, 80% positive and 97% negative predictive values, and 93% accuracy.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Proliferating metastatic cells frequently extend into the nodal hilum, thereby obstructing the hilum. The intact hilum contains fat tissue, blood vessels, and efferent lymphatic vessels, and thus the loss of fat tissue in the hilum has been an important imaging feature highly suggestive of metastatic nodes [1, 2]. However, conventional MRI inconsistently detects this feature. The high-resolution images on MR microscopy coils enable efficient assessment of the nodal structures on T1-weighted and fat-suppressed T2-weighted images. It is interesting to assume that when metastatic foci are small, and the intact hilar structures are not yet obliterated, the heterogeneous MR signals from proliferating cancer cells may be more critical than the loss of hilar structures in diagnosing metastatic nodes. Our results support this notion (Table 3).
Loss of the hilar fat was most frequently observed in, but not limited to, the metastatic nodes. Hilar fat was lost in 79% of nodal lymphomas and in 46% of benign nodes. Consequently, loss of hilar fat was a significant finding differentiating metastatic from benign nodes but not from lymphomas (Table 1). Previous studies using sonography showed that hilum structures with or without vascular flow signals were frequently lost or narrowed in lymphomas [11, 12]. Our MRI findings were consistent with these previous studies.
The heterogeneity of the internal architecture of the nodes on MR images may be due to the presence of necrosis, cancer nests, and interstitial fibrous tissue, which has been considered to be a pathognomonic feature for metastatic nodes from head and neck squamous cell carcinomas [13]. We found in this study that heterogeneous parenchyma was a significant finding for discriminating metastatic nodes from lymphomas and benign nodes (Table 1). Central nodal necrosis was reported to occur in 32% of metastatic nodes [14]. Enhanced CT may be more effective than unenhanced CT or enhanced MRI in detecting nodal necrosis [6]. However, a recent report showed that no significant difference occurred in sensitivity or specificity for detecting nodal necrosis between enhanced CT and enhanced MRI [15].
In our study, high-resolution imaging using a microscopy coil showed that the hyperintense areas were surrounded by rims or areas that were hypointense relative to the residual nodal parenchyma. Histologic evaluation confirmed that such hypointense areas corresponded to cancer cell nests, which were composed of proliferating cancer cells and coagulation necrosis, and to surrounding fibrous tissue. On the other hand, central hyperintense areas were caused by liquefaction necrosis. Hypointense foci with or without hyperintense areas on fat-suppressed T2-weighted images were noted in 67% of the metastatic nodes and in 14% and 9% of lymphomas and benign nodes, respectively (Table 2). Considering the reportedly low incidence of central nodal necrosis in metastatic nodes in the neck [14], a combined use of these MR microimaging features (hyperintense focus of central nodal necrosis plus hypointense area) on fat-suppressed T2-weighted images may be a better MR criterion than central nodal necrosis alone for the diagnosis of metastatic nodes.
We found in this study that heterogeneous parenchyma was not an MRI feature limited to metastatic nodes, but was also observed in nodal lymphomas and even in benign nodes. Our finding that heterogeneous architecture of the nodes occurred in approximately 30% of nodal lymphomas was unexpected. The observed heterogeneity in the nodal architecture may be due to the presence of necrotic areas. Necrosis is well described in Hodgkin's disease, but it has been considered to be uncommon in head and neck non-Hodgkin's disease. King et al. [16] reported that 27% of non-Hodgkin's disease in the neck was associated with necrosis. Thus, the notion that the homogeneous MRI architecture is characteristic of nodal lymphomas is open to dispute.
ADC measurement may shed light on other aspects of the diseased nodes and may provide additional information distinct from the morphologic features. Among the three nodal diseases, ADCs were highest in metastatic nodes and lowest in lymphomas (Fig. 7). A previous study showed that ADCs were significantly different among metastatic nodes, nodal lymphomas, and benign nodes [8, 17]. In our study, we confirmed that the ADC measurements significantly and independently contributed to the discrimination of metastatic nodes and nodal lymphomas in the neck (Tables 3 and 4). We conclude that a node that has an ADC value equal to or less than 0.51 x 10-3 mm2/sec is highly suggestive of lymphoma.
The MR microimaging technique has some limitations in the evaluation of head and neck lymph nodes: First, MR microimaging requires additional time for the examination and evaluation of the image data. Second, the small field of view makes it difficult to see the surrounding structures and the exact location of lymph nodes. Third, the technique does not apply to deep nodes. A rapid and sensitive MRI technique, such as a STIR turbo sequence coupled with parallel imaging (SENSE [sensitivity encoding]), would be useful to survey suspicious nodes in the neck before an extensive examination with a microscopy coil.
In the present study we examined relatively larger nodes that were 8 mm or more in the long-axis diameter on axial images. MR microscopy may be useful for the diagnosis of nodes that are relatively large, but size criteria alone are not effective in differentiating metastatic nodes in the neck. However, an answer to the question whether MR microscopy is also useful in the diagnosis of smaller (< 8 mm) nodes, and in the effective detection of micrometastasis, must be revealed by future studies.
In conclusion, high-resolution MRI using a microscopy coil can effectively characterize the morphologic details of benign and metastatic nodes in the neck without using gadolinium enhancement. This technique provided better results than previous studies using MRI and CT [7]. In particular, a palpable node in the neck that exhibits heterogeneous parenchyma on T1- or fat-suppressed T2-weighted images, or both, and also has an ADC value equal to or greater than 0.73 x 10-3 mm2/sec is suggestive of a metastatic node, whereas a node that has an ADC value equal to or less than 0.51 x 10-3 mm2/sec is suggestive of lymphoma. Therefore, the proposed MRI technique may be the primary technique for palpable nodes in the neck.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Tanitame, K. Sasaki, T. Sone, S. Uyama, M. Sumida, T. Ichiki, and K. Ito Anterior Chamber Configuration in Patients with Glaucoma: MR Gonioscopy Evaluation with Half-Fourier Single-Shot RARE Sequence and Microscopy Coil Radiology, August 11, 2008; (2008) 2483071556. [Abstract] [Full Text]
|
|
|
|
 |
|
 Y. Kimura, M. Sumi, N. Sakihama, F. Tanaka, H. Takahashi, and T. Nakamura MR Imaging Criteria for the Prediction of Extranodal Spread of Metastatic Cancer in the Neck AJNR Am. J. Neuroradiol., August 1, 2008; 29(7): 1355 - 1359. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Kawai, M. Sumi, and T. Nakamura Turbo Short {tau} Inversion Recovery Imaging for Metastatic Node Screening in Patients with Head and Neck Cancer AJNR Am. J. Neuroradiol., June 1, 2006; 27(6): 1283 - 1287. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
