| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Shinshu University School of Medicine, 3-1-1 Asahi,
Matsumoto 390-8621, Japan.
2
Department of Surgery, Shinshu University School of Medicine, Matsumoto
390-8621, Japan.
Received November 6, 2000;
accepted after revision March 29, 2001.
Address correspondence to S. Takashima.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. MR imaging was performed on the normal trachea of one cadaver and 30 healthy subjects as a standard of reference. Then, MR imaging findings in 67 patients with thyroid carcinoma were reviewed and correlated with surgical and pathologic findings. A logistic regression model was used to determine which MR imaging features were significant for predicting tracheal invasion.
RESULTS. Twenty-three (34%) of the 67 patients had tracheal invasion. Logistic regression model analysis revealed that significant MR characteristics for determining tracheal invasion included soft-tissue signal in the tracheal cartilage (p < 0.001), intraluminal mass (p < 0.001), and degree of tumor circumference around the trachea (p = 0.001). The highest accuracy (90%) for determining tracheal invasion was achieved using a combination of findings. A case was considered positive for tracheal invasion if there was soft-tissue signal in the cartilage, an intraluminal mass, or a tumor that abutted a circumference of the trachea of 180° or greater. Using these factors resulted in seven false-positive diagnoses because soft-tissue signal in the cartilage was sometimes seen in healthy trachea. Although intraluminal mass invariably reflected deep tracheal invasion, soft-tissue signal in the cartilage rarely indicated actual cartilage invasion but rather indicated tumor extension between the cartilaginous rings.
CONCLUSION. Tracheal invasion by thyroid carcinomas can be accurately diagnosed with MR imaging, and using a combination of criteria is the most accurate method of predicting this phenomenon.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Until now, the presence of a mass lesion in the tracheal lumen was the only reliable radiologic sign of tumor invasion of the trachea. Invasion of the superficial tracheal layers has been identified by surgeons only at the time of surgical exploration [5,6,7]. Many reports have discussed MR imaging of various diffuse and nodular thyroid diseases and nodal metastasis from thyroid carcinomas [8,9,10], but, to our knowledge, nothing in the literature has addressed the capability of this technique for diagnosing tumor infiltration of the trachea. Our study evaluated the accuracy of MR imaging in predicting tracheal invasion by thyroid carcinomas, and we propose optimal criteria for identifying this invasion.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study Design in Cadaver
One fresh trachea was resected from a 55-year-old man who died of splenic
rupture. MR imaging was performed 2 hr after his death. The specimen was
imaged with a 1.5-T MR imager (Signa; General Electric Medical Systems,
Milwaukee, WI) using a quadrature head coil. T1-weighted images (TR/TE,
500/16; excitations, 2) and fast spin-echo T2-weighted images (5000/81;
excitations, 2; thickness, 5 mm; gap, 1 mm; field of view, 16 cm x 16
cm; matrix, 320 x 256) were obtained in the axial plane. T1-weighted
images and fast spin-echo T2-weighted images were repeated in the coronal
plane with the same pulse sequences but with a slice thickness of 2.6 mm and a
field of view of 20 x 20 cm.
Then, the specimen was fixed with formalin. A total of 20 consecutive permanent sections of the upper trachea with a 1-mm interval were obtained in the axial plane, and a total of eight consecutive sections of the remaining upper trachea with a 1-mm interval were obtained in the coronal plane. All sections were stained with H and E.
Study Design in Healthy Control Subjects and Patients with Thyroid
Carcinoma
Thirty healthy control subjects with a mean age of 55 years (range, 23-76
years), including 12 men and 18 women, were selected from recent MR imaging
records because they had no evidence of disease in the thyroid gland and the
trachea and because they were of comparable age.
Then, 67 consecutive patients with thyroid carcinomas were evaluated retrospectively. The study included 15 men and 52 women with a mean age of 57 years (range, 20-82 years). Fifty-six patients had papillary carcinomas, four had follicular carcinomas, six had anaplastic carcinomas, and one had metastatic squamous cell carcinoma. These patients had undergone MR imaging before surgery because, on the basis of physical examination and sonographic findings, they were suspected of having invasive thyroid tumor or nodal metastases in the neck. All patients in this study underwent sonographic evaluation before MR imaging. In each case, the surgeons determined whether MR imaging should be performed. They suspected tumor invasion if the movement of the tumor was limited relative to the concerned organ. From 1 to 26 days elapsed between MR imaging and surgery. Tracheal invasion was diagnosed if residual tumor in the trachea was apparent with positive margins (surgical diagnosis) or if tumor invasion was verified in surgically resected specimens (pathologic diagnosis).
MR Imaging Techniques
MR imaging was performed with a 1.5-T Magnetom (Siemens, Erlangen, Germany)
or a 1.5-T Signa (General Electric Medical Systems) MR unit using a volume
neck coil or a Helmholtz coil. T1-weighted images (700/13; excitations, 2, or
900/15; excitations, 1) and conventional spin-echo T2-weighted images
(2000/70; excitations, 1) (n = 37) or fast spin-echo T2-weighted
images (3200/91; excitations, 2) (n = 30) were obtained in the axial
plane. A total of 18 images from the level of the mandibular angle to the
sternal notch were obtained with a slice thickness of 5 mm and an intersection
gap of 1-2 mm. The field of view was 20 cm x 20 cm with a 256 x
192 acquisition matrix. After IV bolus injection of 0.1-0.2 mmol/kg of
gadopentetate dimeglumine (Magnevist; Schering, Berlin, Germany),
contrast-enhanced axial T1-weighted images (700/13; excitations, 2, or 900/15;
excitations, 1) were obtained in 55 of the 67 patients with thyroid carcinomas
and all 30 healthy control subjects.
MR Image Analysis and Statistical Methods
Two radiologists evaluated MR images of the trachea in 30 healthy control
subjects and determined the findings by consensus. The MR imaging features
assessed included tracheal shape, tracheal location, and tracheal-wall signal
intensity (mucous membrane, cartilage, adventitia, and posterior membranous
portion). Tracheal shapes were assigned to one of three classifications:
horseshoe, elliptical, or circular configuration; trachea with a locally
straightened wall; and trachea with an inward concave deformity. The horseshoe
definition was used when the trachea was slightly elongated in the
anterior-posterior dimension relative to the transverse dimension. The
elliptical or circular configuration was assigned when the anteroposterior and
transverse dimensions of the trachea were close to being equal. The locally
straightened shape was defined as being when the trachea was locally flattened
along the anterior curve. Tracheal location was expressed as the shortest
distance between the center of the trachea and the midline of the spinal
body.
Next, the same two radiologists independently evaluated MR images of the 67 patients with thyroid carcinomas without knowledge of the accompanying clinical, surgical, and pathologic findings. Assessed MR findings included the maximum axial diameter of thyroid tumors, the shape of the trachea, the distance of tracheal displacement, the degree of tumor circumference of the trachea, the presence or absence of soft-tissue signal in the tracheal cartilage, tumor extension to the posterior membranous portion of the trachea, and intraluminal mass. The tracheal shape and the distance of tracheal displacement in the 67 patients were assessed in the same way as in the healthy control subjects. The degree of tumor circumference of the trachea with loss of intervening fat plane was classified in four consecutive grades: grade I, 0-89°; grade II, 90-179°; grade III, 180-269°; grade IV, 270° or more. Soft-tissue signal within the tracheal cartilage was determined when intermediate soft-tissue signal was seen instead of the normally low signal or signal void representing the cartilage adjacent to the thyroid tumors. Because the difference in appearance between cartilage and soft tissue was higher in T2-weighted images and in contrast-enhanced T1-weighted images than that in unenhanced T1-weighted images, we used these two images to identify this finding in this study. The intraluminal mass was defined as the presence of apparent soft-tissue mass in the tracheal lumen. In this study, all thyroid cancers enhanced when T1-weighted images obtained before and after administration of contrast material were compared. On contrast-enhanced MR images, slight enhancement was considered to be present if the signal intensity of the tumor was lower than that of the healthy thyroid gland; moderate enhancement was assigned if the signal intensity was similar to that of the healthy thyroid gland; and pronounced enhancement was assigned if the signal intensity of the tumor was higher than that of the healthy thyroid gland.
Interobserver agreement of the two reviewers was assessed by calculation of kappa values for the five sets of category data (the degrees of tumor circumference of the trachea, tracheal shape, presence or absence of soft tissue in the tracheal cartilage, tumor extension to posterior membranous portion of the trachea, and intraluminal mass) and by calculation of correlation coefficients for the two sets of continuous data (the maximum axial diameter of the tumors and the distance of tracheal displacement). The final interpretation for the category data was made on the basis of the consensus of both radiologists, whereas the averaged values of the two reviewers were used for the continuous data.
A stepwise logistic regression model was carried out using the seven parameters as independent variables to determine the significant factors for predicting tracheal invasion. Finally, we proposed the optimal criteria for predicting this phenomenon. Fisher's exact test, Student's t test, or the chi-square test was used for statistical analysis; a p value less than 0.05 was considered statistically significant (SPSS software; Statistical Package for the Social Sciences, Chicago, IL) was used for all of the statistical analyses.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
On coronal MR images, much variation was seen in vertical length of each cartilage ring, ranging from 2 to 5 mm. Membranous portions between the cartilage rings measured 0.5-1 mm and showed intermediate signal intensity on both MR pulse sequences (Fig. 1C). On T2-weighted images, the signal intensity of tracheal mucous membrane was greater than that of the posterior tracheal membrane or of the intercartilaginous membrane, which had signal intensity greater than the tracheal cartilage.
|
On histologic studies, the mucous membrane consisted of ciliated columnar epithelium, connective tissue, and mucous glands (Fig. 1D). Although most of the tracheal rings showed a complete horseshoe shape, some of them were incomplete, fragmentary, or bifuracted. The intervening membranous portions between the cartilage consisted of collagenous fibers, smooth muscle, and minor salivary gland (Fig. 1E). The adventitia, which is 0.2-0.3 mm thick, is composed of perichondrium and dense connective tissue intermingled with elastic fibers.
|
|
MR Findings in Healthy Control Subjects
None of the tracheae of healthy subjects had a locally concave shape. On MR
imaging, the trachea showed a horseshoe or circular configuration in 27 (90%)
of the 30 healthy control subjects and a locally straightened shape in the
remaining three subjects. The trachea was located in the midline in six
subjects (20%), to the right side of the midline in 19 subjects (63%), and to
the left side in five subjects (17%). The distance between the center of the
trachea and the midline of the spinal body were within 4 mm in all 30
subjects.
Signal intensity on unenhanced MR images and the thickness of the mucous membrane and posterior membranous portion of the trachea in the 30 healthy control subjects were similar to the findings in the cadaveric trachea. The adventitia was too thin to identify on MR images in all subjects. Eight subjects (27%) had incomplete fat planes between the cartilage and the thyroid gland. On T2-weighted images of 30 patients, 26 (87%) had a horse-shoe-shaped trachea of low signal intensity and four (13%) showed small areas of higher signal intensity in the cartilage.
On contrast-enhanced T1-weighted images, mucous membrane and the thyroid gland showed intermediate to fair enhancement. However, no discernible enhancement, which could have enabled us to more easily differentiate the tracheal rings from other structures, was found in the cartilage, intercartilaginous membranes, or posterior membranous portion.
Assessment of Patients with Thyroid Carcinomas
Of the 67 patients with thyroid carcinomas, tracheal invasion was diagnosed
in 23 patients (17 papillary carcinomas, five anaplastic carcinomas, and one
metastatic squamous cell carcinoma) either pathologically (n = 19) or
surgically (n = 4). Partial resection of the trachea was performed in
21 of the 67 patients. Of these, 19 patients had pathologically verified
tracheal invasion (four intraluminal, nine submucosal, three
intercartilaginous, and three adventitial invasion); two patients had no tumor
invasion but had only reactive collagenous proliferation of the
adventitia.
MR Imaging Findings
Kappa values for agreement between for the two reviewers were 0.83 for
intraluminal mass, 0.81 for the degrees of tumor circumference, 0.78 for the
tracheal shape, 0.78 for the tumor extension to the posterior membranous
portion, and 0.68 for soft-tissue signal in the tracheal cartilage. All these
values indicated almost perfect or substantial agreement
[11]. Pearson's correlation
coefficient was 0.99 (p < 0.001) for the maximum axial diameter of
the tumors and 0.96 (p < 0.001) for the distance of tracheal
displacement, which indicated strong agreement of the measurement between the
two radiologists.
Of the 67 thyroid carcinomas, 24 were in the right lobe, 18 in the left lobe, 11 in the right lobe and isthmus, five in the left lobe and isthmus, five in both lobes and the isthmus, and four in both lobes of the thyroid gland. A mean maximum axial diameter of 67 thyroid carcinomas was 3.5 ± 1.5 cm (one SD), ranging from 1.1 to 8.4 cm. On unenhanced T1-weighted images, 32 tumors (48%) appeared hyperintense, 29 (43%) isointense, and six (9%) hypointense relative to the healthy thyroid parenchyma. On T2-weighted images, 57 lesions (85%) appeared hyperintense, six (9%) hypointense, and four (6%) isointense. Of the 55 patients in which contrast-enhanced T1-weighted images were obtained, enhancement was pronounced in 46 tumors (84%), moderate in six (11%), and mild in three (5%). All thyroid carcinomas were identified with one or a combination of the three MR pulses.
The maximum axial diameters (p < 0.001) and the distance of displacement of the trachea (p = 0.008) in patients with tracheal invasion were significantly greater than those in patients without invasion (Table 1). The incidence of soft tissue in the cartilage (p < 0.001), tumor extension to the posterior membranous portion (p < 0.001), and intraluminal mass (p < 0.001) was significantly greater in patients with tracheal invasion than in those without invasion. There were significant differences for the distribution of deformity (p < 0.001) and tumor circumference of the trachea (p <0.001) between the patients with tracheal invasion and those without invasion.
|
Diagnostic Statistics
A logistic regression model revealed that the soft tissue in the cartilage
(p < 0.001), intraluminal mass (p < 0.001), and tumor
circumference of the trachea of 180° or greater (p = 0.001) were
the only significant factors for predicting tracheal invasion. Of the three
factors, soft tissue in the cartilage was most accurate (87% accuracy) with
91% sensitivity and 84% specificity (Table
2). There were no false-positive diagnoses for intraluminal mass
(100% specificity), but its sensitivity was low (48%). With tumor
circumference, the highest accuracy of 81% with 100% specificity was attained
with 180° or greater, but its sensitivity was 43%. Using the three
predictors, combined criteria of soft tissue in the cartilage, intraluminal
mass, or a tumor circumference of 180° or greater, produced the greatest
accuracy (90%) with 100% sensitivity and 84% specificity. Using this protocol,
criteria produced seven false-positive diagnoses.
|
Surgical—Pathologic Findings in Correlation with MR
Findings
Intraluminal mass was found in 11 (16%) of the 67 patients on MR images.
All 11 patients had coexistent soft-tissue signal in the cartilage, and
pathologic findings showed tumor invasion in the submucosal areas (n
= 7) (Fig.
2A,2B,2C,2D)
or in the mucous membrane (n = 4).
|
|
|
|
Soft-tissue signal within the tracheal cartilage was detected in 28 patients (42%) on T2-weighted images with (n = 11) or without (n = 17) associated intraluminal mass. Contrast-enhanced T1-weighted images revealed soft-tissue signal within the tracheal cartilage in 24 (86%) of the 28 patients. Of the 17 patients without associated intraluminal mass, tumor invasion was identified pathologically in six patients (35%; two each had submucosal, intercartilaginous, and adventitial invasion) or surgically in four patients (24%). Seven patients (41%) had no tracheal invasion. Of the 23 patients having tracheal invasion, cartilage invasion was pathologically identified only in two patients (9%); both patients had intraluminal masses that were visable on MR images and coexistent submucosal tumor extension was seen on pathologic studies.
Of the 28 patients with soft-tissue signal in the cartilage, 24 patients underwent contrast-enhanced MR imaging. Of these 24 patients, synchronous enhancement of the soft tissue in the cartilage with thyroid carcinomas was seen in 14 patients (58%). All 14 patients had pathologic findings of tracheal invasion (Fig. 3A,3B,3C,3D). Although no synchronous enhancement was found in the other 10 patients, tracheal invasion was identified in six (60%) of the 10 patients.
|
|
|
|
Tumor circumference of 180° or greater was seen in 10 patients (15%), and all 10 had pathologic (n = 7) or surgical (n = 3) evidence of invasion to the trachea. The tumor reached submucosal areas in two patients, intercartilaginous areas in two, and adventitia in three. Nine of the 10 patients had coexistent intraluminal mass or soft-tissue signal in the cartilage; although no such coexistent MR findings were seen in the remaining patient, at pathology this patient proved to have adventitial invasion (Fig. 4A,4B,4C,4D).
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this study, we performed MR imaging in a cadaveric trachea with a thinner slice section and a smaller field of view than routine MR studies to obtain a detailed MR view of the healthy trachea in correlation with histologic findings. Next, we evaluated MR features of the trachea in the healthy control subjects as a standard of reference. Most of the healthy tracheae showed a horseshoe or a circular shape. Tracheal walls consisted of two layers with different signal intensity; each layer was approximately 1-mm thick in axial MR scans. The inner layer corresponding to the mucous membrane and submucosal tissue showed higher signal intensity with enhancement on contrast-enhanced MR images, whereas the outer layer comprising cartilage rings, intercartilaginous membrane, posterior membranous portion, and adventitia showed lower signal intensity without discernible enhancement. The adventitia was too thin to identify on MR images.
Although most of the healthy tracheal cartilage showed homogeneous low-signal intensity in axial T2-weighted images, some had small areas of higher signal intensity in the low-signal structures. In the cadaver study, cartilage rings of 2-5 mm in vertical length having lower signal intensity appeared alternately with intercartilaginous membranes of 0.5-1 mm having higher signal intensity. Some of the tracheal rings were incomplete, fragmentary, or bifurcated. Axial MR scanning planes were not necessarily parallel to the planes of the tracheal rings. A 5-mm-thick section used in our clinical study contained at least one pair of the tracheal ring and intercartilaginous membrane. Therefore, we think that the MR appearance of the tracheal rings was a product of volume averaging and that the signal intensity depends on the proportion of these two components of different signal intensity in the slice section.
In our series, soft-tissue signal in the cartilage, intraluminal mass, and the degrees of tumor circumference to the trachea were the only statistically significant predictors for tracheal invasion on MR images. Our study, like others documented in the literature, showed that the presence of intraluminal mass was highly specific to tracheal invasion (100% specificity). This finding always indicated deep tumor infiltration, either in the submucosal areas or in the mucous membrane, in pathologic studies. However, the sensitivity was low (48%), because more superficially confined invasion and a subtle submucosal invasion were missed with this MR finding.
Soft-tissue signal in the cartilage showed relatively high accuracy (87%) for tracheal invasion in our series. However, this MR feature represented actual tumor invasion of the cartilage in only 7% (2/28) of the patients with this sign. The other patients with this MR finding proved to have tumor infiltration to the intraluminal (14%, 4/28) or the submucosal (32%, 9/28) areas or intercartilaginous membrane (7%, 2/28) or to have adventitial invasion (11%, 3/28) in pathologic specimens. Tracheal invasion was discovered at surgery in three patients. When thyroid carcinomas infiltrated to the trachea, they usually invaded the intercartilaginous membranous portions and widened the space between the tracheal rings. We think that as the proportion of the voxels of the tumor of higher signal intensity increased in relation to the voxels of the cartilage rings of lower signal intensity, localized high-signal-intensity areas within the low-signal-intensity structures became more distinct on T2-weighted images and contrast-enhanced T1-weighted images. We think these two pulse sequences can provide enough information for soft-tissue differentiation between tumor and healthy trachea.
The prevalence of the soft-tissue signal in the cartilage without tracheal invasion (7/67 patients, 10%) that caused false-positive diagnoses in our series of thyroid carcinomas was not statistically significantly different (p = 0.733) from that found in the healthy control subjects (4/30 subjects, 13%). We think a false-positive fraction of approximately 10% is unavoidable for this criterion with present scanning technique. This finding also included two false-negative cases because a subtle invasion of the intercartilaginous membranous portions or adventitia was not sufficient to change the MR signal intensity. However, the soft tissue in the cartilage with synchronous enhancement with thyroid carcinomas was highly specific to tracheal invasion and invariably indicated tracheal invasion by the tumors. Thus, we suggest that contrast-enhanced MR imaging should be performed in patients with findings suggestive of tracheal invasion on unenhanced MR images. In the future, such a study of MR imaging with the use of higher spatial resolution and thinner section thickness should be attempted to further clarify the characteristics of the soft-tissue signal in the cartilage and to better differentiate tumor invasion from the healthy tracheal cartilage.
Although sensitivity (43%) of the sign was low, a tumor circumference of 180° or greater was highly specific (100% specificity) and always indicated tracheal invasion regardless of the depth of the tumor invasion of the trachea. This MR appearance without intraluminal mass and without soft tissue in the cartilage likely correlates with adventitial tumor invasion. We combined the predictors of soft tissue in the cartilage, intraluminal mass, and a tumor circumference of 180° or greater. When any one of these combined criteria was positive, we achieved an accuracy of 90% with 100% sensitivity and 84% specificity. This criterion included seven false-positive cases because of the presence of soft tissue signal in the cartilage. Four of these seven patients had contrast-enhanced MR images; no synchronous enhancement of the areas of the soft tissue signal was found in any of the four patients. Therefore, we think that contrast-enhanced MR imaging may improve the specificity of the optimal criteria.
Even though the number of patients in this study having tracheal invasion in each layer is relatively small, we can summarize our observations as follows: Tumor circumference of 180° or greater around the trachea in the absence of other MR findings indicated adventitial invasion of the trachea; intraluminal mass represented tumor invasion into deep layers of the trachea, either in the submucosal areas or in the mucous membrane; and the soft-tissue signal in the cartilage indicated healthy trachea or tumor invasion of either submucosal areas or cartilage or of intercartilagenous membrane or the adventitia.
In conclusion, although the prediction with MR imaging of the precise depth of tumor invasion to the trachea was unreliable, presence or absence of tracheal invasion was accurately diagnosed with this procedure. Using this technique will contribute to surgical planning and prediction of prognosis of patients with thyroid carcinomas.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  D.M. Yousem, K. Gad, and R.P. Tufano Resectability Issues with Head and Neck Cancer AJNR Am. J. Neuroradiol., November 1, 2006; 27(10): 2024 - 2036. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Takashima, F. Takayama, J. Wang, S. Kobayashi, and M. Kadoya Using MR Imaging to Predict Invasion of the Recurrent Laryngeal Nerve by Thyroid Carcinoma Am. J. Roentgenol., March 1, 2003; 180(3): 837 - 842. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
