| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Head and Neck Imaging |
1 From the Endocrine (M.L.D.F., C.B., M.S., M.L.B.), Radiology (S.C., A.T.), Surgery (D.B., P.C., F.T.), and Nuclear Medicine Units (G.B., L.V.) of the Department of Clinical Physiopathology, and the Pathologic Anatomy Unit (A.A.), University of Florence, Viale Pieraccini 6, 50139 Florence, Italy. Received January 14, 1998; revision requested April 2; final revision received April 28, 1999; accepted June 1. Supported in part the Associazione Italiana per la Ricerca sul Cancro (AIRC). Address reprint requests to M.L.B. (e-mail: m.brandi@dfc.unifi.it).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: First, computed tomographic (CT), magnetic resonance (MR), ultrasonographic (US), and technetium 99m methoxyisobutyl-isonitrile (MIBI) scintigraphic images in 49 patients with primary hyperparathyroidism were retrospectively evaluated. A single-blind, prospective study that included 16 patients with primary hyperparathyroidism was then conducted. MR, US, scintigraphic, and color Doppler US images of the neck were obtained and analyzed.
RESULTS: In the retrospective study, CT, MR imaging, and US had low sensitivity (13%, 17%, and 27%, respectively) and specificity (39%, 65%, and 65%, respectively). Scintigraphy had 57% sensitivity and 85% specificity. In the prospective study, the use of latest-generation MR and US equipment and the participation of experienced operators led to improved sensitivity and specificity for these techniques. The combination of US and scintigraphy resulted in improved sensitivity (96%), specificity (83%), and positive and negative predictive values (88% and 94%, respectively), relative to the results obtained with either method alone. Doppler US was of little help in the setting of small glands.
CONCLUSIONS:
The combination of 99mTc MIBI scintigraphy and US performed by well-trained operators with up-to-date instruments appeared to be the best diagnostic tool for the preoperative diagnosis of parathyroid disease.
Index terms: Parathyroid, CT, 274.1211 • Parathyroid, hyperparathyroidism, 274.531 • Parathyroid, MR, 274.121411, 274.12143 • Parathyroid, neoplasms, 274.363 • Parathyroid, radionuclide studies, 274.12175 • Parathyroid, US, 274.12981, 274.12983
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Indeed, the results of many studies (6–8) have clearly indicated the low accuracy of the different methods, which include ultrasonography (US), scintigraphy, and computed tomography (CT); therefore, some authors (6) have suggested that surgery is the only sensitive approach to aid in detection of an abnormal parathyroid gland. In many of the studies published so far, one or two techniques have been evaluated, often in nonhomogeneous groups of patients. Therefore, the optimal choice and sequence of diagnostic procedures, which may necessitate the use of sophisticated, expensive techniques and well-trained operators, is not clear. The lack of specific diagnostic criteria for identification of parathyroid glands on scintigraphic and other types of radiologic images contributes to the divergent views reported in the literature (6–15) on the utility of preoperative diagnostic procedures.
In the first part of this study, we retrospectively evaluated cases of pHPT diagnosed and surgically treated at our center. Our analysis was limited to findings in patients with abnormal parathyroid glands localized in the neck and treated only once with surgery. We performed blinded interpretation of images obtained with at least two of the following modalities: US, CT, magnetic resonance (MR) imaging, and technetium 99m 2–methoxyisobutyl-isonitrile (MIBI) scintigraphy. The aims of the study were (a) to assess the utility of routine diagnostic procedures that are readily available at any medical center with respect to surgical success rate in uncomplicated cases of pHPT and (b) to evaluate the bias associated with each procedure, according to the surgical and histopathologic results, to determine radiologic signs that may be helpful for parathyroid gland localization.
On the basis of the results of the retrospective analysis, a group of patients were included in the second part of our study, which was a single-blind, prospective investigation. The aims of the second part of the study were (a) to determine specific criteria that may be useful for detection (ie, presence of a nodule) and characterization (ie, parathyroid nature of a nodule) of abnormal parathyroid glands and for volumetric assessment of nodules; (b) to evaluate the importance of operator experience in evaluating parathyroid disease; and (c) to analyze the potential of up-to-date US and MR imaging and standardized scintigraphic methods with respect to equipment and protocols. No effort was made to select specific equipment because the aim of this part of the study was to acquire information on the utility of examinations performed at various centers.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
All selected patients had been evaluated at different radiologic and nuclear medicine university centers and by different operators (including C.B., S.C., A.T., G.B.), who used at least two of the following methods: US (n = 47), CT (n = 27), MR imaging (n = 24), and 99mTc MIBI scintigraphy (n = 36). As a standard procedure, the patients were referred for the different techniques at the same time, with the only indication to the operators being the presence of biochemical and clinical signs of pHPT. The examinations were performed in random order, and the resultant images were interpreted independently and in blinded fashion by each operator; a clinician (M.L.D.F.) analyzed the results of the examinations at the same time (at the end of the diagnostic procedures), with no further discussion with the operators.
Prospective study.—All patients with pHPT diagnosed at our endocrine unit from January 1995 to December 1995 were enrolled in a prospective study. Written informed consent was obtained, and the study protocol was approved by the institutional committee. One year after surgery, 16 women (age range, 54–71 years; mean age, 61 years ± 7) met the five inclusion criteria (same criteria as for the retrospective study).
In all patients, surgical findings confirmed the presence of one or more abnormal parathyroid glands in the anterior region of the neck, with one exception—a woman with a diagnosis of "unidentified" (Table 1). This patient had biochemical features of pHPT, a huge goiter, a history of irradiation in the neck region because of chronic pharyngitis (30 years earlier), and US and MR imaging evidence suggestive of abnormal parathyroid tissue in the neck. At surgery, no abnormal parathyroid gland was found, and on images, the lesion diagnosed as a parathyroid nodule appeared to be a lymph node. However, similar to the other patients, calcium and parathyroid hormone levels decreased to normal after surgery and remained in the normal range for up to 1 year. We concluded, therefore, that pathologic parathyroid tissue may have been present and was removed at surgery but that identification of such tissue was hampered by the dimensions of the goiter and the consequences of irradiation of the region.
|
Retrospective Study Imaging Techniques
CT studies.—CT was performed with advanced second- and third-generation CT scanners (Tomoscan 350, Philips Medical Systems, Eindhoven, the Netherlands; CT 8000, GE Medical Systems, Milwaukee, Wis). Scanning was performed before and after intravenous administration of iodinated contrast medium (iohexol, Omnipaque 300, Schering, Berlin, Germany; iopamidol, Iopamiro 300, Bracco, Milan, Italy). Contiguous 2- or 5-mm-thick sections were obtained at 250–400 mAs.
MR imaging studies.—MR images were obtained with a 0.5-T superconducting magnet (Gyroscan; Philips Medical Systems) equipped with a head coil or collar-shaped surface coil. Spin-echo T1-weighted (480–600/12–18 [repetition time msec/echo time msec]) and T2-weighted (1,700–2,1,700–2,100/90–100) images were obtained. Spin-echo T1-weighted images were also obtained after intravenous administration of a paramagnetic contrast medium (gadopentetate dimeglumine [0.2 mmol/kg], Magnevist; Schering). Contiguous 5-mm-thick T1-weighted and 8-mm-thick T2-weighted transverse and sagittal multisection studies were obtained from the angle of the mandible to the clavicle.
US studies.—US was performed with electronic 5- and 7.5-MHz transducers with a multicrystal linear array (model AU 450, Esaote Biomedica, Genoa, Italy; model SSD650, Aloka, Tokyo, Japan). Transverse and sagittal scans were obtained from the level of the mandibles to the medial end of the clavicles.
Scintigraphic studies.—Scintigraphy was performed by means of rapid intravenous administration of 99mTc MIBI (240 mBq). Images were obtained at 10 and 90 minutes after administration of 99mTc MIBI. The image field included the retromastoid region superiorly and the sternal xiphoid process inferiorly. A static acquisition for 10 minutes was performed with a circular gamma camera and a lowenergy parallel-hole collimator (matrix, 128 x 128; zoom acquisition factor, 1.5).
Prospective Study Imaging Techniques and Diagnostic Criteria
MR imaging studies.—MR images were obtained with 0.5-T unit (Gyroscan T5-II/3; Philips Medical Systems) with a 12-mT gradient by using a surface coil or quadrature receive coil. Contiguous T1-weighted turbo spin-echo images (450–650/12–15 [repetition time msec/effective echo time msec) were obtained before and after intravenous administration of a 0.1–0.2 mmol/kg bolus of gadopentetate dimeglumine. Dynamic studies were performed by using the Key-Hole technique (Philips Medical Systems), with acquisitions at 0, 25, 50, 75, 125, and 250 seconds after administration of contrast material (2,205/11–14, 30° flip angle). T2-weighted turbo spin-echo images (3,000–3,000–3,400/100–120 [effective], 5-mm section thickness) also were obtained. MR imaging was performed in the transverse, sagittal, and coronal planes. The neck region was imaged from the mastoid processes to the clavicle.
On MR images, a parathyroid gland was considered to be abnormal when the signal intensity pattern of the suspected soft-tissue lesion was different from that of thyroid parenchyma, thyroid nodules, or lymph nodes. A typical pattern of parathyroid signal intensity could not be identified on the basis of data from the retrospective study. Location of a soft-tissue mass in the retrothyroid region, within the dorsal part of the thyroid, or caudal with respect to the gland contributed to the confidence of the diagnosis, as did a location close to the carotid artery sheath. Normal parathyroid glands were not detectable. For the dynamic MR studies, a parathyroid nature also was postulated for nodules with early and transient contrast enhancement, which was suggestive of arterial vascularization.
US studies.—US of the anterior portion of the neck was performed with a real-time B-mode system equipped with a 7.5-MHz linear-array probe (model AU 560; Esaote Biomedica) and a mechanically moving 10-MHz scan-head transducer oscillating in an oil bath (AU 420; Esaote Biomedica). The transverse resolution was 0.5 mm, the lateral resolution was 1–2 mm, the field of view was 3 x 5 cm, and the depth of penetration was 4–5 cm. Transverse and sagittal scans were obtained from the level of the mandible to the clavicle, with the patient's neck hyperextended. Acoustic gel was used.
Parathyroid glands were measured in three dimensions, and the US appearance was analyzed with respect to echotexture, homogeneity of the inner structure, configuration, and location. US images were considered to be positive for a parathyroid nodule when the images depicted a relatively hypoechoic (with respect to thyroid nodules) or anechoic homogeneous area with an oval shape, regular borders, aligned in a craniocaudal direction, and located in the retrothyroid region within the posterior portion of thyroid parenchyma or in proximity to the carotid artery sheath. In larger nodules, a lack of homogeneity, cystic changes, necrotic areas, or a tendency toward a round shape also were evident (6).
Color Doppler US studies.—Color Doppler US was performed with a 7.5-MHz linear-array transducer (model AU 590 Asynchronous; Esaote Biomedica) (the power module was not available) with a pulse repetition frequency of 500–800 Hz and a color gain of 5–15 Hz. More intense color signals were also evaluated by using pulsed-wave spectral analysis. The criteria for parathyroid lesions were as follows: (a) a polar distribution of color with an arterial pattern and virtually no perinodular peripheral distribution and (b) a diffuse intraparenchymal distribution of color signals without an arterial-pole or perinodular distribution.
Scintigraphic studies.—The patients were positioned supine with the head and neck extended and immobilized. A bolus containing 240 mBq of 99mTc MIBI (Cardiolite, Du Pont–Pharma, Brussels, Belgium; Elumatic, Oris-CEA, Gif-sur-Yvette, France), prepared according to the manufacturer's instructions, was injected intravenously. A 10-minute dynamic acquisition (20 frames) was performed. The acquisition began immediately after the bolus injection and at 120 minutes. We used a circular gamma camera and a low-energy parallel-hole collimator (matrix, 128 x 128; zoom acquisition factor, 1.5). The image field included the retromastoid region superiorly and the sternal xiphoid process inferiorly. The differential washout properties of MIBI lead to a marked change in the uptake ratio of parathyroid lesions relative to that in thyroid tissue between early (10–15-minute) and delayed (2-hour) images (16), which allows identification of abnormal parathyroid glands in cases of coexistent thyroid disease. Abnormal ectopic parathyroid glands can also be visualized on early images. We increased the delayed data acquisition time to 120 minutes, rather than the standard 90 minutes as was used in the retrospective study, to improve demonstration of parathyroid tissue in patients with a slightly slower washout time and coexistent thyroid disease.
Verification of Imaging Results
The results of the preoperative localization procedures were correlated with the findings at surgical exploration. Data from each technique were analyzed in relation to histopathologic findings (adenoma or hyperplasia), anatomic location in the neck, gland size, and coexistent thyroid disease. Pitfalls in detection (presence vs absence of a nodule) and characterization (parathyroid origin of a nodule) were considered separately. A study was classified as true-positive if the defined abnormality corresponded to the surgical and histopathologic findings for exact location, size, and origin. False-negative results were obtained when a nodule could not be detected and/or pathologic parathyroid tissue was erroneously characterized as thyroidal or nodal in origin. False-positive results were obtained when a nodule was detected but there was no surgical confirmation of its presence or when a nodule was erroneously characterized as a pathologic parathyroid gland.
Surgical evaluation.—Detailed descriptions of the numbers, anatomic sites, and sizes of parathyroid and nonparathyroid nodules were obtained from surgical reports. For each gland, the diameter was measured along three axes; gland volume was determined by using the formula (
/6)xyz for an ellipsoid with diameters x, y, and z in the three orthogonal planes. The longest diameter was used for correlation with imaging studies because of the usually ellipsoid shape of the glands.
Histopathologic evaluation.—Paraffin-embedded sections stained with hematoxylin and eosin were examined without knowledge of imaging results. Clinical history, laboratory data, and surgical notes were taken into account for the final diagnosis of parathyroid disease. Briefly, criteria for a diagnosis of adenoma included the presence of a single enlarged gland together with evidence of three normal-sized glands and at least a second (and often a third) histologically normal gland. The diagnosis of hyperplasia was based on the presence of at least two enlarged glands or one enlarged gland and a second normal-sized or minimally enlarged gland with abnormal cellularity (17).
Statistical Analyses
For the retrospective and prospective studies, true-positive, true-negative, false-positive, and false-negative findings were assessed for all localization methods. The accuracy, sensitivity, specificity, positive predictive value, and negative predictive value for parathyroid and nonparathyroid nodular lesions also were determined. Other statistical tests were not performed for the retrospective study owing to heterogeneity of the data.
The Fisher exact test (18) was performed with data from the prospective study to compare rates between different imaging methods. There was no correction for simultaneous pairwise comparisons.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The sensitivity and specificity for CT, MR imaging, and US were lower than those reported by some authors (6,7,10), whereas 99mTc MIBI scintigraphy showed a better sensitivity (57%) and specificity (85%), with a positive predictive value of 91% (Table 2).
|
|
The results obtained with the three imaging techniques are shown in Table 3. All techniques had good sensitivity and specificity. US and scintigraphy had approximately equivalent sensitivity, specificity, accuracy, and positive and negative predictive values; these values were lower for MR imaging, although the differences were not significant. The characteristics of the parathyroid nodules studied with the three methods are shown in Figure 2.
|
|
False-negative US images (eight of 24 nodules) consisted almost exclusively of characterization errors, with the exception of one case of an ectopic nodule located beneath the carotid artery sheath. In fact, two nodules in one patient were intrathyroidal in location, mimicking a goiter; two nodules in a patient with chronic autoimmune thyroiditis were hyperplastic; two retrothyroid nodules were associated with a huge nodular goiter; and one lesion was a large cystic and necrotic nodule (Table 4). Negative nodules were associated mainly with hyperplastic glands (six of eight nodules). Five (62%) of the eight nodules that were not characterized were approximately 2 cm in longest diameter; however, of the 16 nodules correctly identified at US, 12 (75%) were small (longest axis 
2 cm; in six of these, longest axis 1 cm), which thus suggests that for US, lesion size was not a limiting factor. The only false-positive US finding of a nodule was that of a retrothyroid lymph node in a patient with goiter, thyroiditis, and a history of neck irradiation. Of the eight nodules for which US findings were false-negative, seven were detected at scintigraphy.
|
Doppler US results could not be evaluated in one patient with parathyroid hyperplasia, nodular goiter, and thyroid carcinoma and in one patient with an ectopic parathyroid nodule that was not detected at US. Of 35 nodules studied, 21 were of parathyroid origin, 13 were of thyroid origin, and one was a lymph node. In 12 (57%) of 21 parathyroid nodules examined with this technique, Doppler signals could not be detected; these 12 nodules included three cases of adenoma and nine cases of hyperplasia. In two (10%) of the 21 nodules, Doppler signals were present but were not specific. In the remaining seven (33%), we could identify the characteristic pattern that has been described in the literature (10) and previously observed by us. Six of these seven lesions were characterized as adenomas, and four of these, as well as the only hyperplastic nodule in this group, had a longest axis that was more than 1 cm. Nine of 12 parathyroid nodules with no Doppler signals were characterized as hyperplastic lesions (four had a longest axis of <1 cm). The presence of thyroid disease did not influence the results. One false-positive finding of a nodule was confirmed to be a lymph node. In conclusion, Doppler US had low sensitivity (33%) and accuracy (57%) but high specificity (93%) and positive predictive value (87%).
The association of more than one technique (defined as results of analyses of techniques in concordance with the correct diagnosis) resulted in a marked although nonsignificant reduction in sensitivity (42% reduction for MR imaging and scintigraphy and for US and scintigraphy, 33% reduction for US and MR imaging, 29% reduction for MR imaging, US, and scintigraphy). The increase in specificity that could be obtained by association of two techniques was small and nonsignificant, owing to the high specificity achieved with each method alone (78% for MR imaging, 94% for US, and 89% for scintigraphy).
In consideration of the relatively high sensitivity and specificity of each technique used alone, statistical analysis of the combination of more than one technique (pathologic parathyroid nodule localized with at least one of the two or three studies) was performed to assess whether an improvement in any of the evaluated parameters could be achieved. Results indicated that the combination any one method with one or two of the other methods resulted in an increase in sensitivity (79% for MR imaging and US vs 50% for MR imaging alone [P < .02]; 75% for MR imaging and scintigraphy vs 50% for MR imaging alone [P < .05]; 96% for US and scintigraphy vs 67% for US alone [P < .05], 71% for scintigraphy alone [P < .02], and 50% for MR imaging alone [P < .001]).
The specificity of the combination of US and scintigraphy (83%) was significantly higher than that obtained with combinations of either MR imaging and US (P < .05) or MR imaging and scintigraphy (P < .01). The combination of US and scintigraphy also resulted in increases in the accuracy (90% vs 67% for MR imaging and US and 59% for MR imaging and scintigraphy [not significant]), positive predictive value (88% vs 68% for MR imaging and US [not significant] and 62% for MR imaging and scintigraphy [P < .03]), and negative predictive value (94% vs 54% for MR imaging and scintigraphy [P < .02]); therefore, the combination of US and 99mTc MIBI scintigraphy appeared to be the most reliable technique.
The pitfalls associated with US and scintigraphy in cases where a positive result was obtained with one technique (either false-positive or false-negative) are shown in Table 4. The combination of US and scintigraphy led to a correct diagnosis of 23 of 24 nodules. Five of six cases where US was positive and scintigraphy was negative involved nodules with all three axes smaller than 10 mm (the sixth case was of a 10 x 10 x 20-mm fusiform nodule). Positive scintigrams allowed diagnosis of one ectopic gland not detected at US and six nodules not characterized at US (size, 5–35 x 5–15 mm). The only case in which both techniques provided negative results proved to be a large hyperplastic gland with wide cystic and necrotic areas in a patient with four large hyperplastic glands mimicking a goiter; two other parathyroid glands in this patient also were not characterized at US owing to findings suggestive of a thyroid goiter. Errors in US characterization appeared to be due to intrathyroid location or associated thyroid disease. Positive US and negative scintigraphic findings of nodules were associated with small gland size (n = 5) or a coexistent thyroid nodule overlying the parathyroid nodule (n = 1); all were correctly diagnosed at US. False-positive findings of a parathyroid gland included a lymph node (US characterization error) and goiter with or without a surgically proved nodule.
The combination of MR imaging and US or MR imaging and scintigraphy did not result in a significant increase in any of the parameters evaluated, as compared with US alone or scintigraphy alone (Table 3), which suggests that MR imaging was less useful in these combinations.
The combination of all three techniques did not lead to significant increases in sensitivity, specificity, accuracy, or positive or negative predictive values, as compared with the increases observed with the combination of US and scintigraphy (Table 3).
The association of thyroid disease and location in the neck appeared to be relevant for localization techniques, in particular for US (four of eight negative parathyroid nodules were intrathyroid or ectopic, and thyroid disease was slightly more frequent).
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Thyroid disease was more frequently associated with nodules not detected at CT or MR imaging than with correctly identified nodules (Fig 1), which suggests that the presence of thyroid disease contributed to the low sensitivity and specificity of these techniques. Parathyroid nodules not detected at CT and MR imaging were mainly retrothyroid in location, which indicates that the anatomic location was less of an influence on lesion detection than was nodule size.
Scintigraphy had better specificity and positive predictive values. As for CT, MR imaging, and US, parathyroid hyperplasia was more often present in the group of nondetected nodules (17 of 23). Indeed, true-positive nodules usually were larger (21 of 30 with longest axis > 1 cm), and most (20 of 30) were shown to be adenomatous lesions. Coexistent thyroid disease appeared not to influence the interpretation of the scintigraphic image. In conclusion, the results of our retrospective analysis showed an unsatisfactory sensitivity and specificity for all of the techniques investigated; this was especially evident for US, CT, and MR imaging, particularly in the absence of specific and extensive operator experience with parathyroid gland evaluation.
To determine accurate and reliable diagnostic criteria for parathyroid tissue that could be adopted for complicated cases with lower surgical success rates (eg, recurrent or persistent pHPT at second intervention or previous surgery for thyroid disease), the errors in detection and/or characterization observed in this retrospective analysis were used to outline a protocol for a single-blind prospective study. To improve US specificity (improve correct characterization of parathyroid nodules), Doppler US was combined with gray-scale US. CT was not included because the CT results obtained in the retrospective study showed low sensitivity, specificity, and positive and negative predictive values, similar to those of MR imaging (Table 2), and the use of CT added radiation dose to the patients.
In comparison with the results of the retrospective study, 99mTc MIBI scintigraphy in the prospective study showed increases in sensitivity, accuracy, and negative predictive value and similar specificity and positive predictive value, with no differences between adenoma and hyperplasia among the positive and negative nodules. In five of seven negative parathyroid nodules and in five of 17 positive nodules, the longest axis was less than 1 cm, with no differences in plasma calcium and parathyroid hormone levels between these two groups (data not shown). Interestingly, three of the five nodules that were negative at scintigraphy were in two patients with hyperplasia and at least one other parathyroid nodule that was positive at scintigraphy; this result suggests that glandular metabolic activity may influence the detection limit of scintigraphy. The role of coexistent thyroid disease was not calculated, owing to the small number of false-positive and false-negative results in this series. The two false-positive findings were of a thyroid nodule in nodular goiter and a case of chronic autoimmune thyroiditis with no surgical and histologic confirmation of the presence of a parathyroid nodule. The possibility of uptake of 99mTc MIBI by a hyperplastic thyroid nodules and Hürthle cell adenoma has been demonstrated (11).
US showed better results in the prospective study than in the retrospective study. Indeed, the sensitivity, specificity, accuracy, and positive and negative predictive values were in the same range as those for scintigraphy, although no specific echotextural patterns were recognizable, and identification was reliant on the site of the lesion. The major pitfalls of US were associated with the detection of ectopic nodules, the characterization of intrathyroid nodules, or the presence of thyroid disease. The use of the 10-MHz transducer resulted in an increase in the total number of detected nodules but did not provide any additional information for characterization.
The limitations of US do not appear to be related to the small size of a parathyroid nodule but rather to characterization of such a nodule. Therefore, we included Doppler US analysis in this study. In cases where Doppler signals were present, the specificity of this technique was high (93%, with only one false-positive finding due to a lymph node), although the number of parathyroid nodules analyzed for Doppler signals was too small for us to draw final conclusions. Interestingly, the specific Doppler signals seemed to be more often represented in parathyroid adenomatous lesions (six of seven nodules) with a longest axis that was greater than 1 cm; therefore, this technique was not helpful, at least in our investigation, in cases of smaller glands. In fact, 12 of 21 nodules in our series showed no detectable signals, and only two of eight nodules with a longest axis that was more than 1 cm showed a specific pattern.
MR imaging showed an increase in all of the parameters evaluated relative to those from the retrospective study; in particular, sensitivity increased from 17% to 50%, probably due to the decrease in image acquisition time. However, the spatial resolution of MR images of the neck is undoubtedly lower than that of US images (leading to the possibility of detection errors), and the sharp contrast resolution that can be obtained does not appear to markedly improve characterization of parathyroid nodules, owing the variability in signal intensity among parathyroid glands and nodules, thyroid nodules, and lymph nodes. Moreover, the results of dynamic contrast-enhanced MR imaging did not lead to a significant increase in the number of nodules detected or improvement in characterization. Indeed, the specificity and positive predictive value were higher with respect to those statistics in the retrospective study but lower than those reported for US and 99mTc MIBI scintigraphy (although not significantly, owing to the small number of cases).
The combination of two techniques that both had high sensitivity and specificity resulted in an increase in sensitivity. In particular, the combination of US and scintigraphy resulted in the best balance between the parameters evaluated (Table 3). In fact, with scintigraphy, it was possible to characterize nodules not characterized at US and to detect ectopic glands, whereas with US, it was possible to detect lesions that may have been overlooked at scintigraphy owing to their small dimensions, low metabolic activity, and/or the presence of necrotic or cystic areas (Table 4).
In contrast to the combination of US and scintigraphy, the combination of MR imaging with US or with scintigraphy led to a significant increase in sensitivity with respect to MR imaging alone (P < .02) but did not influence the results obtained with US or scintigraphy. This reinforces the hypothesis of the low utility of MR imaging for demonstration of neck lesions. Finally, the combination of US, scintigraphy, and MR imaging did not improve the quality of data relative to results obtained with the combination of US and scintigraphy.
From the analysis of these findings, a list of the limitations for each localization technique used alone was determined (Table 5). Figure 3 depicts a potentially useful model for localization of parathyroid lesions, with 99mTc MIBI scintigraphy of the neck and mediastinum as the first approach. Neck US should always follow, whether the scintigrams are positive or negative, to rule out false-positive and false-negative results and in the setting of multiglandular disease. In cases of negative scintigrams, accurate US evaluation of the causes of error should facilitate correct selection of other methods, thus helping to optimize criteria for detection and characterization, even in cases of small lesions.
|
|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, M.L.B.; study concepts, M.L.B., S.C., C.B.; study design, M.L.B., S.C.; definition of intellectual content, M.L.D.F., S.C.; literature research, M.L.D.F.; clinical studies, M.L.D.F., D.B., P.C., F.T., A.A., G.B., L.V., A.T.; data acquisition, M.L.D.F.; data analysis, M.L.D.F., C.B., S.C.; statistical analysis, M.L.D.F., S.C.; manuscript preparation and editing, M.L.D.F., S.C.; manuscript review, M.L.B., M.S.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
