Neuroradiolog y / Head and Neck Imaging Original Research

Mihailovic et al. Detection of Lymph Node Recurrences in Thyroid Carcinoma Neuroradiology/Head and Neck Imaging Original Research

Downloaded from www.ajronline.org by 113.197.110.71 on 07/15/13 from IP address 113.197.110.71. Copyright ARRS. For personal use only; all rights reserved

MRI Versus 131I Whole-Body Scintigraphy for the Detection of Lymph Node Recurrences in Differentiated Thyroid Carcinoma
Jasna Mihailovic1 Mladen Prvulovic2 Miodrag Ivkovic 3 Branko Markoski 3 Dobrivoje Martinov 3
Mihailovic J, Prvulovic M, Ivkovic M, Markoski B, Martinov D

Keywords: differentiated thyroid carcinoma, lymph node, MRI, recurrence, whole-body scintigraphy DOI:10.2214/AJR.09.4172 Received December 23, 2009; accepted after revision April 14, 2010. This work was presented in part at the 13th International Society of Magnetic Resonance Meeting, Miami Beach, FL, May 2005.
1 Department of Nuclear Medicine, Oncology Institute of Vojvodina, Institutski put 4, Sremska Kamenica 21204, Serbia. Address correspondence to J. Mihailovic (jasnans@eunet.rs). 2

OBJECTIVE. The aim of this study was to compare the clinical usefulness of MRI and radioiodine (131I) whole-body scintigraphy for the detection of lymph node metastases in differentiated thyroid carcinoma (DTC). MATERIALS AND METHODS. After surgery and 131I therapy, 40 patients with DTC underwent 131I whole-body scintigraphy and MRI. Each patient was clinically suspected of having or already had evidence of nodal recurrences (conrmed by laboratory studies, cytologic analysis, or whole-body scintigraphy). Planar whole-body scintigraphy was done after administration of 111 MBq of 131I, and MRI was done using spin-echo T1- and T2-weighted imaging, T1-weighted spin-echo imaging with fat suppression, and STIR sequences. RESULTS. MRI detected nodal metastases as partly or entirely cystic and as heterogeneously enhanced on contrast-enhanced T1-weighted images. Hyperintense cystic areas appeared on T1- and T2-weighted images and STIR sequences in 57% of cases. Nodal metastases showed extracapsular spread in 24% of patients. MRI results were true-positive in 76%, true-negative in 90%, false-negative in 24%, and false-positive in 11% of cases, whereas 131I whole-body scintigraphy results were true-positive in 71%, true-negative in 91%, and falsenegative in 29% of cases. There were no false-positive results of 131I whole-body scintigraphy. False-negative whole-body scintigraphy was induced by tumor dedifferentiation. The sensitivity, specicity, accuracy, positive predictive value, and negative predictive value of MRI were 76.2%, 89.5%, 82.5%, 88.9%, and 77.3%, respectively, and the corresponding values for 131I whole-body scintigraphy were 71.4%, 100%, 85%, 100%, and 76%, respectively. CONCLUSION. Whole-body scintigraphy is more specic than MRI in the detection of nodal metastases in patients with DTC. The principal value of MRI is in non-iodine-avid recurrences and in evaluation of mediastinal foci. ifferentiated thyroid carcinoma (DTC) usually has a good prognosis if it is identied during the early stages of disease. Nevertheless, metastases may appear during follow-up. About 2030% of patients develop recurrences over several decades, and twothirds of recurrences appear within the rst decade after initial treatment. Recurrent thyroid cancer occurs most frequently in the cervical lymph nodes. Papillary thyroid carcinoma develops lymph node metastases more frequently than follicular carcinoma does, with an incidence of 2565% [1, 2]. The most important tools for routine follow-up of DTC are diagnostic radioiodine (131I) whole-body scintigraphy, thyroglobulin (Tg) measurements (with or without thyrotropin stimulation), and neck ultrasound [3]. Both stimulated Tg measurements and ul-

MRI Center, Oncology Institute of Vojvodina, Sremska Kamenica, Serbia.

3 University of Novi Sad, Technical Faculty, Mihajlo Pupin Zrenjanin, Zrenjanin, Serbia.

AJR 2010; 195:11971203 0361803X/10/19551197 American Roentgen Ray Society

trasound with ne-needle aspiration (FNA) and cytologic analysis of the aspirate have high sensitivity for detecting nodal metastases [4]. On the other hand, measurement of Tg levels and 131I whole-body scintigraphy have high specicity in the detection of DTC recurrence because of the ability of DTCs to secrete Tg and to concentrate iodine. However, one-third of recurrent DTCs with increasing serum Tg levels lost their iodine avidity and appeared negative on 131I scans [1]. For those patients, ultrasound is the preferred imaging technique for detecting nodal recurrences in the central and lateral neck, but interference with bone and air inuences diagnostic limitations in the mediastinal space [5]. Therefore, additional imaging techniques are necessary for the assessment of nodal metastases. MRI is a diagnostic tool that helps in detecting mediastinal metastatic lymph nodes,

AJR:195, November 2010

1197

Mihailovic et al. especially non-iodine-avid metastases. This modern imaging method is preferable to CT because it allows multiplane evaluation, has better tissue contrast, allows no radiation to reach the neck, and can be performed with a paramagnetic contrast agent that does not interfere with subsequent 131I therapy [5]. MRI of thyroid cancer has been described in the literature [512], along with comparisons between MRI and ultrasound [13] and between MRI and CT [1416]. Most recently, the role of MRI in evaluating nodal recurrences in patients who have undergone thyroidectomy with increasing Tg levels has been reported [17]. However, the potential of MRI versus whole-body scintigraphy in the evaluation of recurrent thyroid carcinoma has not been signicantly investigated. To our knowledge, there are only a few data reported comparing MRI to whole-body scintigraphy in the evaluation of recurrent DTC. MRI was found to be more sensitive than thallium-201 scintigraphy in detecting recurrence of DTC [18]. On the other hand, MRI has been reported to be less specic than 131I whole-body scintigraphy in detecting recurrent or persistent papillary carcinoma. It was recommended that MRI should be indicated, in addition to ultrasound, for patients suspected of having non-iodine-avid neck or mediastinal recurrence before a surgical procedure [19]. FDG PET is an alternative method that could be used in non-iodineavid recurrences, but it cannot supersede 131I whole-body scintigraphy [1]. In this study, MRI was used in addition to 131I whole-body scintigraphy and Tg level measurements as a routine diagnostic tool for regular monitoring of DTC. The aim was to compare the imaging efcacy of MRI versus that of whole-body scintigraphy in the detection of lymph node metastases in DTC. Materials and Methods Patients
Between January 2004 and December 2008, we retrospectively reviewed the medical records of patients with DTC who were clinically suspected of having recurrent disease. There were 40 patients who had previously undergone total thyroidectomy and 131I therapy; they underwent both MRI of the neck and upper mediastinum and 131I whole-body scintigraphy. This study included patients who were being monitored for disease regularly, with laboratory, cytologic, and scintigraphic evidence of lymph node recurrences and clinical ndings of palpable cervical lymph nodes. Scintigraphic evidence included pathologic 131I uptake outside the thyroid bed as determined by 131I whole-body scintigraphy (diagnostic and/or posttherapeutic). Cytologic evidence was based on aspirate obtained by FNA. Laboratory evidence was based on elevated serum Tg level. disease). Chemotherapy was performed for two patients and was combined with radiotherapy for one of those patients. External beam radiotherapy was performed for two patients.

Downloaded from www.ajronline.org by 113.197.110.71 on 07/15/13 from IP address 113.197.110.71. Copyright ARRS. For personal use only; all rights reserved

Laboratory Testing Initial Treatment

A surgical procedure and 131I therapy were considered as initial treatment. A near-total or total thyroidectomy was performed for all 40 patients. Lymph node dissection was performed for 23 patients (unilateral in 11 and bilateral in 12 patients; two of those 12 patients had regional lymph node metastases combined with distant lung metastases). Postoperative diagnosis of DTC was based on the pathologic report. The histologic classication was performed in accordance with World Health Organization guidelines [20]. Surgical treatment was done in different surgical centers in the country, but 131I treatment performed in our institution. Radioiodine ablation (activity of 3.7 GBq of 131I was performed for the patients with cancer staged as N0M0, whereas 5.55 GBq and 7.4 GBq was applied to the patients with cancer staged as N1M0 and N1M1, respectively) performed at a median interval of 9 months (range, 196) after the primary surgery. The initial treatment in our institution is performed according to our protocol of diagnostics, treatment, and control of malignant thyroid tumors [21]. Laboratory tests included measurements of levels of Tg, Tg antibodies, free triiodothyronine, and thyroid-stimulating hormone (TSH). Tg levels less than 5 g/L were considered normal when radioimmunoassay was performed (until April 2006), or less than 1.00 g/L when an electrochemiluminescence immunoassay method was performed using the Roche Elecsys 2010 System (from April 2006 onward).

Diagnostic Imaging Methods

Whole-body scintigraphy Radioiodine whole ody scans were performed after withdrawal of b levothyroxine for a 1 month and after conrmation of TSH levels of greater than 30 mU/L. Patients were advised to adopt a low dietary intake of iodine 15 days before the scan. Planar whole-body scintigraphy and spot images in anterior and posterior view were done with a gamma camera after tracer doses of 111 MBq of 131I were administered. Until November 2004, a single-headed gamma camera (Orbiter 75, Siemens Healthcare) using a parallel iodine medium-energy all-purpose collimator with a scan speed of 10 minutes per frame and matrix size of 128 128 pixels was used. Radioiodine scintigraphy included several partial planar images of the body in anterior view (head and neck, mediastinum and thorax, and abdomen and legs) and posterior thoracic image. Since November 2004, much better technical quality of scans has been obtained with the use of a dual-headed gamma camera (E-cam, Siemens Healthcare) equipped with an iodine highenergy collimator, which replaced the old camera. Scan speed was 6 cm/min for 200-cm scan length with matrix size of 250 1024 pixels. Posttherapeutic 131I whole-body scintigraphy was performed 7296 hours after the 131I treatment. The retrospective interpretations of the whole-body scintigraphy were done independently by two experienced nuclear medicine physicians who were blinded to the results. If Tg serum level was increased without uptake in a metastatic site on the posttherapeutic 131I whole-body scintigraphy, metastases were classied as non-iodine-avid metastases. Negative 131I whole-body scintigraphy was determined if there was no uptake inside and/or outside the thyroid bed, in patients with normal Tg serum level. The interpretation of the 131I uptake in the thyroid bed was difcult and tricky in differentiation between thyroid remnants and nodal foci. We were not strongly convinced that uptake in or near the thyroid bed was always in thyroid remnants. We assumed that normal thyroid remnants and

Follow-Up
Overall, the median clinical follow-up of the patients was 57.6 months (range, 2180 months). After the initial treatment, routine monitoring was performed every 3 months during the rst year, at 6-month intervals thereafter, and annually after 5 years. We did not perform routine ultrasound in follow-up of patients with DTC. Patients with palpable neck lymphadenopathy suspected of having recurrent disease underwent ultrasound-guided FNA and cytologic analysis. Routine diagnostic 131I wholebody scintigraphy was performed 6 months after the initial treatment. Further diagnostic 131I whole-body scintigraphy was performed only for patients suspected to have metastatic lymph nodes or distant metastatic deposits (suspicious by palpation of cervical lymph nodes and/or by elevated Tg levels on hormonal treatment). If recurrences were conrmed (by 131I whole-body scintigraphy, elevated Tg level, or cytologic report), additional surgery or 131I treatment with subsequent posttherapeutic 131I whole-body scintigraphy was performed. Twenty-three patients received additional 131I treatments once or several times after the ablation. Radioiodine treatments were repeated until the 131I whole-body scintigraphy results were negative and Tg was reduced to normal levels or to persistently low levels (in patients with stable

1198

AJR:195, November 2010

Detection of Lymph Node Recurrences in Thyroid Carcinoma

slightly elevated Tg levels (which decreased to normal values after the ablative 131I therapy) were present in patients with previously negative MRI results. The lymph node metastases were suspected if 131I uptake was detected in the neck region in patients with markedly increased Tg levels and positive MRI results. MRIMRI of the neck and upper mediastinum was performed in patients with positive diagnostic 131I whole-body scintigraphy, elevated Tg level, or palpable cervical lymph nodes. Usually, MRI was done before surgery or 131I therapy at intervals no longer than 2 months. However, for eight patients, diagnostic 131I whole-body scintigraphy and MRI were performed 6 months after the 131I treatment after 4 weeks of L-thyroxine withdrawal. Patients with mildly elevated Tg levels who showed the presence of thyroid remnants on diagnostic scinitigraphy and MRI subsequently underwent additional 131I therapy. MRI was performed using a 1.5-T scanner (Avanto Tim, Siemens Healthcare). The following sequences were used: conventional spin-echo T1weighted images in the coronal plane with TR/TE of 824/8.7, eld of view (FOV) of 240 mm, matrix of 256 256, and slice number of 25; T2-weighted images in transversal plane with 5,200/103, FOV of 200 mm, and matrix of 192 256; T1-weighted spin-echo images with fat suppression in the axial section; and STIR sequences in axial and the coronal section. Biplanar contrast-enhanced T1-weighted images were obtained in some patients with the aim of differentiating between thyroid remnants and recurrent disease. Paramagnetic contrast medium (gadopentetate dimeglumine) was applied as an IV bolus injection at a dose of 0.1 mmol/kg. MR characteristics for STIR sequences were 7,000/82, inversion time of 150 milliseconds, FOV of 240 mm, slice thickness of 3 mm, and slice number of 25. Parameters for the T1-weighted sequence with fat suppression were 755/11, FOV of 200 mm, matrix of 192 256, slice thickness of 3 mm, slice number of 25, and distance factor of 10%. Retrospectively, two independent radiologists performed a qualitative analysis of the MR images of the neck and upper mediastinum in a blinded manner. Images were analyzed to evaluate the morphology and signal characteristics of the postoperative and postablative thyroid bed and of metastatic lymph nodes. Images were analyzed by comparing the signal intensity (SI) of lymph nodes with the SI of muscle and fat. The SI of lymph nodes was dened qualitatively as low if the signal was less than or equal to that of muscle, as intermediate if the signal was greater than that of muscle but less than that of fat, and as high if the signal was greater than or equal to that of fat. The qualitative assessment of MRI also included detection of enlarged lymph nodes, growing inltrative lymph node, extracapsular spread of the lymph node, and invasion or displacement of adjacent anatomic structures, such as blood vessels and trachea. The lymphadenopathy was considered as metastatic according to the following criteria: general inhomogeneity of lymph nodes, central necrosis or complete cystic appearance, presence of extracapsular spread, and nodal size (minimal transversal diameter) exceeding 10 mm as recommended by Response Evaluation Criteria in Solid Tumors (RECIST) [22]. Smaller nodes with similar morphologic characteristics or size-criteria nodes without central necrosis or extracap sular spread were dened as suspicious. All other lymph nodes were classied as benign. The nal analysis was made by a separate member of the team and included comparison between MR and 131I images, laboratory tests and histopathologic and cytologic results.

TABLE 1: Characteristics of Patients (n = 40)

Characteristic Sex Female Male Histologic diagnosis Papillary Follicular Hrthle cell carcinoma Metastases at the time of diagnosis No metastases (N0M0) Regional metastases (N1M0) Distant metastases (N0M1) Combined regional and distant metastases (N1M1) 15 (37.5) 23 (57.5) 0 (0) 2 (5) 38 (95) 1 (2.5) 1 (2.5) 25 (62.5) 15 (37.5) No. (%) of Patients

Downloaded from www.ajronline.org by 113.197.110.71 on 07/15/13 from IP address 113.197.110.71. Copyright ARRS. For personal use only; all rights reserved

Statistical Analysis
The sensitivity, specicity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) in the detection of lymph node metastases for both MRI and whole-body scintigraphy were calculated with standard formulas [23].

Results The study included 40 patients (25 females and 15 males) with a mean age of 38 years (range, 1173 years). Patients characteristics are shown in Table 1. In our cohort, there were more females than males (62.5% vs 37.5%), more papillary carcinomas than follicular carcinomas and Hrthle cell carcinomas (95% vs 2.5% and 2.5%, respectively), and more patients younger than 45 years than patients who were older (70% vs 30%). Initially, regional

lymph node metastases were present in 57.5% of patients and were combined with lung distant metastases in 2% of patients. Lymph node metastases were not present at the time of diagnosis in 37.5% of patients. During the course of the disease, nodal metastases developed in eight patients, and distant metastases appeared in three patients. The lymph nodes were grouped according to the guidelines of the American Joint Committee on Cancer; the terminology divides the lymph node groups into a series of seven levels of prognostic importance [14]. MR images showed cervical lymph node metastasis as partly cystic and partly solid. Solid parts varied from hypointense to hyperintense on T1- and T2-weighted images. Cystic areas were hyperintense on T1- and T2weighted images and on contrast-enhanced

TABLE 2: Definitive Diagnoses of Lymph Nodes

Denitive Diagnosis Metastatic lymph nodes Well-differentiated thyroid carcinoma Poorly differentiated thyroid carcinoma Total Benign lymph nodes Lipomatosis Sinus histiocytosis Lymphadenitis reactiva Total Overall total 1 1 3 5 21 0 0 0 0 2 0 0 0 0 3 13 3 16 2 0 2 0 3 3 Surgery and Fine-Needle Aspiration Autopsy and Histologic Analysis and Cytologic Analysis Histologic Analysis

AJR:195, November 2010

1199

Mihailovic et al. STIR images for 12 (57.1%) of 21 patients with conrmed metastases (MRI was positive in 10 patients and suspicious in two patients). Extracapsular spread was detected on MRI for ve (23.8%) of 21 patients with conrmed metastatic lymph nodes. Malignant lymphadenopathy showed heterogeneous enhancement on contrast-enhanced T1-weighted images in all patients when contrast medium was administered. The 40 patients were divided into two groups according to whether they had benign or malignant pathologic abnormalities; only 21 of them underwent surgery. Conrmation of malignant disease was made for 21 patientshistologically for 19 patients (16 patients after surgery and three patients after autopsy) and cytologically after FNA for two patients. The remaining 19 patients had benign pathologic abnormalities, which were histologically conrmed for ve patients (after the surgery) and were based on clinical follow-up, including no evidence of recurrent disease, for 11 patients (Table 2). Figure 1 shows a patient with cervical lymph node metastases that were clearly detected with both MRI and 131I whole-body scintigraphy. Figure 2 shows a patient with nodal metastasis that was clearly detected by MRI, whereas 131I whole-body scintigraphy showed no uptake (i.e., an empty scan) because of a dedifferentiated tumor. Figure 3 shows an example of false-negative MRI report (with a lymph node< 10 mm) and corresponding true-positive 131I whole-body scintigraphy. As shown in Table 3, patients were divided into two groups according to the presence or absence of lymph node metastases. Group 1 included 21 patients with conrmed nodal metastases and elevated Tg levels. MRI detected metastatic lymph nodes in 16 patients (clearly positive results were found in 12 patients, whereas MRI ndings suspicious for recurrence were found in four patients), six of them with non-iodine-avid recurrence. In 76.2% of cases, there were true-positive MRI results (including four suspicious ndings). Metastases were missed in ve (23.8%) of 21 patients on MRI (false-negative results) but were detected on posttherapeutic 131I wholebody scintigraphy. All of these ve patients had nodes smaller than 10 mm on MR images. According to the anatomic location of recurrences, there were cervical metastases in 11 patients, mediastinal metastases in two patients, and cervical metastases combined with mediastinal localization in eight patients. The results of whole-body scintigraphy were true-

Downloaded from www.ajronline.org by 113.197.110.71 on 07/15/13 from IP address 113.197.110.71. Copyright ARRS. For personal use only; all rights reserved

Fig. 1 35-year-old woman with papillary thyroid carcinoma (T2aN1aM0/I), after total thyroidectomy with right neck dissection and ablative 131I therapy, with elevated Tg levels on hormonal therapy. A and B, MR STIR sequence image in coronal plane ( A ) and MR T1-weighted turbo spin-echo fat-suppressed image in axial plane (B) show bilateral cysticlike metastatic lymph nodes (arrows ) in neck with hyperintense signal characteristics. Lymph nodes are clustered, sized 10 30 35 mm, located in level 4 on right side, and characterized by inltration. C and D, Posttherapeutic 131I whole-body scintigraphy anteroposterior view (C) and posteroanterior view (D) show bilateral foci of 131I uptake in neck (arrows ), corresponding to metastatic lymph nodes detected on MRI.

positive in 15 (71.4%) of 21 patients and falsenegative in the other six patients (28.6%) whose nodes did not concentrate 131I but who had positive MRI ndings. Three of these six patients were surgically treated, and they were histologically conrmed to have poorly differentiated carcinoma.

Group 2 included 19 patients with benign nodal pathologic abnormalities, all of them considered as negative for recurrences on 131I whole-body scintigraphy (100% true-negative). Among them, ve patients with clinically palpable lymph nodes underwent surgery and had benign histologic results. Two (10.5%) of the

1200

AJR:195, November 2010

Detection of Lymph Node Recurrences in Thyroid Carcinoma

Downloaded from www.ajronline.org by 113.197.110.71 on 07/15/13 from IP address 113.197.110.71. Copyright ARRS. For personal use only; all rights reserved

Fig. 2 37-year-old woman with papillary thyroid carcinoma (T1bN1aM0/I), after total thyroidectomy followed with two cycles of 131I therapy, with elevated Tg level on

hormonal therapy. A and B, MR STIR sequence in coronal plane ( A ) and MR T1-weighted turbo spin-echo axial fat-suppressed image (B) show metastatic lymph node (arrows ), located in upper mediastinum, on left paratracheal side. Node is round in shape, is 17 14 15 mm in size, and is clearly marginated with no sign of inltration. Node is characterized as hyperintense with heterogeneous structure. C, Posttherapeutic 131I-whole-body scintigraphy anteroposterior view shows no radioiodine uptake due to nonavid metastases.

patients who underwent surgery had false-positive results on MRI. In spite of negative imaging studies and negative Tg measurements, the other three patients underwent surgery at their own request. MRI excluded metastases in the other 17 patients (89.5% true-negative). There were no false-positive results with 131I whole-body scintigraphy, whereas nonablated thyroid remnants were detected in eight of 19 patients. All patients with recurrent disease had signicantly elevated Tg levels, whereas all patients with negative Tg levels had no evidence of metastatic disease. Eight patients with mildly elevated Tg levels were found to have thyroid remnants on whole-body scintigraphy, which were subsequently treated with return of their Tg levels to normal. MRI clearly identied thyroid remnants as hyperintense in STIR sequences, with enhancement on contrast-enhanced T1-weighted images. In detection of malignant lymph

nodes, we found that MRI had sensitivity, specicity, accuracy, PPV, and NPV of 76.2%, 89.5%, 82.5%, 88.9%, and 77.3%, respectively. On the other hand, 131I whole-body scintigraphy had sensitivity, specicity, accuracy, PPV, and NPV of 71.4%, 100.o%, 85.0%, 100.0%, and 76.0%, respectively. With statistical analysis, suspicious MRI results were included in the group considered as true-positive. Discussion Because nodal metastases may appear late during the course of disease, patients with DTC require life-long follow-up. The purpose of the follow-up is timely detection of local or regional recurrences. However, the prognostic importance of regional lymph node metastases is still controversial. Some nd that their presence has no impact on recurrence or survival. On the contrary, Maz-

zaferri and Kloos [2] detected signicantly higher 30-year cancer mortality rates in patients with nodal metastases compared with patients without nodal metastases (10% vs 6%; p < 0.01). Because the presence of lymph node metastases is associated with a 50% reduction in 5-year survival, the detection of nodal recurrences in patients with DTC is of signicant importance [14]. In this study, MR images of the neck and upper mediastinum were analyzed qualitatively according to several criteria. We detected lymph node metastases (if > 10 mm in short axis) in 76% of cases and missed recurrences in 24% of cases (if size was < 10 mm). Our criteria for malignant lymph nodes were in accordance with RECIST recommendations [22]. Nodal size exceeding 10 mm has been suggested as indicator of cervical recurrence in other studies, too [10, 18]. However,

Fig. 328-year-old man with papillary thyroid carcinoma (T1aN1aM0/I), after total thyroidectomy followed with two cycles of 131I therapy. A, MRI STIR sequence in coronal plane shows 9-mm cervical lymph node (arrow ), located in right supraclavicular region. B, Posttherapeutic 131I-whole-body scintigraphy anteroposterior view shows focus of pathologic 131I uptake located on right side of neck (arrow ).

B
1201

AJR:195, November 2010

Mihailovic et al. TABLE 3: MRI and Whole-Body Scintigraphy Detecting Lymph Node Metastases in Patients With Differentiated Thyroid Carcinoma
Group of Patients Imaging Procedure, Result MRI Negative Positive Whole-body scintigraphy Negative Positive 6 15 19 0 5 16 17 2 71.4 100.0 85.0 100.0 76.0 Group 1: Metastatic Lymph Nodes Group 2: Benign Lymph Nodes Diagnostic Statistics Sensitivity (%) Specicity (%) Accuracy (%) 76.2 89.5 82.5 PPV (%) 88.9 NPV (%) 77.3

Downloaded from www.ajronline.org by 113.197.110.71 on 07/15/13 from IP address 113.197.110.71. Copyright ARRS. For personal use only; all rights reserved

NoteNPV = negative predictive value, PPV = positive predictive value.

because of the large number of false-negative results, a conventional size criterion (> 10 mm transverse diameter in the upper neck and 9 mm elsewhere) is not a sensitive indicator for the detection of nodal metastases [12, 18, 19]. More than half of the metastatic lymph nodes of papillary thyroid carcinoma are less than 3 mm in diameter [24]. Lymph nodes are considered as metastatic if they were 0.70.8 cm, or if there are grouped lymph nodes (three or more) with 0.5 cm of size for each of them [25]. Some authors have declared a size criterion to be an unsafe indicator of malignancy because there is signicant overlap between unspecic and metastatic lymph nodes [9]. In 57% of patients, we detected hyperintense necrotic recurrences on T1- or T2weighted images and fat-suppressed T1weighted images. Similarly, cystic areas showing high SI on both T1- and T2-weighted images have been described in other studies [5, 11, 13]. Contrast-enhanced fat-suppressed T1-weighted images improve recognition of nodal necrosis [9]. According to the postmortem histopathologic analysis in three patients, we explained high SI by high Tg content and hemorrhage. Hyperintense necrotic metastases on T1- and T2-weighted images due to hemorrhage and high Tg content has been described elsewhere [8, 12, 15]. In 24% of our patients, MRI detected extracapsular spread of metastatic recurrences. This criterion has already been reported as a signicant factor indicating malignant lymphadenopathy [14, 16]. In our study, clustered lymph nodes were detected in some patients with large cervical lymph nodes. In the literature, the presence of clusters is suggested to be one of the criteria for malignant lymphadenopathy [14]. We detected malignant lymph nodes in the mediastinal region in 10 patients (combined with cervical recurrences in eight of them), which indicates that MRI is helpful in the de-

tection of mediastinal recurrence in patients with DTC. This nding indicates that, as a result of excellent tissue contrast, MRI enables better evaluation of deep cervical and mediastinal regions. Although sonography has 96% sensitivity for locoregional metastases in patients with elevated Tg levels [4], MRI is superior for the evaluation of mediastinal lymph nodes where ultrasound has limited penetration [5, 6, 10, 19]. Recently, it has been reported that MRI of retropharyngeal and parapharyngeal area should be performed, especially in patients with increasing Tg levels and negative or inconclusive neck ultrasound with a history of 131I therapy and neck dissection [17]. We did not have any false-negative or false-positive Tg results in spite of the fact that Tg may be undetectable in 20% of patients receiving L-thyroxine treatment who have isolated lymph node metastases [1]. In patients who underwent total thyroidectomy and radioablation, serum Tg measurement has been the reference standard for early detection of recurrent DTC [26]. Our results show that MRI has high accuracy (83%) and very high specicity (90%) and PPV (89%) in the detection of lymph node metastases. Similar results were reported by other authors, with sensitivity that varied between studies. For the differentiation between tumor recurrences and postoperative scar, MRI has a PPV of 82%, an NPV of 86%, and diagnostic accuracy of 83% [10]. For detection of cervical nodal metastases in patients with DTC, MRI showed high sensitivity of 95%, accuracy of 83%, and PPV of 84%, whereas specicity was poor (51%) [11]. Results of another MRI study indicated less sensitivity (59%) in the detection of neck metastases because false-negative ndings were obtained in 41% of cases, whereas specicity, accuracy, and PPV were high (100%, 82%, and 100%, respectively) [12]. According to

our results, 131I whole-body scintigraphy showed high specicity, accuracy, and PPV. When the two imaging techniques are compared, this study convincingly shows that MRI has higher sensitivity (76.2% vs 71.4%) and NPV (77.3% vs 76.0%) than 131I wholebody scintigraphy, but that MRI has lower specicity (89.5% vs 100%), accuracy (82.5% vs 85%), and PPV (88.9% vs 100.0%) than 131I whole-body scintigraphy. Non-iodine-avid recurrences, which were detected on MRI, showed false-negative results on posttherapeutic 131I whole-body scintigraphy in 29% of our patients. As already reported, whole-body scintigraphy has excellent accuracy for the detection of DTC metastases, but it is limited only for metastases accumulating iodine [27]. The loss of a tumors ability to concentrate 131I is explained by tumor dedifferentiation with reduced or absent sodium-iodine symporter function [28]. Although it is expensive and therefore not easily available, PET/CT is useful in localizing the metastases that do not accumulate iodine in patients with elevated Tg levels [29]. There are several limitations to this study. This study was retrospective in nature and excluded some imaging methods that could be additionally performed, such as ultrasound initially. Moreover, the sonography results could be compared with both MRI and wholebody scintigraphy images. There is also a possible selection bias in that we enrolled patients suspected of having recurrence on the basis of clinical information. Furthermore, it is most likely that, for some of our patients, the nodal pathologic analysis would have changed if we had changed the size criteria from greater than 10 mm to greater than 7 mm or 8 mm. Consequently, such a change should reduce false-negative MRI reports. However, on the basis of results of this study, we suggest the following algorithm for

1202

AJR:195, November 2010

Detection of Lymph Node Recurrences in Thyroid Carcinoma the appropriate management and posttreatment follow-up of patients with DTC. In accordance with the American Thyroid Association recommendations [30], we suggest that both ultrasound and determination of Tg level be performed routinely. Whole-body scintigraphy should be performed only if patients with ultrasound-negative or inconclusive results show persistently elevated Tg levels. The additional benet of whole-body scintigraphy is in determination of the tumors iodine avidity; it makes this technique important in the decision-making process for possible future treatment of metastatic disease with 131I. Although it is not economically feasible to apply MRI to all patients with DTC as screening for deep extramucosal neck metastases, it should be done for diagnostic evaluation of mediastinal areas. Morphologic information, such as extracapsular spread, suggests that MRI is the desirable method to perform in patients who are planning to undergo surgical resection. If FDG PET is available, it should be included for detection of nodal recurrences in Tg-positive and 131I-negative patients. In conclusion, MRI cannot replace 131I whole-body scintigraphy as a standard procedure in the follow-up of patients with DTC. MRI is most useful in patients with non-iodine-avid recurrences, where it has a high specicity and accuracy for nodal detection and exact localization. This imaging method also has a signicant role in the detection of lymph node metastases in the mediastinal area. Despite excellent morphologic characterization of metastatic nodal recurrences, MRI cannot reliably make a differentiation between benign and malignant lymph nodes. References
1. Schlumberger M, Pacini F. Thyroid tumors, 3rd ed. Paris, France: Editions Nucleon, 2006 2. Mazzaferri EL, Kloos RT. Carcinoma of follicular epithelium: radioiodine and other treatments and outcomes. In: Braverman LE, Utiger RD, eds. Werner and Ingbars the thyroid: a fundamental and clinical text, 9th ed. Philadelphia, PA: Lippincott Williams & Wilkins, 2005:934966 3. Thyroid Carcinoma Task force. CE/AAES medical/surgical guidelines for clinical practice: management of thyroid carcinoma. American Association of Clinical Endocrinologists. American College of Endocrinology. Endocr Pract 2001; 7:202220 4. Pacini F, Molinaro E, Castagna MG, et al. Recombinant human thyrotropin-stimulated serum thyroglobulin combined with neck ultrasonography has the highest sensitivity in monitoring differentiated thyroid carcinoma. J Clin Endocrinol Metab 2003; 88:36683673 5. Miyakoshi A, Dalley RW, Anzai Y. Magnetic resonance imaging of thyroid cancer. Top Magn Reson Imaging 2007; 18:293302 6. Weber AL, Randolph G, Aksoy FG. The thyroid and parathyroid glands: CT and MR imaging with pathology and clinical ndings. Radiol Clin North Am 2000; 38:11051129 7. Gotway MB, Higgins CB. MR imaging of the thyroid and parathyroid glands. Magn Reson Imaging Clin N Am 2000; 8:163182 8. Branstetter BF, Weissman JI. Head and neck. In: Edelman RR, Hesselink JR, Zlatkin MB, Crues JV, eds. Clinical magnetic resonance imaging, 3rd ed., Vol. II. Philadelphia, PA: Elsevier, 2006:21152135 9. Risse JH. Magnetic resonance imaging. In: Biersack HJ, Grunwald F, eds. Thyroid cancer. Berlin, Heidelberg, Germany: Springer-Verlag, 2005:251282 10. Auffermann W, Clark OH, Thurnher S, Galante M, Higgins CB. Recurrent thyroid carcinoma: characteristics on MR images. Radiology 1988; 168:753757 11. Gross ND, Weissman JL, Talbot JM, Andersen PE, Wax MK, Cohen JI. MRI detection of cervical metastasis from differentiated thyroid carcinoma. Laryngoscope 2001; 111:19051909 12. Takashima S, Sone S, Takayama F, et al. Papillary thyroid carcinoma: MR diagnosis of lymph node metastases. AJNR 1998; 19:509513 13. King AD, Ahuja AT, To EWH, Tse GMK, Metreweli C. Staging papillary carcinoma of the thyroid: magnetic resonance imaging vs ultrasound of the neck. Clin Radiol 2000; 55:222226 14. Wippold FJ II. Neck. In: Lee JKT, Sagel SS, Stanley RJ, Haiken JP, eds. Computed body tomography with MRI correlation, 3rd ed. Philadelphia, PA: Lippincott-Raven, 1998:107182 15. Som PM, Brandwein M, Lidov M, Biller HF. The varied presentation of papillary thyroid carcinoma cervical nodal disease: CT and MR ndings. AJNR 1994; 15:11231128 16. Yousem DM, Som PM, Hackney DB, Schwaibold F, Hendrix RA. Central nodal necrosis and extracapsular neoplastic spread in cervical lymph nodes: MRI imaging versus CT. Radiology 1992; 182:753759 17. Kaplan SL, Mandel SJ, Muller R, Baloch ZW, Thaler ER, Loevner LA. The role of MR imaging in detecting nodal disease in thyroidectomy patients with rising thyroglobulin levels. AJNR 2009; 30:608612 18. Ohnishi T, Noguchi S, Murakami N, et al. Detection of recurrent thyroid cancer: MR versus thallium-201 scintigraphy. AJNR 1993; 14:10511057 19. Toubert ME, Gorse FC, Zagdanski AM, et al. Cervicomediastinal magnetic resonance imaging in persistent or recurrent papillary thyroid carcinoma: clinical use and limits. Thyroid 1999; 9:591597 20. DeLellis RA, Lloyd RV, Heitz PU, Eng C, eds. World Health Organization classication of tumours: pathology and geneticstumours of endocrine organs. Lyon, France: IARC Press, 2004:49133 21. Stefanovic LJ, Guduric B, Sljapic N, et al. Protocols of diagnostics, treatment and control of malignant thyroid tumors at the Institute of Oncology in Sremska Kamenica [in Serbian]. Onc Arch 1994; 2:101107 22. Schwartz LH, Bogaertsb J, Fordc R, et al. Evaluation of lymph nodes with RECIST 1.1. Eur J Cancer 2009; 45:261267 23. Patton DD. Introduction to clinical decision making. Semin Nucl Med 1978; 8:273282 24. Noguchi S, Noguchi A, Murakami N. Papillary carcinoma of the thyroid. Part I. Developing pattern of metastasis. Cancer 1970; 26:10531060 25. Van den Brekel MWM, Castelijns JA, Stel HV, et al. Occult metastatic neck disease: detection with US and US-guided ne-needle aspiration cytology. Radiology 1991; 180:457461 26. Haq MS, Harmer C. Non-surgical management of thyroid cancer. In: Mazzaferri EL, Harmer C, Mallick UK, Kendall-Taylor P, eds. Practical management of thyroid cancer: a multidisciplinary approach. London, UK: Springer-Verlag, 2006:171191 27. Freudenberg LS, Bockisch A, Jentzen W. 124I Positron emission tomographic dosimetry and positron emission tomography/computed tomography imaging in differentiation thyroid cancer. In: Biersack HJ, Grunwald F, eds. Thyroid cancer. Berlin, Heidelberg, Germany: Springer-Verlag, 2005:127 138 28. Mazzaferri EL. Management of differentiated thyroid carcinoma patients with negative whole-body radioiodine scans and elevated serum thyroglobulin levels. In: Mazzaferri EL, Harmer C, Mallick UK, Kendall-Taylor P, eds. Practical management of thyroid cancer: a multidisciplinary approach. London, UK: Springer-Verlag, 2006:237251 29. Freudenberg LS, Antoch G, Frilling A, et al. Combined metabolic and morphologic imaging in thyroid carcinoma patients with elevated serum thyroglobulin and negative cervical ultrasonography: role of 124I-PET/CT and FDG-PET. Eur J Nucl Med Mol Imaging 2008; 35:950957 30. Cooper DS, Doherty GM, Haugen BR, et al. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid 2009; 19:11671214

Downloaded from www.ajronline.org by 113.197.110.71 on 07/15/13 from IP address 113.197.110.71. Copyright ARRS. For personal use only; all rights reserved

AJR:195, November 2010

1203