At our institution, all patients who undergo total thyroidectomy routinely undergo diagnostic ¹³¹I scanning 5–6 weeks after the operation to complete staging before the first dose of radioiodine therapy is administered. Based on staging information incorporating clinical, histopathologic, and imaging data and whole-body dosimetric calculations, the ¹³¹I therapy dose is adjusted according to the patient's risk classi fication. Our protocol typically involves administration of a low-dose radioiodine (1.1 GBq [30 mCi]) for thyroid remnant ablation, medium dose (3.7–5.5 GBq [100–150 mCi]) for regional nodal disease, and high dose (7.4–11.1 GBq [200–300 mCi]) for distant metastasis.

Whole-body and static neck and chest images with additional neck pinhole images as required are obtained 24 and 48 hours after administration of a diagnostic 37-MBq (1 mCi) ¹³¹I dose under thyroid hormone withdrawal (hypothyroid ¹³¹I scan protocol) or a 150-MBq (4 mCi) dose with recombinant human thyroid-stimulating hormone (thyrotropin alfa, Thyrogen, Genzyme) stimulation. In a selected patient group, SPECT/CT images are acquired 48 hours after ¹³¹I administration for further evaluation of activity foci seen on planar images. We have found that SPECT/CT provides useful information for the following indications: characterization of equivocal central neck activity, identification of regional lymph node metastasis, anatomic localization of distant metastatic foci, evaluation of suspected physiologic mimics of disease, and assessment of discrepancies between ¹³¹I planar imaging and histopathologic or biochemical data. Forty-eight hours after administration of a therapeutic ¹³¹I dose, a posttherapy scan is obtained to assess for additional foci of activity compared with the diagnostic scan.

Fifty-six studies of 53 patients (41 women, 12 men; mean age, 47.3 years; range, 17–83 years) with DTC referred to the clinic between February 2005 and March 2007 who underwent ¹³¹I planar imaging and SPECT/CT were included in this retrospective study. This group included 48 diagnostic ¹³¹I studies before first radioiodine therapy, four diagnostic ¹³¹I studies with recombinant human thyroid-stimulating hormone stimulation, and four posttherapy ¹³¹I studies. The histologic types were papillary thyroid carcinoma (n = 48), follicular carcinoma (n = 3), and Hürthle cell carcinoma (n = 2). One patient was excluded from the study because SPECT/CT data were not retrievable from the archive. This patient had regional lymph node metastasis confirmed at SPECT/CT. All patients with diagnostic ¹³¹I studies also underwent posttherapy planar imaging. However, posttherapy SPECT/CT was not performed in most cases because additional foci were rarely identified. Two patients underwent both diagnostic and posttherapy SPECT/CT, and two patients underwent only posttherapy SPECT/CT for evaluation of differences between radioiodine distribution on the diagnostic and posttherapy planar images.

Iodine-131 whole-body planar and static neck and chest images were acquired in the anterior and posterior projections with a dual-head gamma camera (ECAM, Siemens Medical Solutions) with parallel-hole high-energy collimators and a 20% energy window set at 364 keV ± 15%. The table speed for the whole-body images was 9 cm/s. The static image acquisition time was 10 minutes to 500 K counts (matrix size, 256 × 256). The field of view for static planar images was 59.1 × 44.5 cm.

SPECT/CT images were obtained with a dualhead hybrid gamma camera with inline CT capability (Symbia T6, Siemens Medical Solutions). SPECT images were accquired with 64 steps (20 s/stop), noncircular orbit over 360°, and a 128 × 128 matrix. Standard reconstruction technique entailed 3D ordered subset expectation maximization iterative reconstruction with eight iterations and four subsets and a CT-based attenuation-corrected algorithm applied to the SPECT images. Filtered back-projection was used in a small number of reconstructions. Low-dose CT parameters consisted of 130 kV and 100 mAs. Reconstruction was performed at 5-mm slice thickness on a 512 × 512 matrix and a field of view of 53.3 × 38.7 cm. Patients were imaged with their arms down; immobilizers were not used.

The incremental value of ¹³¹I SPECT/CT over planar imaging was determined with focus-based analysis of all discrete ¹³¹I activity identified on planar and SPECT/CT images. The ¹³¹I scans were viewed by two nuclear medicine physicians designated reader 1 and reader 2. Reader 1 was blinded to clinical and biochemical data to avoid bias in lesion classification with the intent that interpretation be based solely on imaging properties. Reader 2, the nuclear endocrinologist, was unblinded and highly familiar with all patients and provided an independent reading with full access to clinical, biochemical, and histopathologic results.

Reader 1 interpreted planar ¹³¹I images and classified foci as central neck (thyroid bed or neck) or distant and further evaluated lesions according to intensity, location, nature, and reader confidence. Pinhole images were available in 10 cases. The superior resolution of pinhole images allowed resolution of intense ¹³¹I activity into several discrete foci or confirmation of the presence of very faint planar ¹³¹I uptake. All foci were reinterpreted after review of ¹³¹I SPECT/CT images with software (MedImage, MedView) that allowed display of side-by-side coregistered images and fused images. Incremental diagnostic information from SPECT/CT and limitations of SPECT/CT technique, including misregistration and poor image quality, were recorded.

Interpretations by reader 1 were compared with a reference standard prepared by reader 2. This standard included comprehensive interpretation of histopathologic reports, biochemical data, clinical details, and ¹³¹I planar and SPECT/CT findings. All patients had histologic records on the surgical specimen that included information regarding tumor size, surgical margins, and extrathyroidal tumor spread. Thirty patients underwent central neck compartment dissection and had histologic records on the presence or absence of nodal metastasis. This reference standard was verified with clinical and imaging follow-up findings. Imaging follow-up included neck sonography or diagnostic CT 2 months after radioiodine therapy and repeated ¹³¹I scintigraphy 6–12 months after therapy. Biochemical markers were evaluated 2 and 6 months after therapy.

Evaluation of reader confidence was based on a subjective 5-point scale (1, no confidence; 2, equivocal; 3, possible; 4, probable; 5, certain). Reader 1 reported increased reader confidence with SPECT/CT compared with planar imaging in the diagnosis of 104 of the 147 foci; confidence did not change regarding 32 foci and decreased regarding 11 foci on the basis of suboptimal SPECT/CT images. Of the 17 distant foci, there was no change in reader confidence regarding eight foci, but reader confidence increased regarding nine foci owing to superior anatomic localization and additional findings on CT.

Two months after radioiodine therapy, 47 patients underwent sonography of the neck or diagnostic CT including the neck region. Iodine-131 scintigraphy was performed on 42 patients 6–12 months after therapy. The group of 11 patients without follow-up ¹³¹I scans included six patients who underwent clinical follow-up with no evidence of recurrent disease to prompt further imaging, including the four patients imaged with the recombinant human thyroid-stimulating hormone protocol. A further four patients underwent 2-month follow-up imaging after which they relocated out of area. One patient with regional disease detected on SPECT/CT did not undergo follow-up. Clinical follow-up is summarized in Table 3.

Twenty-nine of 53 patients had the SPECT/CT finding of thyroid remnant. Twenty-eight of these patients had no evidence of residual functioning thyroid tissue after radioiodine therapy, either normal findings at sonography of the neck and ¹³¹I scanning and normal thyroglobulin levels (n = 19) or normal findings at clinical examination and normal thyroglobulin levels (n = 9). This follow-up finding strongly indicates that SPECT/CT information is correct in assignment of ¹³¹I foci as thyroid bed or thyroglossal duct remnant, especially in view of the low-risk histologic surgical specimens (small tumors with negative surgical resection margins) in most of these patients. One patient with SPECT/CT findings of thyroid remnant had evidence of recurrent neck disease. This patient had received medium-dose ¹³¹I therapy because the presence of positive lymph nodes was documented in the postsurgical histopathology report. The SPECT/CT findings may have been normal because of limited spatial resolution or the presence of non-iodine-avid disease.

Fifteen of 53 patients had the SPECT/CT diagnosis of regional disease. Clinical follow-up of six of the 15 patients revealed stable or progressive regional disease after radioiodine treatment. The other eight patients who underwent follow-up had normal imaging findings and thyroglobulin levels and were considered to have controlled nodal metastasis. Despite lack of biopsy confirmation, the presence of ¹³¹I-avid lymph nodes on SPECT/CT was convincing evidence of the diagnosis. Distant metastatic lesions were found in nine of 53 patients, including five patients with pulmonary metastasis, two patients with bone metastasis, and two patients with both lung and bone metastasis. Iodine-131 scanning and CT follow-up showed stable pulmonary disease in three patients, an interval decrease in size of lung nodules in one patient, and complete resolution of lung disease in one patient after ¹³¹I therapy. Two patients with bone metastasis had complete resolution of disease according to ¹³¹I scanning and CT findings, and the two patients with lung and bone metastasis had progressive disease according to imaging and biochemical data.

The availability of ¹³¹I SPECT/CT has had great impact on our nuclear therapy clinical practice. Planar ¹³¹I imaging has traditionally suffered from low resolution, and a paucity of anatomic information along with a long list of physiologic variants makes image interpretation challenging. The use of multiple maneuvers to aid in differentiation of physiologic from pathologic foci of activity, including swallowing water, separate-day imaging, oblique and lateral imaging, washing the patient's skin, removing and scanning clothing, correlating with other imaging techniques, and remaining vigilant for pitfalls such as dentures, handkerchiefs, and other sources of contamination, is essential for accurate diagnosis. In our experience, ¹³¹I SPECT/CT has facilitated rapid, accurate, and confident assessment of radioiodine activity outside the expected biodistribution. On SPECT/CT images, central neck activity is characterized as thyroid remnant or locoregional disease, and the number of equivocal foci on planar assessment is reduced. In the detection of distant metastatic disease, the superior lesion localization and additional CT-derived anatomic information obtained with SPECT/CT increase reader confidence, assisting clinical management decisions.

The utility of ¹³¹I SPECT/CT for investigation of DTC has been evaluated in studies in which the patient groups underwent predominately or only posttherapy iodine studies. In early studies, SPECT and CT scans were compared with software-fused SPECT/CT [11] and SPECT was compared with SPECT/CT [12]. Although it is not routine clinical practice to use SPECT with ¹³¹I studies, both groups of investigators found SPECT/CT had incremental diagnostic value due to improved lesion localization and characterization. Tharp and colleagues [13] reported their experience with ¹³¹I SPECT/CT in a mixed group comprising 17 patients who received a diagnostic dose (eight before ablation) and 54 patients who underwent therapeutic-dose scanning. Those investigators evaluated the incremental value in SPECT/CT over planar imaging, which is the usual clinical imaging method, and found additional value for 57% (41 of 71) of patients, with a substantial impact on clinical management. SPECT/CT improved localization and characterization of foci inside and outside the neck, including equivocal foci on planar ¹³¹I images. The SPECT/CT findings led to a change in treatment approach (dose of radioiodine therapy or direction of surgical management). Our study confirmed these findings of incremental value with diagnostic ¹³¹I SPECT/CT. The high interobserver agreement between the blinded reader and the unblinded reader who provided the reference standard suggests that ¹³¹I SPECT/CT is accurate for characterization of activity foci identified on planar images and when applied to a diagnostic ¹³¹I study allows completion of TNM staging and risk stratification based on the patient's disease burden.

Despite the benefits of ¹³¹I SPECT/CT, we encountered technical difficulties and challenges early in our experience when applying this technology to diagnostic radioiodine scintigraphy. The low intrinsic activity of diagnostic ¹³¹I scans may be poorly sampled with SPECT and unresolved after tomographic reconstruction owing to inadequate count statistics. This problem occurred both regionally and distantly in a small proportion (< 5%) of our patients with 3D ordered subset expectation maximization iterative reconstruction technique and might be overcome with the use of filtered back-projection reconstruction. We are exploring different parameters, such as matrix size, number of steps, and number of seconds per step position, with the goal of optimizing SPECT acquisition for ¹³¹I imaging with general trade-offs among spatial resolution, adequate count activity per pixel, and patient movement with longer acquisition times.

Considerable misregistration of functional and anatomic data was encountered in only one case. In general, integrated SPECT/CT cameras image patients in the same bed position during the same session to achieve image fusion without the use of external fiducial markers. In practice, however, patient movement on the bed continues to occur and may be reduced with immobilizers. The difficulty with ¹³¹I SPECT/CT, unlike ¹⁸F-FDG PET/CT, is the lack of background tissue activity on functional images, which facilitates recognition of misregistration on FDG imaging. We use salivary gland activity as a check for registration by ensuring alignment of activity to the correct anatomic location of the glands. Misregistration may also be related to the fusion software used and may arise from the use of different header information on tomographic data from different vendors.

We have found planar images, including pinhole collimation, very useful for interpretation of SPECT/CT images. Problems encountered with SPECT of low-level ¹³¹I activity included limited spatial resolution that blurred two adjacent foci seen distinctly on planar images, reconstruction of septal penetration activity into discrete foci with iterative reconstruction, streak artifact related to filtered back-projection reconstruction of intense ¹³¹I activity, and attenuation of central neck activity by the shoulders of patients with non-attenuation-corrected filtered back-projection images. We therefore support the view that SPECT/CT images should be interpreted in conjunction with planar images and not in isolation.

Our study of ¹³¹I SPECT/CT was conducted with a selected group of patients. Therefore, incremental diagnostic value might be expected to occur in a higher proportion of cases than if SPECT/CT had been used for every patient. However, that ¹³¹I SPECT/CT allows more accurate risk stratification of patients with DTC may favor its routine use. SPECT/CT test performance parameters such as sensitivity, specificity, and accuracy could not be calculated directly owing to the lack of a histologic reference standard for assessing ¹³¹I-avid lesions. After ¹³¹I therapy, thyroid remnant, local residual disease, and regional lymph node metastasis in the central part of the neck may resolve; therefore, the true nature of lesions can be inferred with follow-up but not confirmed. The reference standard included SPECT/CT as a major component in determining the nature of physiologic and pathologic activity. This limitation is inherent to most other studies in which an attempt is made to determine the incremental value of SPECT/CT over planar imaging.

Iodine-131 SPECT/CT provides incremental diagnostic information compared with that obtained at planar imaging. This information can be used for accurate characterization of central neck activity, superior localization of distant metastatic lesions, evaluation and rapid confirmation of suspected physiologic mimics of disease, and increasing reader confidence.