Measuring the Extent of Total Thyroidectomy for Differentiated Thyroid Carcinoma Using Radioactive Iodine Imaging: Relationship With Serum Thyroglobulin and Clinical Outcomes

Importance  Despite performing total thyroidectomy (TT), postoperative radioactive iodine (RAI) imaging often demonstrates the presence of residual thyroid tissue within the operative bed.

Objective  To measure the extent of TT using postoperative RAI imaging and assessing serum thyroglobulin (Tg) level for patients with differentiated thyroid carcinoma (DTC).

Design, Setting, and Participants  We evaluated 245 patients undergoing TT for clinically staged cT1-3N0M0 DTC, who underwent diagnostic postoperative RAI imaging.

Interventions  Total thyroidectomy.

Main Outcomes and Measures  On the basis of quantitative measurements, RAI uptake (RAIU) in the thyroid bed of 0.2% of administered activity was selected as the cutpoint to determine the presence or absence of thyroid remnant.

Results  By postoperative RAI imaging, TT in 106 patients (43%) resulted in RAIU of less than 0.2%. In the remaining 139 patients (57%), there was measurable iodine-avid thyroid tissue and/or tumor in the thyroid bed (n = 117 [84%]), the neck (n = 4 [3%]), or both (n = 18 [13%]). For the entire study population, mean 24-hour RAIU was 0.62%. Stimulated serum Tg levels were obtained in 232 of 245 patients (95%). Measurable stimulated Tg level (≥1 ng/mL) (to convert to micrograms per liter, multiply by 1) was found in 26 of 102 patients (25%) without thyroid remnant and in 87of 133 patients (65%) with thyroid remnant (P < .001).

Conclusions and Relevance  A goal of postthyroidectomy RAIU of less than 0.2% helps maximize the likelihood of an unmeasurable postoperative Tg level, potentially simplifying follow-up evaluation and reducing the use of postoperative RAI in order to facilitate surveillance.

Total thyroidectomy (TT) is the treatment of choice for most differentiated thyroid carcinomas (DTCs) larger than 1.5 to 2.0 cm in diameter.¹ Total thyroidectomy eliminates multifocal disease, decreases the incidence of primary site recurrence, and facilitates the uptake of therapeutic postoperative radioactive iodine (RAI) by known or putative metastasis. However, the extent of surgery for DTC has been, and continues to be, a controversial subject. Some clinicians believe it is impossible to perform a true TT because of RAI uptake (RAIU) often seen after surgery.²

In 1983, Fratkin et al³ compared preoperative and postoperative RAI scans (using iodohippurate sodium I 131 [¹³¹I]) to study the cervical distribution of residual thyroid tissue in 24 patients who had undergone putative TT. Every patient in this study had evidence of RAIU, whether in the thyroid bed, the neck, or both. However, surgical technique was not standardized, with 18 patients reported as having undergone a TT and 6 patients receiving a “near-total” thyroidectomy. Thus, it was difficult to determine how effectively RAI imaging estimated the extent of thyroidectomy, especially given the heterogeneity of surgical treatment and high prevalence of RAIU remnant. In 2007, Salvatori et al⁴ revisited this question, reporting on a much larger cohort of patients treated systematically with TT and then imaged. In this study, the presence of a thyroid remnant was determined based on an RAIU of 1% or greater, and most patients (93%) had a thyroid remnant following TT for DTC.

The American College of Radiology and the Society of Nuclear Medicine (ACR-SNM) has since standardized the calculation of RAIU, using the following formula, which is based on scintillation measurements from the neck and extremity, normalized against phantom and background counts⁵:

Therefore, a range of cutpoints have been recommended to determine the presence or absence of thyroid remnant, which can be used to estimate the completeness of TT. Furthermore, in addition to RAI imaging, serum thyroglobulin (Tg) level can be used in the absence of demonstrable disease as a surrogate biochemical measure to determine the extent of TT.⁶ Thus, there are now at least 2 complementary measures that might be used to infer more accurately the extent of surgery.

With an emerging awareness of the long-term adverse effects⁷ of RAI therapy in patients with low-risk DTC,^8,9 we have examined our own experience to estimate the incidence of “total thyroidectomy” using RAI imaging and correlate these findings with serum Tg and clinical outcomes.

We evaluated patients who underwent TT at the University of Texas MD Anderson Cancer Center (F.C.H., E.G., E.M.S., J.E.L., R.S.W., and G.L.C.) from January 1, 2001, to February 20, 2012, for clinically staged cT1-3N0M0 DTC and who received diagnostic postoperative RAI imaging. Staging was based on preoperative clinical and radiographic staging. All patients underwent radiographic evaluation of the thyroid and neck by ultrasonography, except for 5 who had previously undergone computed tomography (CT) that had identified the primary thyroid cancer. Patients with poorly differentiated tumors and those with tumors classified as T4 were excluded. A retrospective review was performed after approval by the institutional review board of the University of Texas MD Anderson Cancer Center, which granted a waiver of patient informed consent. Follow-up time was the time from first appointment at our center until the date of last contact or death.

A total of 245 patients had quantitated uptake on RAI imaging and were included in this study. Imaging and uptake was performed with either ¹²³I or ¹³¹I, administered orally as sodium iodide. Radioactive iodine images were visually reviewed by 2 experienced nuclear medicine physicians (S.R. and E.M.R.), and uptake values were confirmed in the electronic medical record. As discussed in the ACR-SNM guidelines,⁵ low measurement values for RAIU were taken to represent a negative scan. However, there is currently no consensus regarding what constitutes a negative scan because of limitations of the current scanning technology and the difficulty of managing background signal. Therefore, for this study an uptake value of less than 0.2% of total administered activity was defined as the cutpoint to determine the presence or absence of measurable thyroid remnant.

Patient data were abstracted from the electronic medical record and then entered twice into a Microsoft Access database (Microsoft Inc). Edited data were independently verified for consistency and then analyzed in collaboration with a statistician (H.Y.L.) from the Department of Biostatistics, University of Texas MD Anderson Cancer Center. Differences in proportions of patients were tested by means of the Pearson χ² test or Fisher exact test. P<.05 was considered statistically significant. A paired t test (with adjustment for unequal variances where necessary) was used to compare continuous variables. Otherwise, all tests were 2-sided. All analyses were conducted in SAS statistical software (version 9.1; SAS Institute Inc). For the entire study population, median follow-up time was 43.5 months.

Demographics, staging, and surgical treatment are outlined in Table 1. Most patients (59%) underwent TT performed alone. A third of the patients underwent central compartment dissection (unilateral or bilateral). Sixty-five percent of patients had pathologically staged pT1 (45%) or pT2 (20%) stage disease. Surgical complications were rare. Transient hypocalcemia developed in 45 patients (18%), although 2 (1%) developed permanent hypoparathyroidism requiring long-term medical therapy. Four patients (2%) developed temporary vocal cord paralysis (duration <6 months), and 2 patients (1%) developed permanent unilateral vocal cord paralysis. In 2 patients, preoperative vocal cord paralysis, presumed to be independent of the thyroid neoplasm, was identified via systematic preoperative laryngoscopy. Two patients (1%) developed postoperative hematoma requiring evacuation. A single patient developed seroma, which resolved with conservative measures.

All patients underwent postoperative whole-body RAI imaging: 143 (58%) with ¹²³I and 102 (42%) with ¹³¹I. In 106 patients (43%), the measured RAIU in the thyroid bed was less than 0.2%. At the time of the diagnostic RAI study in this group, the mean thyrotropin level was 89.8 μU/L (median, 83.5 μU/L). The remaining 139 patients (57%) demonstrated visible and measurable iodine-avid thyroid tissue and/or tumor. This activity was seen in the thyroid bed–paratracheal echelon in 117 patients (84%), in the lateral neck in 4 patients (3%), or in both sites in 18 patients (13%). For those patients with visible RAIU, the mean 24-hour iodine uptake was 0.99% of administered activity (median, 0.6%; range: 0.2%-7.0%). Most patients (207 of 245 [84.5%]) had RAIU of less than 1%. Thirty-nine patients (16%) had RAIU between 1% and 5%, and only 2 patients (1%) had RAIU above 5%. For the entire study population, the mean 24-hour RAIU was 0.59%.

A total of 198 patients (81%) had postoperative RAI ablation, administered at a mean of 9 weeks after surgery, at a mean dose of 93 mCi (3.4 GBq). Serum Tg levels were obtained in 233 of 245 patients (95%). For these patients, measurable Tg level (≥1 ng/mL) (to convert to micrograms per liter, multiply by 1) was found in 26 of 102 patients (25%) without measurable thyroid remnant and in 87 of 133 patients (65%) with measurable thyroid remnant (P < .001). There was no significant relationship between the iodine isotope (¹³¹I vs ¹²³I) used for the diagnostic scan and the presence of RAIU remnant, recurrence, or serum Tg level.

Rates of local and/or regional failure were low and were independent of RAIU, whether seen in the thyroid bed, central compartment, or lateral neck (Table 2). No patient had local recurrence in the thyroid bed. However, several patients had regionally persistent or recurrent disease identified via ultrasonography and rising Tg level: 2 patients had central compartment disease; 1 patient had lateral neck recurrence 2 years following surgery and postoperative RAI imaging; 2 patients had detectable thyroid remnant; and 1 had no postoperative thyroid remnant. A single patient developed pulmonary metastases 3 years 8 months after TT, with rapid progression to bone. Her tumor was a 5-cm encapsulated pT3cN0 papillary thyroid cancer, with a baseline chest roentgenogram negative for metastasis. At the time of thyroidectomy, 2 lymph nodes were sampled and both were without evidence of metastasis. This patient had no detectable thyroid remnant, nor pulmonary or osseous lesions by RAI imaging, but did have measurable Tg level both prior to ablation (11 ng/mL) and at 6 months (10 ng/mL) and 12 months (16 ng/mL). At age 65 years, she died from metastatic papillary thyroid cancer, 5 years 10 months after TT, despite additional RAI imaging and protocol-based systemic therapy.

Total thyroidectomy is often recommended as the treatment of choice for DTC. The use of the term total in describing the extent of thyroidectomy implies complete macroscopic removal of all grossly apparent thyroid tissue; however, postoperative RAI imaging has called into question the extent and completeness of thyroidectomy in many patients who undergo so-called total thyroidectomy. We show herein that approximately 4 in 10 patients undergoing TT at our center have RAI evidence of complete removal of functioning thyroid tissue. For the remaining patients, there is evidence by RAIU suggesting the presence of residual thyroid tissue, nodal disease, or both, although the majority had RAIU of less than 1%. Absence of RAIU correlated with undetectable Tg level, suggesting that very low RAIU correlated with true TT.

In fact, Randolph and Daniels¹⁰ first suggested that “total thyroidectomy” might be a misnomer.¹⁰ Fratkin et al³ documented the high rate of RAIU-defined remnants following thyroidectomy, finding multiple foci of functioning thyroid tissue suggesting incomplete resection despite reported TT. This study suggested that surgeons may frequently fail to recognize the extent of the superior poles or pyramidal lobe, while Salvatori et al⁴ emphasized that the area of the posterolateral suspensory ligament of Berry may pose difficulties with its tight adherence to the trachea and proximity to the recurrent laryngeal nerve. Nonetheless, the surgeon must balance the desire to remove all gross residual disease and surrounding normal tissue, with the risk of nerve injury. The surgeon should be just as mindful of the external branch of the superior laryngeal nerve during the superior pole dissection, as during the mobilization of the Berry ligament and careful protection of the recurrent laryngeal nerve.

These intraoperative challenges may lead to either unintentional or deliberate retention of functioning thyroid tissue during surgery, resulting in the finding of foci of iodine-active thyroid tissue on postoperative imaging. These studies suggest that closer attention paid by the surgeon to the aforementioned locations may result in successful resection of the entire thyroid gland. Clearly, the skill, experience, and attention of the surgeon to the patient’s specific anatomic variations and the extent and invasiveness of the cancer present is what ultimately determine the potential of performing a true TT. Studies that incorporate the review of radionuclide assessment of the thyroid remnant and serum Tg level may provide important new information to perfect the surgeon’s art, although this new perspective of a complete thyroidectomy with the risks of complications.

But do these surgical fine points matter? As in our series, most patients with DTC have excellent long-term outcomes, in terms of both disease control and postoperative laryngeal and parathyroid function. To demonstrate improvements in functional outcomes, reduced rates of locoregional recurrence, and/or improved survival, a larger study over a longer time interval would be needed to discern differences between various surgical approaches, ie, “total thyroidectomy” vs “near-total thyroidectomy.” We acknowledge this point but emphasize that an important potential benefit to more meticulous surgical technique that maximizes the chance for TT may potentially reduce the need for postsurgical RAI ablation.

Radioactive iodine therapy has been used for decades in the treatment of benign and malignant thyroid diseases, and in many centers it is used routinely in a dose range of 100 to 150 mCi (3.70-5.55 GBq) for ablation of residual iodine-avid uptake within the thyroid bed identified after thyroid surgery. Ablative treatment with RAI may eliminate normal thyroid remnant or residual malignant disease, whether in the thyroid bed, paratracheal or lateral lymph nodes, or distant sites. Eliminating any residual functioning thyroid tissue or disease facilitates the use of serum Tg as more specific indicator of disease recurrence or persistence, thereby simplifying surveillance.

Radionuclide scanning does not correlate completely with a “biochemical” cure,¹¹ even though preablation stimulated Tg levels are lower and less frequently detectable in those with no RAI evidence of remnant thyroid. In our study, we were able to further demonstrate that patients with negative postsurgical RAI scans (as defined by regional uptake in the thyroid bed of <0.2%) were far more likely to have an undetectable stimulated Tg. While there is no evidence that those with RAI evidence of TT have a better prognosis, the sample size of this cohort is too limited and the follow-up not long enough to allow a robust comparison of disease-specific outcomes.

Some have suggested that postoperative RAI scanning can be eliminated altogether in favor of stimulated Tg measurement. Ibrahimpasic and colleagues¹² have advocated that routine postsurgical RAI ablation may be unnecessary in patients with low-risk and intermediate-risk thyroid cancers in which the stimulated Tg level was negative. Not only was the future use of Tg measurement not affected by the lack of RAI ablation in these patients, but there was also no difference in disease-specific survival or recurrence-free survival between the 2 groups.

However, we believe there is still valuable information to be gained from RAI imaging, at least in selected patients. Although the risk of distant metastases is low in patients with small-volume and/or node-negative disease, that risk is not zero. We find that through the use of single-photon emission computed tomography/CT in questionable cases, false positive scans can be minimized and true sites of distant disease can be well evaluated. In addition, the RAI scan allows us to determine in a short time frame whether the patient should receive an ablative dose of ¹³¹I, ie, visible residual uptake in the thyroid bed or measured uptake greater than 0.2%. The information from this scan (performed with reference standards to quantify the regional uptake in the neck) can then be applied to predict radiation protection measures after potential therapy. For example, a patient with 2.0% retention of RAI in the neck will need to avoid close contact with members of the public for a much longer period than a patient with 0.5% retention.

One limitation of our study is the variation in the use of iodine isotopes for diagnostic RAI imaging. Most patients in this study (58%) received ¹²³I for the first postoperative thyroid scan; however, earlier in our practice, ¹³¹I had been used more frequently. Given the controversy regarding the potential for “stunning” with the use of ¹³¹I for diagnostic scans,¹³ our practice gradually transitioned toward a more routine use of ¹²³I for these scans. Nonetheless, the use of ¹³¹I may have introduced a small but potential confounder when interpreting the effect of the final posttreatment result by RAI; however, we found no effect between the use of iodine isotope (¹³¹I vs ¹²³I) and the presence of RAIU remnant or serum Tg.

Nonetheless, information from the RAI scan continues to be used by our surgeons, endocrinologists, and nuclear medicine physicians to determine the administered activity when RAI ablation is indicated. Studies have suggested that lower doses of RAI (30-50 mCi, 1.1-1.85 GBq) may be as effective for remnant ablation as higher doses, with no difference in survival when compared with higher doses.^7,14 We incorporated this information along with the visual and measured RAIU in the neck to help decide whether ablation is necessary, and if so at what dose. Schneider et al¹⁵ were able to demonstrate an additional dimension of the RAI scan, finding that higher activity in the thyroid bed was associated with a higher risk of disease recurrence. Because of the low rate of recurrence in our patient population, we were not able to confirm such an association.

The role of RAI imaging and treatment in the postsurgical setting will continue to evolve. From the 1983 report by Fratkin et al³ to the present day, technical refinements in RAI imaging have permitted the multidisciplinary thyroid cancer team to better study these issues and how a more complete surgery might have an impact on adjuvant therapy.

Further improvements, such as new scanners with higher resolution, especially in combination with CT imaging, will refine postoperative assessment and better guide decision making. Sodium pertechnetate Tc 99m has a sensitivity of 81% by patient and 61% by site for detection of thyroid remnant following thyroidectomy.¹⁶ Instead of scanning 48 to 72 hours after a dose of ¹³¹I, sodium pertechnetate Tc 99m scanning was performed 10 minutes after radiotracer injection. Finally, the use of positron emission tomography (PET)/CT (fludeoxyglucose F 18 PET/CT) imaging has been proposed by Dietlein et al.¹⁷ The use of novel PET radiotracers such as ¹²⁴I may provide even more precise localization of metabolically active thyroid disease.¹⁸

Finally, while it might be tempting to recommend deferring standard ¹³¹I therapeutic treatment for TT patients with RAIU of less than 0.2% and unmeasurable Tg level, such a change should be studied in the setting of a prospective multicenter registry trial with rigorously defined parameters and carefully defined end points. Such a study should incorporate novel imaging techniques as well as the latest biochemical means to assess serum Tg level.¹⁹

Following TT, mean RAIU was less than 1% of total administered activity, and 43% of patients had no detectable thyroid remnant by RAI imaging. A detectable stimulated Tg level in the thyroid remnant–positive group was higher than that of the thyroid remnant–negative group (65% vs 25%; P < .001). A goal of postthyroidectomy RAIU of less than 0.2% helps maximize the likelihood of an unmeasurable postoperative Tg level, potentially simplifying follow-up evaluation and reducing the use of postoperative RAI imaging to facilitate surveillance in low-risk patients.

Corresponding Author: F. Christopher Holsinger, MD, Division of Head and Neck Surgery, Department of Otolaryngology, Stanford University School of Medicine, 875 Blake Wilbur Dr, CC-2227, Stanford, CA 94304 (holsinger@stanford.edu).

Submitted for Publication: May 17, 2013; final revision received January 17, 2014; accepted February 10, 2014.

Published Online: April 3, 2014. doi:10.1001/jamaoto.2014.264.

Author Contributions: Holsinger and Lin had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Holsinger, Cabanillas, Lin, Busaidy, Clayman, Rohren.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Holsinger, Ramaswamy, Lang, Lin, Clayman, Rohren.

Critical revision of the manuscript for important intellectual content: Holsinger, Cabanillas, Busaidy, Grubbs, Rahim, Sturgis, Lee, Weber, Clayman, Rohren.

Statistical analysis: Holsinger, Lang, Lin, Clayman.

Obtained funding: Holsinger.

Administrative, technical, or material support: Holsinger, Ramaswamy, Rahim.

Study supervision: Holsinger, Busaidy, Rohren.

Conflict of Interest Disclosures: None reported.

Previous Presentation: This study was presented at the 2013 Annual Meeting of the American Society of Head & Neck Surgery; April 10, 2013; Orlando, Florida.