| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
and
Davor Eterovi
Department of Nuclear Medicine (V.M., D.E.), University Hospital Split, 21 000 Split, Croatia; and Department of Medical Physics and Biophysics (D.E.), Split University School of Medicine, 21 000 Split, Croatia
Address all correspondence and requests for reprints to: Davor Eterovi, Department of Medical Physics and Biophysics, Split University School of Medicine; 21 000 Split, Croatia. E-mail: eterovic{at}bsb.mefst.hr.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objectives: We hypothesized that hypoechogenic glands, with large, densely packed cells, are more radiosensitive than normoechogenic glands, in which much radiation is wasted on more abundant colloid. We evaluated this hypothesis in a cohort of patients with Graves’ disease.
Design: This was a prospective trial of patients recruited during 4 yr and followed up 1 yr after radioiodine therapy.
Setting: This trial was held in a university hospital-outpatient clinic.
Patients: A total of 177 consecutive patients with first presentation of Graves’ disease (28 males), 23–76 yr old, who relapsed after antithyroid therapy were included in the study.
Intervention: The patients were assigned to an ablative target-absorbed dose of 200 Gy (n = 78) or randomly to 100 or 120 Gy of nonablative dose (n = 99).
Main Outcome Measures: The measures were incidences of hyperthyroidism, euthyroidism, and hypothyroidism at 12-month follow-up.
Results: At follow-up there were 25 hyperthyroid, 44 euthyroid, and 108 hypothyroid patients. Compared with 96 patients with a hypoechogenic gland, in 81 patients with a normoechogenic gland, there were more hyperthyroid (22 vs. 7%) and euthyroid (41 vs. 11%), but less hypothyroid outcomes (37 vs. 81%; P < 0.0001). The other independent predictor of increased radioresistance was the large gland volume.
Conclusion: In patients with Graves’ disease, normoechogenic and large glands are associated with increased radioresistance.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Given the amount of radioactivity applied, the target-absorbed dose depends on gland mass and radioiodine intrathyroid kinetics (uptake and biological half-time). According to thyroid echography and measuring tracer radioiodine kinetics, both of these parameters can be assessed in a single individual, to allow calculating the activity needed to produce the intended target-absorbed dose. However, the same absorbed dose (radiation energy deposited in unit mass of gland tissue) may induce widely different effects in different individuals (3). Thus, at present it seems impossible to titrate the individual doses to obtain the euthyroid state in a great majority of cases. In an individual patient, the risk of recurrent disease, requiring repetition of RIT, should be weighted against the risk of permanent hypothyroidism, requiring lifetime substitution.
In light of these problems and extra efforts associated with the assessment of gland size and radioiodine kinetics, some clinicians use a simpler fixed-dose regime, or come halfway, in accounting only for the gland size (4, 5, 6). Obviously the target tissue absorbed dose approach should not be abandoned or underestimated before the possible predictors of gland radiosensitivity are evaluated. Several reports are consistent in demonstrating that large glands are more resistant to radiation than small glands (7, 8, 9, 10, 11, 12, 13). There are also occasional observations that age (14, 15, 16), gender (14, 15), severity of hyperthyroidism (15), the baseline serum thyroglobulin level (11), and thyroid stimulating antibody activity (8), as well as antithyroid drug therapy before or after RIT (17, 18, 19, 20), predict its outcome. However, some of these studies report merely statistically, but not clinically relevant predictors (8, 11); lack the power (8, 16); do not account for the confounding due to varying absorbed dose (14, 15, 16); or use uncritically the multivariable methods (13, 14, 15, 16), producing in effect unreliable results (21).
Moreover, nobody evaluated the thyroid echogenicity as a predictor of RIT outcome, although there is a plausible background to support this hypothesis. Thyroid echogenicity reflects its follicular structure (22). Echoes are primarily generated at the interface between solid tissue and interstitial colloid. Glands with less intensive echoes are likely to have a greater percentage of cells in a given volume, whereas more echoes point toward less abnormal follicular structure, with more colloid, smaller, and less abundant cells. The biological effects come from irradiation of thyroid cells (primarily the follicular epithelial cells), whereas energy deposited in colloid and other extracellular structures has no effect. Based on these considerations, we hypothesized that gland echogenicity reflects its radiosensitivity and, thus, predicts the outcome of RIT. We evaluated this hypothesis in a cohort of patients with Graves’ disease.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
A total of 177 patients with Graves’ disease (149 women and 28 men) are presented in this report. Aside from satisfying the eligibility criteria, they were consecutive patients with Graves’ disease, who presented for the first time at the Department of Nuclear Medicine, University Hospital Split, during the period of 2001–2004, and subsequently relapsed after at least 12-month antithyroid drug treatment and had complete follow-up data after RIT.
Our first-line therapy for the patients with Graves’ disease is antithyroid drug treatment. All patients were treated with methimazole. The patients with disease still active at least 1 yr (from 12–29 months, median 18) after introduction of conservative therapy were advised to have RIT. In patients with related cardiac manifestations or advanced age, the goal was a relief from hyperthyroidism. They received the definitive (ablative) doses of radioiodine that are expected to cure the hyperthyroidism in a great majority of cases, which, however, often results in persistent hypothyroidism (n = 78). The others received the nonablative doses of radioiodine, which are likely to produce euthyroidism but cannot guarantee the cure from hyperthyroidism (n = 99). Thus, in these patients the eventual need to repeat RIT was allowed because of having a solid chance to avoid hypothyroidism and, thus, the lifetime substitution therapy. Because there is no consensus which nonablative dose yields the best results, we tested the two absorbed doses of 100 and 120 Gy. We assumed that the difference of 20 Gy would be enough to produce the clinically relevant differences, yet not too great to support the trivial hypothesis that the greater the dose, the greater the irradiation effect.
To be eligible for this study, the patient had to be without thyroid nodules, with mild or no clinical signs of ophthalmopathy, and sign the agreement to receive the suggested activity of radioiodine and participate in this study, after being informed of the possible therapy dose-dependent outcomes and study goals. Pregnancy was excluded in women in reproductive age. The protocol was approved by the institutional review board.
Clinical and laboratory assessment
The diagnosis of Graves’ disease was based on symptoms and clinical signs of hyperthyroidism, on laboratory results of elevated levels of circulating thyroid hormones triiodothyronine and T4, and on suppressed TSH values and elevated TSH receptor antibodies. The 4- and 24-h radioiodine uptake and ultrasound examinations were also performed in all patients at presentation and follow-ups.
Ultrasound examination
Thyroid ultrasonography was performed with an instrument (Aloka SSD-500; Aloka, Tokyo, Japan) equipped with a 7.5-MHz linear transducer. The standard scanner adjustments included: brightness gain of 72 dB, near gain –7, far gain 2.3, and focal depths at 2 and 3 cm. All examinations were performed by the same, experienced physician (V.M.). As part of a routine necessary to calculate the activity of radioiodine needed to produce the planned target dose, the thyroid mass was estimated from thyroid volume. The volume of each lobe was calculated as its length x width x thickness x 
/6, the volume of isthmus as length x width x thickness, while the thyroid density was assumed to be 1.06 g/ml. As an assessment per investigation, the thyroid echogenicity was ranked in four grades, as described by Zingrillo et al. (23). For the purpose of statistical analyses, the four grades were collapsed into two as: 1) normoechogenicity to mild hypoechogenicity and 2) moderate to marked hypoechogenicity. The ultrasound examination was performed at d 1–3 after cessation of antithyroid therapy.
RIT
At d 3 after cessation of antithyroid therapy, the patients received the tracer doses of radioiodine, and its kinetics was measured the next 5–7 d to enable an estimation of radioiodine effective half-life (T1/2). The next day, i.e. on d 8–10 after cessation of drug therapy, the patients were given the oral therapeutic dose of iodine-131. The nonablative group patients were randomly assigned to the intended target dose of either 100 or 120 Gy. The randomization was simple and unblinded. The ablative group patients were assigned to the intended target dose of 200 Gy. The I-131 activity needed for the intended target dose was calculated using Marinelli’s formula (24):
 |
Follow-up and outcome endpoints
The symptoms of thyrotoxicosis due to irradiation thyroiditis were controlled by ß-blockers. The patients were followed up monthly for the first 3 months, then at least every 3 months. The antithyroid drugs were restarted when there was evidence of persistent or recurrent disease. In this report the outcomes at 12 months are reported: 1) hyperthyroidism (overt hyperthyroidism, treated with methimazole from time when resolved from irradiation thyroiditis, or subclinical hyperthyroidism with suppressed TSH only); 2) euthyroidism (normal thyroid hormones and TSH without medication, with or without previous departures from euthyroid state); and 3) hypothyroidism (overt hypothyroidism, treated with T4, or a subclinical one).
Power and sample size requirements
The study was powered to have an 80% chance to detect as statistically significant (P < 0.05) the 50% difference in the incidence of permanent hyperthyroidism between the patients with hypoechogenic and normoechogenic thyroid. From the pilot study, we assumed that the 12-month incidence of permanent hyperthyroidism in the whole sample would be about 20%. Assuming equal group sample sizes, the anticipated incidences would be 30 and 15%, which would require about 120 patients in each group. The study was terminated earlier upon observing that the impact of the studied predictor on the outcomes was much greater than anticipated.
Data analyses
The 
2 test and ANOVA (or Kruskal-Wallis test in the case of age) were used for categorical and numerical variables, respectively. The adjustment for the intended target dose was done using ANOVA with covariates (analysis of covariance). When the therapy outcome was dichotomized (failure = recurrent hyperthyroidism yes/no), the failure ratios for categorical predictors were calculated. The adjustment for the covariates was done by logistic regression, followed by conversion of odds ratios to relative risks, as suggested by Zhang and Yu (25) for studies with common outcomes, like this one. Maximally, two independent variables were allowed in the logistic regression model to have at least 10 outcomes per variable, as suggested by Concato et al. (21). Therefore, before entering in the model, the variables were screened for their confounding potential by univariate tests, i.e. they had to be associated with both the outcome (failure) and the main predictor. Although the conventional borderline P value of 0.05 was assumed, in light of Bayesian interpretation, the results related to P values larger than 0.05 were commented if expected on sound clinical or scientific grounds. The analyses were performed with SPSS 11.0.1 software (SPSS, Inc., Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The baseline characteristics were typical for patients with Graves’ disease: there were about five times more women than men, with a median age of 48 yr. About 44 and 56% of the patients comprised the ablative and nonablative groups (Table 1
). There were 81 (45.7%) patients with a normoechogenic gland, which included 25 patients with mild hypoechogenicity rank, and 96 (54.3%) patients with a hypoechogenic gland, including 70 with moderate and 26 with marked hypoechogenicity. At 12-month follow-up, there were 25 hyperthyroid, 44, euthyroid and 108 hypothyroid patients. With increasing target dose, the proportion of hyperthyroid outcome decreased, whereas the thyroid volume reduction and proportion of hypothyroid outcome increased, as expected (data not presented). Table 2 shows which of the studied variables differed between patients with a different outcome of RIT. Together with the intended target dose, the strongest predictor of the therapy outcome proved to be the thyroid pretreatment echogenicity, followed by the thyroid pretreatment volume. Normoechogenic and large glands were more radioresistant than hypoechogenic and small glands. Intrathyroid radioiodine effective T1/2 and the thyroid uptake of radioiodine differed slightly among the three outcome groups, whereas the patient age, gender, and T4 before the therapy did not.
|
|
presents the associations of thyroid echogenicity and RIT outcome in each of the three intended target dose groups. The predictive potential of thyroid pretreatment echogenicity was observed in all dose groups. However, the therapy goal was satisfactorily achieved only in some patients’ subgroups.
|
The intended target dose of 120 Gy was most adequate in patients with a normoechogenic gland: the majority of outcomes were euthyroidism, with similar proportion of hyperthyroid and hypothyroid outcomes. If hyperthyroidism and hypothyroidism are given the same weight of adverseness in this group of patients, this is probably "the best that could be done." However, in patients with basal thyroid hypoechogenicity, both intended doses of 100 and 120 Gy induced hypothyroidism in about 70% of cases, suggesting that they required less than 100 Gy the absorbed dose in the gland (Table 3
). There were 17 patients with marked gland hypoechogenicity in this nonablative group (nine and eight of them received 100 and 120 Gy in the gland, respectively). Among them there were 12 hypothyroid, three euthyroid, and two hyperthyroid outcomes.
Patients who received ablative doses
In this group the overall results were better, although the goal of relief from hyperthyroidism was almost 100% fulfilled only in patients with a hypoechogenic gland. This group included nine patients with marked gland hypoechogenicity, and all of them were hypothyroid at 12-month follow- up. Thus, it appears that the ablative doses above 200 Gy may be considered in patients with a normoechogenic gland (Table 3).
Multivariable analyses
In the next analyses, summarized in Tables 4 and 5, we defined the persistent hyperthyroidism as the only adverse outcome (or failure), with all patients included, however. We also attempted to refine the assessments of the predictive potential of gland echogenicity and its volume on RIT outcome by accounting for the possible confounding effect of other patients’ characteristics.
|
|
). However, the radioiodine uptake correlated with basal thyroid volume (linear r = 0.47; P < 0.0001), which was not true for the radioiodine effective T1/2 (r = –0.08; P = 0.27). This suggests that the apparent impact of radioiodine uptake on thyroid radiosensitivity was at least partly due to its association with basal thyroid volume. Thus, to avoid the overadjustment bias in the multivariable analyses of the predictors of thyroid radiosensitivity, the effective T1/2 was considered a covariate in the case of thyroid echogenicity as the main predictor, whereas radioiodine uptake was tested for the confounding effect in the case of thyroid baseline volume as the main predictor. This left us with only two independent variables to be tested in each of the models, which was safe regarding the overfitting bias.
It turned out that normoechogenicity tripled the risk of failure, whereas a large gland doubled it, when compared with a hypoechogenic and small gland, respectively. The impact of normoechogenicity appeared even larger than almost doubling the intended target dose, from 100–120 to 200 Gy. The adjustment for the possible confounders suggested an even greater predictive potential of thyroid echogenicity (Table 4).
Because the individual patient presents with both attributes, the thyroid echogenicity and its size, it is important to investigate the possible combinations of these strong, independent predictors in affecting the outcome of RIT. Although our study was not powered for these detailed analyses, some trends can be seen: the smallest risk of failure was observed in the reference category of a small, hypoechogenic gland. The large gland increased this risk for 30% in the case of a hypoechogenic gland, but doubled it (increasing the relative risk from two to four) in the case of a normoechogenic gland. The adjustment for the intended target dose slightly modified these estimates. Although only some of these figures can claim the statistical significance (Table 5), the overall test of interaction of these two predictors of the therapy failure was significant (P = 0.025). Thus, it appears possible that the impact of gland size is substantial only in the case of normoechogenic glands, whereas hypoechogenic glands are more radiosensitive, with a relatively small influence of the gland size.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Considering the effect of the gland size, our data agree with most of the previous reports on this issue. Large glands appear to be more radioresistant than the small ones, requiring more radiation per unit mass. The independent, clinically relevant effects of other tested variables were not noted. A larger study could reveal whether the effect of gland size is confined to normoechogenic glands only.
At presentation the hyperactive glands are hypoechogenic in the vast majority of cases (26). The reasons are hypertrophy and hyperplasia of epithelial follicular cells, and acceleration of hormone kinetics, which reduces the colloid volume and, thus, the reflecting surfaces. However, the antithyroid medication alters the follicular structure so that before applying RIT, the patients present with a wide spectrum of echogenicity, with increased fraction of less hypoechogenic and normoechogenic glands (23). This explains why we had encountered similar fractions of normoechogenic and hypoechogenic glands in our sample of patients with Graves’ disease. The occasional observations that antithyroid medication is associated with increased radioresistance (17, 18, 19, 20) now find the common background with our findings. It remains to be answered as to why the adverse effects on RIT were consistently reported for propylthiouracil (19, 20), but not carbimazole, or its metabolite methimazole (10, 17, 18, 27), that we used.
To explain our findings, one should consider the kinetics of iodine and the physical properties of the I-131 nucleus, in conjunction with the gland histology. Iodine anions are captured by the thyroid epithelial cell, and it requires only a couple of minutes that the organified form appears in the extracellular colloid, as proven in the rat model (28). Thereafter, the iodine atom, incorporated in the thyroglobulin molecule, has a T1/2 from around 20–60 d (29), which translates to 5- to 7-d intrathyroid effective T1/2 of radioactive iodine. Thus, the vast majority of ionization events originate from the colloid droplets. The radiation dose is mainly due to the ß-minus particle, whereas the accompanying 
-ray contributes negligibly (29). The biological effects are due to irradiation of cellular tissue, primarily the follicular epithelial cells, whereas energy deposited in colloid and other extracellular structures has no effect. In the normal gland, the droplets are relatively large, with some of them over 0.5 mm in diameter (30). The range of I-131 ß-particles in soft tissue varies from 0 to about 2 mm (29), depending on their energy, whereas their penetration depth, due to multiple scatters, can be much shorter. Thus, some lower energy electrons never leave the original droplet, producing no biological effects. The likelihood of these events depends on the distribution of droplet sizes and overall colloid content. The hyperfunctioning gland is mostly cellular, with few, relatively small colloid spaces. On that occasion, the dose distribution is expected to be uniform across the thyroid tissue. Thus, assuming equal average gland absorbed dose, the less echogenic, mostly cellular gland is expected to receive a greater intracellular absorbed dose, which is uniformly distributed across the gland volume, than the more echogenic gland, with nonuniform dose distribution, favoring the colloid droplets.
On the contrary, we find no plausible explanation for the observed radioresistance of large glands observed here as well as in other studies. This effect appeared independent of the thyroid echogenicity, i.e. the large glands were equally often hypoechogenic and normoechogenic. However, the gland echogenicity seemed to moderate the effect of the gland size on the outcome of RIT, confining it mostly to normoechogenic glands. However, an echogram is two-dimensional information, so that the eventual association of thyroid echogenicity and its volume may have been underestimated in this study.
This study was not powered for the detailed analyses, so that some inferences, in particular on the interaction of thyroid echogenicity and thyroid size, remain the speculative ones. We were also unable to quantify the intensity of thyroid ultrasound reflections, relying on the experienced judgment of the operator. Expressing the thyroid echogenicity on the continuous scale would certainly refine our assessments of the underlying histology, which, as we strongly believe, determines the thyroid radiosensitivity. Considering the external validity of the study, one should observe that our patients come from the population with sufficient iodine intake achieved by salt iodination with 25 mg potassium iodide/kg since 1996 (31). Because iodine intake impacts on thyroid physiology, our results cannot be lightly extrapolated to populations with different iodine intake. However, it seems possible that by accounting for the gland echogenicity, one accounts for the iodine intake-dependent differences in gland histology, providing the common denominator for populations with different iodine intakes.
In future research the nonablative target doses well below 100 Gy should be evaluated in the case of a small, hypoechogenic gland, whereas 120 Gy could suffice for medium-sized normoechogenic glands. As for the ablative doses, 200 Gy could suffice for the hypoechogenic gland, whereas substantially greater doses should be tested in patients with a normoechogenic gland.
In conclusion, the thyroid echogenicity and its size must be considered in determining the optimal target dose of radioiodine for therapy of patients with Graves’ disease. The more echogenic the gland, the greater the absorbed dose required, especially in the case of large glands. Failure to account for these two strong predictors of gland radiosensitivity precluded the possibility to titrate the dose of radioiodine to individual patients.
|
Footnotes |
|---|
First Published Online July 3, 2007
Abbreviation: RIT, Radioiodine therapy.
Received April 18, 2007.
Accepted June 22, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  D. Eterovic, V. Markovic, A. Punda, and Z. Antunovic 131I Radiation Dose Distribution in Metastases of Thyroid Carcinoma J. Nucl. Med., May 1, 2009; 50(5): 833 - 834. [Full Text] [PDF]
|
|
|
|
|
|
D. Eterovic, Z. Antunovic, V. Markovic, and D. Grosev Planning of 131I Therapy for Graves Disease Based on the Radiation Dose to Thyroid Follicular Cells J. Nucl. Med., December 1, 2008; 49(12): 2026 - 2030. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
