Introduction

Multinodular Toxic Goiter (MTG) is a hyperthyroid syndrome due to thyroid hormones overproduction by one or more autonomously functioning nodules. MTG as a cause of hyperthyroidism is strictly related to the prevalence of multinodular goiter, which in turn depends upon iodine deficiency (prevalence averaging 5 % of hyperthyroidism in areas with sufficient iodine intake and 20 % in areas of iodine deficiency). Iodine deficiency stimulates thyroid cell proliferation, eventually resulting in development of somatic activating mutations of TSH receptor [1]. Genetic background and female gender, in addition to iodine deficiency, are risk factors for its development.

A large prospective epidemiological study, recently conducted in Denmark, in an area with mild to moderate iodine deficiency, reported the following prevalence among the most frequent causes of hyperthyroidism: MTG 44 %, Graves’ disease (GD) 37 %, and toxic adenoma (TA) 6 % [2].

Radioiodine therapy with 131I-iodide is one of the treatments of choice of MTG, together with surgery [3–5]. In fact, antithyroid drugs, as a rule, are unable to achieve hyperthyroidism cure. Radioiodine is selectively taken up by thyroid tissue, resulting in volume and function reduction.

Two different approaches are used to choose the 131I-iodide activity to be administered: a fixed “maximal” activity (typically 600 MBq) or an activity derived from a dosimetric calculation, that is, the computation of the lowest activity which is able to deliver an acceptable radiation dose to the target tissue [6–8]. Although these methods are both effective, the As Low As Reasonably Achievable (ALARA) principle should be always pursued, avoiding unjustified radioactivity exposition to the patient himself, to people living/working near him and to environment.

Dosimetric calculation implies the measurement of 131I-iodide uptake by external counting after administration of a tracer activity and evaluation of functioning target tissue volume.

In this study, we applied a multi-parameter dosimetric approach to a group of patients with MTG. This algorithm was created and clinically validated for GD [9–11]. We successfully applied this protocol to single hyperfunctioning thyroid nodule patients [12]; yet there are no published studies about its use in the MTG treatment. The present study highlighted a fair therapeutic efficacy with a relatively low incidence of hypothyroidism.

Materials and methods

Patients

We prospectively evaluated a group of 93 consecutive patients (28 males and 65 females, aged between 43 and 84, median 71 years) with MTG and functionally autonomous areas as detected by thyroid 99 mTcO4 − scintigraphy. In all of them, the thyroid scan showed a partial or complete functional inhibition of extra-nodular tissue. Patients with a single autonomously functioning thyroid nodule, without other clinically relevant US-detectable nodules, were excluded from the study.

Among them, 22 patients (24 %) had overt hyperthyroidism and 71 (76 %) subclinical hyperthyroidism.

All the patients with subclinical hyperthyroidism who underwent radioiodine treatment had TSH below 0.1 mU/L. They complained of cardiovascular symptoms/signs, such as tachycardia or arrhythmias, or had progressively worsening osteopenia or osteoporosis. In some cases, an additional indication was to obtain a clinically significant total thyroid volume reduction. Patients with clinical hyperthyroidism underwent, as a rule, a short course of low-dose methimazole therapy (10–20 mg daily) before radioiodine: antithyroid drugs were discontinued at least 3 weeks before measurement of thyroid 131I-iodide uptake (see below). Follow-up visits were systematically performed at 7, 15, 30 days and 6 months after radioiodine therapy, and thereafter every 6 or 12 months (according to clinical conditions and thyroid function). Patients who were still hyperthyroid following the first administration of radioiodine were given a second administration 6 months later, which invariably resulted in hyperthyroidism cure (not shown).

Methods

Free thyroid hormones and TSH were measured in basal conditions (i.e., immediately before treatment), and during follow-up. TSH (3rd generation ultrasensitive assay) and free T3 and T4 assays were carried out through standard immunometric methods.

Shortly before treatment, thyroid 131I uptake was evaluated by direct external measurements at 4th, 24th and between 96th and 120th h following administration of a trace activity (about 370 kBq of 131I-iodide) with a 2″ × 2″ thallium-activated sodium iodide scintillator [NaI(Tl)] probe; an acrylic-water phantom was used for probe calibration.

Total thyroid and nodules volumes were evaluated by means of US [13] in basal conditions and at different times from 131I-iodide therapy (after 1–6–12 months). In order to exclude iodine excess, 24-hour urinary iodine excretion was measured by the Sandell–Kolthoff method [14]. The prescribed absorbed dose to the hyperfunctioning tissue ranged between 250 and 300 Gy for MTG.

Dosimetry and statistical analysis

In order to estimate the activity (A, measured in MBq) to be prescribed for each patient, we referred to the method developed by Traino et al. [9] and reported in the above guidelines, specifically developed for GD treatment. This method takes into account, in addition to 131I kinetics, the volume reduction of functioning thyroid tissue after radioiodine administration (TVR). We applied such a method also to the radioiodine therapy for MTG. The algorithm for TVR is the following:

$$A\left( {MBq} \right) = \frac{{D_{\text{T}} \cdot m_{ 0} }}{{U_{\hbox{max} } \cdot T_{{ 1 / 2 {\text{eff}}}} }}\left( {5,656 - 5,08 \times 10^{ - 5} \frac{{D_{\text{T}} \cdot m_{0} }}{{U_{\hbox{max} } }}} \right),$$
(1)

where D T is the target absorbed dose (Gy); m 0 the target (MTG) initial mass (g) evaluated by US (ellipsoid approximation); U max the maximum uptake of intra-thyroid radioiodine kinetics (fraction) and T 1/2eff is the effective half-life.

As a rule, total thyroid volume derived from US was used for estimating m 0. In 27 MTG patients, who showed only one or two “hot” nodules at thyroid scintigraphy, with near-total functional inhibition of the remaining tissue, we considered only the volume of the hyperfunctioning nodule(s) to estimate the target mass m 0: this volume actually accounted for 8–56 % of total thyroid volume.

When the calculated prescribed activity exceeded the maximum allowed for outpatients by our national law (600 MBq), a standard upper threshold activity (600 ± 5 % MBq) was administered; when it was lower than what can be considered a clinical minimum, a lower threshold activity (about 150 MBq) was administered.

This implied, in some cases, a nodule absorbed dose lower/higher with respect to the one that had been calculated by the algorithm. Another imprecision factor was the exact time of radiopharmaceutical administration: 131I was provided as sodium iodide caps with standard activities, and physical decay was used to reach the desired activity. As a rule, however, dose imprecision did not exceed 10 %.

Each prescribed activity was compared with both the following, more common, dosimetric approaches.

(a) Medical Internal Radiation Dose (MIRD) algorithm based on Snyder formula [15], which supposes a mono-exponential intra-thyroid radioiodine kinetics (symbols as before)

$$A\left( {MBq} \right) = 5,829 \cdot \frac{{D_{\text{T}} \cdot m_{0} }}{{U_{\hbox{max} } \cdot T_{{ 1 / 2 {\text{eff}}}} }}.$$
(2)

(b) Simplified Marinelli–Quimby algorithm (MQ), based on the previous one, with a basic intra-thyroid radioiodine kinetics. U is approximated by a single 24th h radioiodine uptake measurement (U24), and a standardized effective half-life (132 h) is considered [9]. Both Spearman and Mann–Whitney tests were used when appropriate for statistical analysis. The Life-Table method [16, 17] was used to estimate the expected prevalence of hypothyroidism. This method allows the calculation of a cumulative incidence of hypothyroidism at various time intervals, among cured patients, taking into account the variability in the length of follow-up period.

Results

After a follow-up of 6–120 months (median 60), 76 out of 93 patients (82 % of total, 23 male, 53 female) achieved hyperthyroidism cure, including both patients who reached normal thyroid function and those who developed hypothyroidism.

Among them, 59 patients achieved euthyroidism 1–6 months after treatment, while six patients were initially hyperthyroid and achieved euthyroidism later (8–24 months, respectively, one after 8 months; three after 12 months; one after 15 months and one after 24 months). In the subgroup who developed hypothyroidism, five patients started to display reduced thyroid function in the first 6 months after treatment, while seven patients became hypothyroid 10–48 months after treatment (two after 10 months, one after 12 months, one after 18 months, one after 28 months, one after 36 months, and one after 48 months). Seventeen patients (18 % of total, 5 male and 12 female) had persistent or recurrent disease.

In summary, we observed hypothyroidism (both early and late) in 12 patients out of 93 (13 %), euthyroidism in 64 patients (69 %), and persistent disease or relapse in 17 patients (18 %). The prevalence of hypothyroidism was greater in females than in males (11 females vs. 1 male).

Using the Life-Table method to estimate the cumulative incidence of hypothyroidism among cured patients, a 21 % incidence at 8.5 years after treatment was estimated (10 years maximum follow-up period, 79 ± 6 % SD average percentage of euthyroid patients, Table 1).

Table 1 Life-Table method evaluation of euthyroidism probability over time in cured patients
Full size table

Despite the inhomogeneous distribution of radioiodine in the thyroid gland, ultrasonographic follow-up after approximately 1 month demonstrated a volume reduction significantly correlated (R 2 = 0.77, P < 0.001) to what predicted by TVR (see Fig. 1); no significant correlation between administered activity and volume reduction was found. We did not observe significant differences in median thyroid volume reduction at 1 month, since it resulted 12 % (7–51 %) in patients who achieved euthyroidism, 25 % (22–50 %) in patients become hypothyroid, and 21 % (19–41 %) in patients with persistent/relapsed disease. Similarly, no significant differences among the three groups of patients were observed 6 months (8–71 % in patients become euthyroid, 11–75 % in patients become hypothyroid, 11–70 % in patients still hyperthyroid) and 12 months following therapy (10–71 % in patients become euthyroid, 11–85 % in patients become hypothyroid, data on patients with persistent hyperthyroidism are not available, as they were given a further radioiodine treatment). The correlation between observed and expected thyroid volume was maintained at the long-term controls (data not shown).

Fig. 1
figure 1

Correlation between expected (model predicted) and observed (US measures) glandular volume, 30 days after 131I-iodide therapy, in patients with MTG (P < 0.001, Spearman test)

Full size image

The median activity to be administered, calculated with TVR method, was 526 MBq (156–600 MBq), lower than the activities that would have been obtained using other methods (median reductions: −13 % compared to fixed activity method, −38 % compared to MQ, and −15 % compared to MIRD), as shown in Fig. 2. In 25 patients, the activities calculated for administration with TVR method exceeded the activity threshold for outpatients established in Italy, so they were treated with a standard activity (600 ±  5 % MBq), thus receiving a dose lower than 250–300 Gy; among these, 19 patients were cured and only six relapsed, with no significant difference in received dose.

Fig. 2
figure 2

Distribution of radioiodine activity prescription, according to different algorithms (MQ, MIRD, TVR) and actually administered. Dotted line indicates the 600 MBq hystogram class, corresponding to upper limit activity

Full size image

In our study, a widespread distribution of 131I-iodide individual effective half-life is noticeable, although most of the patients’ T 1/2eff was in the range 150–200 h (6–8 days) (Fig. 3). An estimated effective half-life higher than the physical half-life of 131I (192 h) is in fact a paradox, due to the fact that it results from a semi-log fitting based on two points only (2nd and 3rd in vivo measurements), in the hypothesis of a single-order kinetics and a maximum uptake around the 24th h (see also discussion).

Fig. 3
figure 3

Distribution of estimated 131I-iodide effective half-life (T 1/2eff), calculated from patients measurements, using a single-order kinetics approximation (see text for further details)

Full size image

Furthermore, we compared the outcome of patients with relapsed or persistent disease after radioiodine therapy with those who were cured after the first treatment, to assess whether there was any difference in terms of kinetics of radioiodine and risk factors for recurrence.

The median T 1/2eff was 174 h (121–330 h) in relapsed patients and 175 h in cured patients (84–511 h), the median maximum uptake was 51.2 % (15.6–67.1 %) in relapsed patients and 37 % (9.4–68 %) in the cured ones. There were no significant differences in female/male ratio (relapsed 2.4/cured 2.3), in age (mean male age: relapsed 72 years/cured 71.8 years; mean female age: relapsed 69 years/cured 64.4 years), in smokers/ex-smokers prevalence (relapsed 63 %/cured 65 %) and in clinical presentation. The pre-treatment glandular volume calculated with the rule of the ellipsoid was significantly greater (P < 0.0001) in relapsed patients (median initial volume: 72.2 cc) than in patients cured after the first treatment (median initial volume: 32.0 cc) (Table 2).

Table 2 Relapsed/cured patients features after treatment with 131I-iodide
Full size table

The 131I-iodide activity administered to relapsed patients (median 605 MBq; range 307–625 MBq) was significantly greater (P = 0.01) than the dose administered to cured patients (median 471 MBq; range 156–615 MBq); this is probably related to the gland volume differences described above. Of note, three patients, after 6–12 months of treatment, showed a slight positivity of TSH receptor antibodies (TRAb): two of them relapsed and one achieved euthyroidism later (18 months after treatment).

Discussion

In this study, we applied a multiparametric dosimetric approach (TVR), developed by Traino in 2001 for the treatment of GD [8], in patients with MTG and evaluated the results in terms of therapeutic efficacy.

This dosimetric approach does not take into account only the intra-thyroid radioiodine kinetics (uptake and elimination phase) for the calculation of the activity to be administered, but it also includes the volume reduction during and after the treatment of ‘target organ’ (hyperfunctioning tissue), which significantly affects the absorbed dose.

The application of this approach would imply a homogeneous radioiodine distribution within the gland. While this assumption is acceptable for the single hyperfunctioning thyroid nodule, it is much less plausible for MTG considered in the present series, where areas with different radioiodine uptake coexist within the gland. Despite this theoretical limit, the correlation between thyroid volume reduction estimated by applying this method and thyroid volume reduction observed by US, was highly significant and a great majority of patients were cured. In addition, the results obtained show that the use of TVR allows to administer radioiodine activities slightly lower than the “fixed” activities without any decrease of therapeutic effectiveness [18].

In patients (27) who showed only one or two hyperfunctioning nodules at thyroid scintigraphy and a near-total functional inhibition of surrounding tissue, we attempted to obtain a more accurate dosimetry, by taking into account only the nodular volume in the dose estimation calculation. Despite this, four patients relapsed, the effectiveness of the treatment being comparable to that observed for the entire series; to date, none of them developed hypothyroidism.

Half-life determination results crucial for activity estimation, and it is probably in most part responsible for the important differences observed among the aforementioned algorithms (namely, between MQ, that does not take it into account, and the others). Its determination implies at least one additional measuring session for the patient, and this is the main reason behind the use of simplified protocols.

As previously mentioned, the finding of effective half-life higher than the physical half-life of 131I (192 h) is a paradox, due to the use of a single-order kinetics. This is obviously a simplification, that in some cases leads to inaccuracy. In fact, in some cases, a second-order kinetics should be more suitable to describe the actual behavior. This is also evident from the iodide biological uptake, which turns out to be still rising after 24 h. In these cases, a better fit should involve more than two points and a “second order effective half-life” should be considered. However, this is hardly feasible in the daily clinical practice, even if this concept has been applied in selected cases; the improvement of accuracy in the activity calculation did not substantially impact the planned therapy. Thus, in our opinion, a delayed last measurement in the range 96th–120th h minimizes, as a rule, the need for further measurements. In fact, the cumulated activity integral, hence the dose, depends mostly on the first-order exponential.

Our endpoint was the achievement of stable euthyroidism; this objective was attained in the majority of patients (69 %), at least during the observed period. Additional 12 patients (13 %, 11 female, 1 male) were cured from hyperthyroidism, but developed hypothyroidism during follow-up. The estimated perspective cumulative risk of hypothyroidism was performed by the Life-Table method, and showed a risk of 20.7 % at 8.5 years after treatment.

Altogether, these results, both in terms of effectiveness with a single administration (main endpoint) and in terms of minimizing the risk of hypothyroidism (secondary endpoint), are fair, when compared with those reported in literature [3–5] In one long-term follow-up study on patients treated with radioiodine, a prevalence of hypothyroidism of 3 % at 1 year and 64 % at 24 years was shown [19]. It is likely that the administration of lower activities with respect to other dosimetric methods can also offer advantages in terms of reduction of the long-term hypothyroidism incidence.

Twenty-five patients were treated with lower activity than predicted by the TVR method, as the calculated activities exceeded the maximum for outpatients: of them, 19 were cured and only six relapsed, with a similar percentage of cure as the whole set; thus, it is possible to achieve permanent clinical remission with radiation doses lower than what established/scheduled. As a matter of fact, on the basis of our results, it is possible to cure the majority of patients with MTG in an outpatients setting.

As mentioned above, the dosimetric protocol we applied in the present study was originally set up for GD treatment: both the proposers [9, 10] and our group [11] obtained good clinical results in terms of GD cure, that is, functional thyroid ablation and hypothyroidism. In a more recent study, we used the same protocol in single hyperfunctioning thyroid nodule [12]. In such condition, the clinical endpoints (hyperthyroidism cure, preservation of thyroid function) are the same as in the present study, and the results have been even more favorable in terms of hyperthyroidism cure; hypothyroidism incidence was apparently low, although a reliable prospective incidence estimate was not possible due to the low number of patients actually hypothyroid. This is not unexpected, since the model assumes a homogeneous radioiodine distribution within the functioning tissue and fits far better with single hyperfunctioning nodule than with MTG.

Furthermore, we comparatively evaluated two groups: cured and relapsed/resistent (after a single treatment) patients to analyze the possible causes of treatment resistance. We did not find any significant difference in age at the time of therapy, in female/male ratio (for both there is a higher female prevalence), smoking habits, or in presence of comorbidities.

Conversely, pre-treatment thyroid volume was significantly higher in relapsed patients (both genders) than in cured ones; in addition, we did not identify any significant difference in intra-thyroid radioiodine kinetics (T 1/2eff and maximum uptake). As a consequence, the median-administered activity to relapsed patients was significantly higher due to a greater measured target volume, rather than to disadvantageous intra-thyroid radioiodine kinetic.

In order to improve the therapeutic effectiveness, it may be considered to increase the prescribed dose only for greater goiters (namely >70 cc, the median relapsed patients’ volume), although some of them would require an inpatient treatment.

In addition, it should be noted that three patients (two relapsed and one with clinical remission after 18 months) developed positivity for TSH receptor antibodies that may have concurred to the poor response to therapy, as it is described in literature [20].

In any case, the individual radiosensitivity of thyroid gland is probably a critical variable affecting the clinical outcome of radioiodine therapy in single patient, although we are not able to quantify it, and may be responsible, at least in part, of the observed variability.

In conclusion, considering that the radioiodine therapy of hyperthyroidism is definitely the most widely employed nuclear medicine therapy, the results of this study show that an effective cure of MTG can be obtained with relatively low 131I-iodide activities, and probably with a relatively low incidence of hypothyroidism, using the TVR algorithm. We think this approach should be preferred, when possible, to the administration of fixed 131I-iodide amounts.