Introduction

For many years the recommended therapy for differentiated thyroid carcinoma (DTC), with the exception of microcarcinomas, has consisted of (near) total thyroidectomy followed by post-operative radioiodine ablation of thyroid remnant tissue.

Successful ablation leads to a better prognosis with regard to both recurrence-free and overall survival [1] and eliminates the difference in recurrence rate between pre-ablation low- and high-risk patients [2]. Even though results of randomized controlled trials are still missing, the combination of surgery and 131I ablation has proven its value as a safe and very effective treatment and is an integral part of various international guidelines [3–7].

The amount of 131I required to achieve successful ablation is still subject to debate, and several trials have been performed or are currently running to provide valuable answers. Thus far, various activities in a bandwidth between 1 and 3.7 GBq 131I have been reported to result in an optimal success rate of ablation. Higher activity seems to coincide with higher success rates [8], although most studies in the literature find no further improvement in success rates when ablation activities exceed 1,850 MBq (50 mCi) [9–11].

However, the cumulated activity per volume of blood might be a more adequate predictor for therapeutic success than the administered activity [12]. The amount of 131I still available in the blood depends on the 131I excretion rate which varies considerably between individual patients. This results in large differences in effective half-life and consequently in the residence time of the activity per volume of blood. The determinant for a successful 131I ablation is the radiation absorbed dose to the target tissue; the decisive parameters for this are the administered therapeutic activity and the retention of radioiodine in the target volume [13].

Target tissue uptake must be expected to depend on the availability of 131I in the blood, i.e. the 131I residence time of the activity per volume of blood. This residence time correlates well with the radiation absorbed dose to the blood (BD) per administered 131I activity [12]. Consequently, the absorbed dose to the target tissue and thus the probability of successful 131I ablation should be associated with the BD.

The aim of the current study was to investigate the relation between the radiation absorbed dose to the blood and the success rate of ablation, and compare this to the relation between the administered 131I activity and the ablation success rate.

Materials and methods

Patients

We retrospectively reviewed the files of DTC patients who underwent 131I ablation after thyroid hormone withdrawal in our hospital in the time period from 1993 to 2007 in order to identify those without distant metastases and with a thyroid-stimulating hormone (TSH)-stimulated follow-up including available thyroglobulin (TG) level measurement and 131I whole-body scan about 6 months later. Patients who received reduced activities (<1,850 MBq) in order to prevent adverse effects in the presence of large thyroid remnants (defined as showing a thyroid remnant uptake >10% with the method used at the time) were excluded. The study included 449 patients eligible according to these criteria.

Tumour staging

Surgical specimens were analysed and classified according to standards prevailing at the time of initial treatment. The present study used the histological and TNM classification given in the original pathology report (5th edition of the TNM system until 2002 [14], 6th edition from 2003 onwards [15]). Criteria for excluding patients from this study due to the presence of distant metastases were (a) histological evidence of distant metastases or (b) non-histological evidence such as elevated TG levels, a post-therapy 131I whole-body scan showing 131I uptake outside the neck, computed tomography scan or magnetic resonance imaging.

Initial treatment

All patients underwent a total thyroidectomy, with some patients also undergoing additional central and/or lateral lymph node dissection. After thyroidectomy patients were left without thyroid hormone substitution until 131I ablation, which took place 4–6 weeks after surgery. Prior to ablation cervical ultrasound was performed and the 24-h remnant uptake was measured using a small 131I tracer activity <10 MBq in order to avoid any possible thyroid remnant stunning [16–18]. The administered ablation activities ranged mainly from 1,850 to 4,300 MBq with a trend towards on average higher activities in later years. After ablation, all patients received TSH-suppressive levothyroxine (LT4) treatment to maintain serum TSH at <0.1 mIU/l.

Estimation of the absorbed dose to the blood

Due to German regulations on radiation protection, each 131I ablation had to be performed as an inpatient procedure. All patients spent between 2 and 7 days in our dedicated radionuclide therapy ward until the radiation levels fell below the advised limits for discharge. Radiation levels were monitored by ceiling dose rate metres built in 2 m above each patient’s bed. The first measurement was acquired up to 2 h after administration of the ablative activity, after which measurements were performed twice daily at 12-h intervals. The 48-h whole-body 131I retention was determined for each patient from a bi-exponential decay curve fitted to the measurements. This retention value was used to calculate the BD using a method recently introduced by Hänscheid et al. [19]. Briefly, a mathematical relationship between radioiodine retention in the whole body and the radiation absorbed dose to the blood was deduced based on the assumptions that the whole-body activity decays exponentially and that 14% of the whole-body residence time can be attributed to the blood [19]. The mean of the absolute deviations between blood dose estimates obtained with the formalism and actual blood doses was found to be 14% if the whole-body retention was measured 2 days after radioiodine administration [19]. A more elaborate description of the methods used in this study is given in the online Electronic supplementary material.

Follow-up TG and diagnostic whole-body scan

All patients received a TSH-stimulated follow-up evaluation approximately 6 months after 131I ablation. Serum TSH levels were elevated either by LT4 withdrawal, or in more recent years also by intramuscular injection of recombinant human TSH (rhTSH). During TSH stimulation, TG levels were measured and a diagnostic whole-body scan (dxWBS) using 150–300 MBq 131I was acquired accompanied by ultrasound imaging of the neck. Success of ablation was assessed using two different criteria. For the main criterion 1, ablation was considered successful if the 131I dxWBS was negative, TG levels were undetectable and no other evidence of persistent disease was present. For the secondary criterion 2 we only considered the dxWBS results. In cases of known persistent thyroglobulinemia, patients were treated with therapeutic 131I activity without preceding 131I dxWBS. These patients were considered to have unsuccessful ablation according to criterion 1 and were excluded from statistical evaluations in connection with criterion 2. Even though this will cause a slight overestimation of the success rate according to criterion 2, it is not possible to include post-therapy scans in this analysis as the sensitivity of a post-therapy scan is much higher than that of a dxWBS [20].

Laboratory analyses

Because different TG assays were used over the inclusion period of this study, classification of TG levels as “undetectable” was based on the lower detection limit of the assay used in a given follow-up exam rather than on a single cutoff value for all assays. The lower detection limit was 1, 0.3 and 0.2 μg/l in 11, 22 and 67% of the patients, respectively. The presence of antibody interference, e.g. from anti-TG antibodies (TGAb) or heterophile antibodies, was excluded through determination of TG recovery rates (normal range 70–130%). Since undetectable serum TG levels cannot be interpreted as evidence of remission in the presence of antibody interference [21], patients would have been excluded from the study in the presence of insufficient TG recovery during follow-up; no patients however showed either of these two conditions.

Statistics

A p value < 0.05 was considered to indicate statistical significance. The tests used are reported with the results. When normally distributed, data are given as mean ± standard deviation, otherwise median values are used.

Results

Of the 449 patients aged 47 ± 15 years (median 47, range 11–84 years), 122 were male and 327 were female; 98 patients had a follicular and 351 patients a papillary thyroid carcinoma. There were 296 patients initially staged as pN0, 61 as pNx and 92 as pN1.

The median administered 131I activity was 3.38 GBq (range 0.98–7.0 GBq). Three patients received less than 1.85 GBq and three patients with poor histology or locally invasive disease were treated with more than 4.3 GBq (up to 7.0 GBq). The median BD was 304 mGy (range 86–943). BD <200 mGy and BD >450 mGy were observed in 8% of the patients each.

Criterion 1

At the time of follow-up, which was performed a median of 5.9 months after 131I therapy, 254 of 449 (56.6%) patients had successful ablation according to criterion 1. In order to investigate the effects of BD and administered activity on the ablation rates, patients were stratified to intervals of 50 mGy and 0.1 GBq, respectively. Subgroups with less than 30 patients were pooled to decrease statistical variation and to harmonize group sizes. Results listed in Table 1 and illustrated in Fig. 1a (filled dots) indicate that the success of ablation increases with increasing BD (Pearson’s r = 0.94). Spearman’s ρ = 0.82 as a non-parametric measure of statistical dependence shows significance with p = 0.03. Probability for successful ablation is significantly higher for patients with BD >350 mGy (2 × 2 chi-square test: p = 0.03) as compared to lower BDs. In contrast, no dependence of the ablation rate on the administered activity is demonstrable [Table 2, Fig. 1b (filled dots)]. The observed tendency to higher ablation rates for administered activities exceeding 3.2 GBq does not reach statistical significance (2 × 2 chi-square test: p = 0.1).

Table 1 Rates of successful ablation after stratification to intervals of absorbed dose to the blood. Criterion 1 was matched if in addition to visually negative diagnostic whole-body scintigraphy the concurrent TSH-stimulated serum TG level was undetectable. Successful ablation according to criterion 2 was assumed in cases of visually negative diagnostic whole-body scintigraphy only
Full size table
Fig. 1
figure 1

Rates of successful ablation in subgroups of patients after stratification to intervals of absorbed dose to the blood (a) and administered activity (b). Criterion 1 (filled dots) was matched if in addition to visually negative diagnostic whole-body scintigraphy the concurrent TSH-stimulated serum TG level was undetectable. Successful ablation according to criterion 2 (open circles) was assumed in cases of visually negative diagnostic whole-body scintigraphy only. Values are drawn at the median of the correspondent interval

Full size image
Table 2 Rates of successful ablation after stratification to intervals of administered activity. Criterion 1 was matched if in addition to visually negative diagnostic whole-body scintigraphy the concurrent TSH-stimulated serum TG level was undetectable. Successful ablation according to criterion 2 was assumed in cases of visually negative diagnostic whole-body scintigraphy only
Full size table

Criterion 2

Of the 449 patients, 24 received an immediate second 131I therapy without preceding 131I dxWBS. Of the remaining 425 patients, 323 (76.0%) had successful ablation according to criterion 2. After stratification to identical intervals as described before, results are similar to those obtained with criterion 1 [Tables 1 and 2, Fig. 1 (open circles)]: ablation rates increase with BD (Pearson’s r = 0.78, Spearman’s ρ = 0.79, p = 0.04) but not with the administered activity, and the probability for successful ablation is significantly higher for patients with BD >350 mGy (2 × 2 chi-square test: p = 0.02).

Patients with identical activity

In order to rule out a potential effect of the selection of the administered activity on the results, the analysis was repeated for a subgroup of 209 patients with almost identical therapeutic activities of 3.5 GBq ± 5% of 131I (Table 3). The distribution of the BD is shown in Fig. 2 (bars) together with the observed ablation rates according to criterion 1 (filled dots). The BD ranged from 164 to 943 mGy (median 325 mGy) showing minor deviations from a normal distribution. The results obtained for all patients are confirmed in the subgroup of patients receiving 3.5 GBq ± 5%: ablation rates increase with BD (Pearson’s r = 0.93, Spearman’s ρ = 0.90, p = 0.02) and the probability for successful ablation is significantly higher for patients with BD >350 mGy (2 × 2 chi-square test: p = 0.03).

Table 3 Rates of successful ablation of patients treated with 3.5 GBq ± 5% 131I after stratification to intervals of absorbed dose to the blood. Criterion 1 was matched if in addition to visually negative diagnostic whole-body scintigraphy the concurrent TSH-stimulated serum TG level was undetectable. Successful ablation according to criterion 2 was assumed in cases of visually negative diagnostic whole-body scintigraphy only
Full size table
Fig. 2
figure 2

Distribution of blood doses in 209 patients receiving therapeutic activities of 3.5 GBq ± 5%. The ordinate shows both numbers of patients and percentage of successful ablation (filled dots) in five blood dose intervals

Full size image

Multivariate analysis using a binary logistic regression with a forward selection based on likelihood ratios showed that for successful ablation according to criterion 1 an absorbed dose to the blood ≥350 mGy (p = 0.032) and the presence of lymph node metastases (p = 0.01) were independent significant predictors of ablation success, whereas T stage (p = 0.12) and histology (p = 0.06) were not significant and the age at the time of ablation (p = 0.85) was unimportant. For ablation success according to criterion 2 the absorbed dose to the blood was the only significant predictor of ablation success (p = 0.018); neither the lymph node status (p = 0.18) nor T stage (p = 0.41), histology (p = 0.72) or age (p = 0.98) were relevant in this analysis.

Discussion

Many studies have already tried to determine the optimal ablation activity, either under withdrawal or under rhTSH stimulation [9, 10, 22–24]. In most of these studies it was generally found that lower activities of about 1,100–1,850 MBq did not result in a significantly worse rate of successful ablation than did a higher activity of around 3,700 MBq. Accordingly no effect of increasing activity on the ablation rate is verifiable in our collective of 449 patients, 446 of whom received activities of 1,850 MBq or more. An effect of the blood dose on the ablation rate is apparent, which however is in contradiction with findings published recently by Flux et al. [25]. They found that BD was lower in patients with successful ablation than in those with unsuccessful ablation in a study of 23 DTC patients. A potential explanation of this discrepancy could be that the median absorbed doses to blood observed by Flux et al. are substantially lower than the absorbed doses observed by other authors [12, 26, 27]. Although Flux et al. made direct measurements of serum 131I activity, no samples were collected within the first 24 h after administration. Disregard of unnoticed short components in the blood activity functions potentially affecting the blood dose calculation might be a reason for low and unreliable BD values.

As many different criteria have been used in the literature to define successful ablation, absolute values for ablation rates are often difficult to compare. Mostly we can compare our results for criterion 2 (negative 131I dxWBS). Our rate of successful ablation of 76% is of the same order as those reported by e.g. Pacini et al. [28] and Verkooijen et al. [29] or in two studies by Barbaro et al. [30, 31], in which the rate of successful ablation in different groups ranged from 58 to 85%. It is, however, somewhat lower than that reported in studies by e.g. Pilli et al. [23], Chianelli et al. [32], Robbins et al. [33] or de Klerk et al. [34], who reported rates of successful ablation according to criterion 2 ranging from 88 to 100%. Criterion 1 can only be compared to a limited number of studies. The rate of successful ablation of 57% is comparable or slightly better than in a study by Verkooijen et al. [29], who reported a success rate of 43–56%, depending on the ablation protocol used, and two studies by Verburg et al., which reported a success rate of 33–65%, depending on whether or not a pre-ablation diagnostic 131I activity of 37 MBq was given [18] or 61% in total [1].

The current study has used a number of different TG assays over time and whole-body scans were acquired using a number of different cameras; also the method of TSH stimulation for follow-up has, especially in later years, varied. This heterogeneity in data may affect the comparability with external data but not the conclusion of the present study.

Another source of uncertainty is the method of blood dose calculation. Hänscheid et al. [19] were able to prove a good correlation between the BD and blood dose estimates from a single whole-body retention measurement in an investigation mainly based on remnant ablation data in patient groups similar to the collective included in the present study. Therefore, it is unlikely that our results are biased, e.g. by the presence of 131I accumulations in the thyroid remnants. The inherent uncertainty introduced by the blood dose estimate is 14% under study conditions. Retrospective application of this method to data gathered less attentively in routine daily practice will probably enlarge this error. However, statistical errors in blood dose calculation will more likely blur rather than introduce a bias to the effect of the BD on the ablation rates.

Despite the limitations of our study, it is obvious that in our large series of patients the BD is a better predictor of the success of ablation than the administered ablative 131I activity. This is only to be expected; theoretical considerations imply that the residence time of the activity in the blood is a determinant for the remnant uptake and thus the dose to the target tissue [12]. Particularly in patients with a high renal iodine clearance rate, any circulating 131I will be excreted quickly, eliminating its availability for any iodine-avid tissue, whereas, in a patient with slow clearance, the thyroid remnant will have a considerably higher amount of 131I circulating in the blood at its disposal, in turn resulting in a larger thyroid remnant absorbed dose.

The results of the present study, which need to be confirmed in independent investigations, have a multitude of clinical implications. In the debate on the dosage necessary for successful 131I ablation, attention has been focused too much on the amount of fixed activities of 131I that should be administered. The radiation absorbed dose to the thyroid remnant must be considered to be the principal determinant of outcome. It is closely related to the amount of circulating 131I, i.e. the number of radioactive decay events in the blood, which is the main determinant of the BD. The absence of significantly improved ablation rates for higher ablation activities in most studies might be due to the considerable variation in iodine kinetics. This variation degrades the correlation between BD and administered activity, which was found to be only r = 0.36 for the patients included in this study, and deteriorates the correlation between administered activity and remnant absorbed dose.

The question has recently been asked about how low the ablative activity may be when performing rhTSH-stimulated ablation [35]. Although it has been shown that there was no significant difference in ablation rates between levothyroxine withdrawal and rhTSH stimulation [28, 36], it has also been shown that iodine clearance is considerably faster under rhTSH stimulation, resulting in a significantly lower absorbed dose to the blood [12, 26, 37]. On the other hand, the 131I half-life in thyroid remnants is longer under rhTSH stimulation [12, 26, 37]. These differences in 131I kinetics makes it impossible to apply the values from the present study to rhTSH-stimulated 131I ablation.

Individualized therapy ideally targets at the optimal radiation absorbed dose to the thyroid remnant and is based on a pre-therapeutic dosimetry of both red marrow and target dose per activity administered. Usually target dosimetry is not feasible as it requires 131I activities potentially inducing stunning and information on the target mass, the S factor of self-irradiation and the homogeneity of the activity concentration in the remnant. Hänscheid et al. [12] suggested that therapy be individualized by generally adjusting the administered activity according to the residence time of the activity per volume of blood or, as a surrogate, that one aim at a fixed radiation absorbed dose to the blood rather than administering fixed activities. As an initial guess, 400 mGy was stated to be adequate as BD for remnant ablation [12]. This value was modified to 300 mGy in a recalculation published in [19]. The current study suggests that the maximum ablation effect will be attained if the BD is at least 350 mGy. Further improved ablation rates at higher doses cannot be excluded from our data, which indicates that the optimal BD might be a compromise between ablation rate and radiation exposure.

Figure 3 shows the distribution of activities necessary to target a blood dose of 350 mGy for the patients included in this study. Values range from 1.1 to 7.3 GBq with the median at 3.7 GBq. Activity exceeds 5 GBq in 13% of the patients. Using the method described by Hänscheid et al. [19], the activity required to achieve this BD can easily and reliably be determined in a simple pre-ablation dosimetry procedure using <10 MBq of 131I in order to determine the 48-h whole-body 131I retention.

Fig. 3
figure 3

Distribution of activities necessary to target a blood dose of 350 mGy in 449 patients

Full size image

Although the current results clearly show a beneficial effect of a higher absorbed dose to the blood on the rate of successful ablation, it still remains a retrospective study with all the limitations that come with such an assessment. The implications of the present study are however of a magnitude that could significantly change the way we practice nuclear medicine; the best way to ascertain the potential of the concept laid out here would be to perform a randomized trial comparing a fixed activity of e.g. 3,500 MBq vs blood dose-based dosimetry using the methodology described in this paper, aiming for a blood dose of at least 350 mGy.

Although we have no data to analyse this in the current collective, it is of course possible that an increase in blood dose will also lead to an increase in the frequency of early and late side effects such as sialadenitis and xerostomia; in the same randomized study it should of course be studied further whether these risks do not outweigh the benefits.

Conclusion

In the present retrospective study it has been observed that the radiation absorbed dose to the blood was a significant indicator of ablation success, whereas this was not the case for the administered 131I activity in a population of patients receiving >1,850 MBq of 131I for thyroid remnant ablation. Our findings should encourage further research into factors influencing ablation success other than the administered activity.