Radioactive iodine in thyroid medicine-How it started in Sweden and some of today's challenges

Abstract

In Sweden, radioactive iodine for thyroid diagnostics and therapy was introduced by Jan Waldenström (1906–1996) and Bengt Skanse (1918–1963). The paper describes the start of the clinical use of radioiodine, the various iodine isotopes available, measurement techniques and dosimetry. There are still problems to solve in relation to an optimal clinical use of radioiodine. One of the remaining challenges is to get consensus about the goal of the treatment of hyperthyreosis, as well as about a method for individual absorbed dose calculations. Careful dose estimates will prevent unnecessary radiation exposure and constitute a base for a future optimised radioiodine therapy. For the dose calculation, it is important to understand if there is any clinically significant temporary reduction in the ability of thyroid tissue to trap or retain ¹³¹I-iodide following prior administration of a diagnostic activity of ¹³¹I-iodide (stunning of the thyroid). This may be of special concern in connection with treatment of thyroid cancer and its metastases. Finally, the production capacity, availability and delivery of ¹²³I have to be improved to increase clinical access to this radionuclide, which is optimal for diagnostic imaging and which gives lower absorbed dose and therefore also less risk for thyroid stunning than ¹³¹I.

The knowledge that the thyroid has a selective affinity for iodide was established in dogs in 1915 [1]. The first studies of iodide metabolism in thyrotoxicosis were made in 1927 when stable iodide was administered orally and the iodine in the urine determined after different intervals. It was found that most patients with thyrotoxicosis excreted less iodine than normal subjects and this method (the iodine tolerance test) became frequently used before the production of radioactive iodide. In 1938, Hertz et al. [2] described the possibility of radioactive isotope technique into the study of thyroid physiology and the investigations concerning thyroid diseases and iodide metabolism were enormously increased. Soon after ¹³¹I became available from Oak Ridge in 1946 [3], the first treatments were carried out by Chapman et al. [4] and Hertz et al. [5].

The aim of this work is to give some views of the start of the clinical use of radioiodine in thyroid medicine from a Swedish perspective and to identify some current challenges.

How it started in Sweden

During the last World War, Sweden – like many other countries-was isolated from other countries. It became evident that our knowledge regarding the most recent medical development was quite fragmentary. This was true above all regarding the USA. Medical journals did not arrive and personal contacts were very few. Therefore persons in responsible position, among others the Surgeon General Axel Höjer and the Secretary of State Tage Erlander had the idea that scientists should be sent over to USA to collect facts. This was arranged and one of the scientists was Jan Waldenström, at that time professor of theoretical medicine in Uppsala. He had the possibility to visit various laboratories in USA during a seven months period in 1943 and 1944 [6–10] and visited places where medical investigations and education were at the highest level. Boston was at that time the medical capital of USA and Massachusetts General Hospital (MGH) had the highest reputation, both for surgery and for internal medicine. James Howard Means was head of the medicine department. His own main interest was in thyroid disease and he was lucky and also far-sighted in adopting radionuclides for diagnosis. The department of medicine at MGH had close collaboration with the physicists at Massachusetts Institute of Technology (MIT), where the leading person was Robley D Evans. Waldenström [10] characterises him as a physicist with “the most unusual instinct for clinical problems”. Early clinical investigations with radioiodine, initially using ¹²⁸I and later ¹³¹I were made by Evans, Hamilton, Soley, Hertz, Roberts, Means and others [3], [11–14].

Waldenström later arranged a visit at the same laboratory for his pupil and collaborator Bengt Skanse, who with the help of Waldenström got a Rockefeller fellowship, which gave him the possibility to stay in Boston for nearly two years during 1948 and 1949. There he introduced a method to determine the radioiodide uptake in the thyroid by analysing the 24-hour urinary excretion [15]. He also actively took part in other studies related to diagnosis and therapy of thyreotoxicosis and thyroid cancer [3], [11], [12]. After his return to Uppsala, Skanse finished his thesis “Radioactive iodine in the diagnosis of thyroid disease” [1].

In 1949 Waldenström became professor of medicine in Malmö and started there January 1, 1950. Some months later, Skanse joined him and began to introduce his new methods, both for diagnosis and for treatment of thyroid diseases. In parallel, there was an impressively fast development at other university hospitals in Sweden. At Radiumhemmet in Stockholm, Lars-Gunnar Larsson, Agnar Egmark, Lars Jonsson, Inger Ragnhult [16–18] and later Jerzy Einhorn [19] gave very significant contributions to the development of the diagnosis and therapy of thyreotoxicosis and thyroid cancer. For more details about the progress in Sweden, the reader is referred to the detailed historical review by Carlsson [20].

Radioiodine isotopes

There are 36 known isotopes of iodine, of which only one is stable. Some of them are presented in Table I. The first ones used clinically for diagnostic purposes were ¹²⁸I, ¹³⁰I, ¹³¹I and ¹³²I [3], [11–13]. They were produced either at cyclotrons (¹³⁰Te + D) or in nuclear reactors (¹³⁰Te + n) in Oak Ridge and Los Alamos, USA and since 1947 also in Harwell, England. Later ¹²⁵I was used for radioimmunoassay (RIA) and sometimes also for in vivo diagnostics as well as for therapy. Today the only isotope used for radiation therapy is ¹³¹I, which together with ¹²³I also may be used in diagnostic nuclear medicine. ¹²³I has excellent physical properties for an imaging agent. It decays by electron capture with a half-life of 13 hours and emits photons with energy of 159 keV. It also provides lower absorbed dose to the thyroid with the necessary activity than does ¹³¹I. With respect to the risk for stunning, it should therefore be an advantage to use ¹²³I instead of ¹³¹I for imaging and uptake test purposes before a therapy. The major disadvantages of ¹²³I are high cost, and problems with availability and delivery from time to time. Despite these restrictions, ¹²³I should be the iodine of choice for thyroid imaging [21].

Table I.  Some isotopes of iodine, their physical half-life and mode of decay.

Mass-number	Half-life	Radiation/decay	Mass-number	Half-life	Radiation/decay
118	14 min	β^+, EC,	128	25 min	EC, β^−, γ
119	19 min	β^+	129	1.7 10^7 a	β^−, γ
120	1.4 h	EC	130	12 h	β^−, γ
121	2.1 h	β^+	131	8.0 d	β^−, γ
122	3.6 min	β^+	132	2.3 h	β^−, γ
123	13 h	EC, γ	133	21 h	β^−, γ
124	4.2 d	β^+, EC, γ	134	53 min	β^−, γ
125	59 d	EC, γ	135	6.6 h	β^−, γ
126	13 d	β^−, β^+, EC, γ	136	83 s	β^−, γ
127	STABLE		137	24 s	β^−, γ

EC, stands for decay by electron capture, which is normally followed by emission of characteristic X-rays and Auger electrons.

Additionally, ¹²⁴I is used for PET (positron emission tomography) studies, and possibly also other iodine β⁺ emitters, e.g. ¹²²I and ¹²¹I, may be utilised for this purpose in the future.

Measurements of uptake, retention and distribution within the thyroid gland

The usual way to quantify thyroid uptake of ¹³¹I-iodide after i.v. or per-oral administration, has been to place a detector over the thyroid, where the activity very soon concentrates. At early times, the uptake measurements were done with a GM-tube. Skanse [15] instead followed the excretion in urine and in that way he could indirectly estimate the retention in the thyroid. This could be done with 1/200 of the activity needed for thyroid uptake measurements. This method was used up to the time when the more sensitive NaI(Tl) scintillation detectors became available and thyroid uptake measurements using a detector over the neck came into practice again. There was an impressive technical development in Sweden, including the construction of scintigraphs, at Radiumhemmet in Stockholm [22] as well as at the Sahlgrenska University Hospital in Göteborg (Erik Berne and Ulf Jonsson at the company NUKAB). Before that, Johansson and Skanse [23] already in 1952 constructed a small gammacamera with a parallel hole collimator, NaI(Tl) as intensifying screen and photographic film as an imaging detector.

Absorbed dose calculations

Calculations of the absorbed dose to the thyroid at treatment of hyperthyroidism

Radioiodine therapy of hyperthyroidism using ¹³¹I-iodide is still the main form of therapy in nuclear medicine and is performed all over the world. For estimating the absorbed dose to the thyroid, information is needed not only about biokinetic parameters such as uptake and biological half-life, but also about the mass and geometrical shape of the uptake in thyroid tissue as well as decay characteristics for the radionuclide. For ¹³¹I, the major fraction of the emitted energy, 67%, is transferred to photons, while beta particles and electrons carry the remaining fraction. However, for a thyroid with a mass of 20 g, which is the mass of the thyroid of the “reference man” [24], the major part of the absorbed dose is delivered by electrons and beta-particles. Photons contribute to approximately 5% of the total dose, while for a mass of 100 g, which can be considered as a very large thyroid, the photon contribution has approximately doubled, but is still not responsible for more than about 10% of the total absorbed dose.

To obtain an accurate individual absorbed dose estimate for treatment of thyreotoxicosis, the active thyroid mass/volume should be assessed. This is often done by scintigraphic methods. For this purpose ^99mTc-pertechnetate may be used, which has better imaging properties than ¹³¹I and is cheaper and more readily available than ¹²³I, which, as earlier mentioned, should be a still better alternative. If available, SPECT imaging with a rotating gamma-camera is the best way to determine the mass/volume of the thyroid. Otherwise, the volume estimation has to be based on the projected area of the thyroid in an anterior image. Assuming that the thyroid is composed of two ellipsoidal lobes each with axes in the ratio of 2:1:1 the volume (V) can be calculated to approximately:1 with a density of 1 g cm⁻³ this equals the mass of the thyroid in grams, if the total projected area a is given in cm². More specifically assigning axes in the ratio of c:1:1, the volume (mass) becomes:2 With the aid of dose factors for a uniform distribution of ¹³¹I in small spheres, which can be obtained from the Olinda internal dosimetry software (OLINDA/EXM uses the RADAR method of dose calculation [25] and the dose conversion factors as supplied on the RADAR web site (www.doseinfo-radar.com), the dose-rate in Gy/h from 1 MBq ¹³¹I can be estimated to:3 where m is the mass in gram as calculated above. This is an approximation for small spheres. It should deviate less than 5% from the real dose rate, and be good enough to be used also for the case when the activity is distributed within organs of other geometrical shapes, e.g. ellipsoids.

The total number of decays, the cumulated activity, can be calculated from the effective half-time, uptake and administered activity. Values for these parameters should be individually determined (see below). Putting it all together, using Equations 1 and 3 together with data on initial uptake and effective half-time, the absorbed dose D in Gy to the thyroid approximately becomes4 where T_½ is the individual effective half-life in hours, U is the extrapolated initial uptake in the thyroid and A₀ is the administered activity expressed in MBq. As an approximation, it may be assumed that the uptake after 24 hours can be used. Taking it the other way around, the activity, A₀, to be administered in order to achieve a desired absorbed dose is5 In the derivation of this equation a number of approximations has been done, which all are small compared to the usually rather large uncertainties in the individual determination of the thyroid mass, half-time and uptake of iodide. To improve the dosimetry, effort should be concentrated on the assessment of these parameters.

Absorbed dose estimations to organs other than the thyroid for radiation protection purposes

For the estimation of the absorbed dose to different organs and the effective dose, often a biokinetic model is utilised. Such models are designed to be representative for the general population for estimating the population averaged doses or the collective dose. Due to large individual variations, such models can not be used for individual dose estimations, neither to the thyroid, nor to other organs. In connection with therapy, the absorbed dose to the thyroid has to be estimated based on individually assessed parameters, as discussed above.

Biokinetic models for iodine have been published by the ICRP to be used both for occupational and environmental exposure [26] and for nuclear medicine investigations [27]. Other authors have also published models designed for different purposes. For dose estimation in nuclear medicine, the MIRD model by Berman et al. [28] is an important contribution that has been widely used. The ICRP [29] has also published a model for iodine in the pregnant woman and the foetus.

In most models intended for nuclear medicine patients the uptakes in stomach wall and salivary glands are included. This is often not the case for the models intended for dosimetry of occupationally or environmentally exposed persons. On the other hand, models for patient dosimetry may be simplified in such a way that they are unsuitable for dosimetry of the long-lived isotope ¹²⁹I. Due to this discrepancy a more general compartment model, which can be used for different purposes, was recently developed [30], [31]. A remaining problem is that it is not clear to what degree the activity found in the stomach is located in the wall or in the content. Knowledge of this distribution is critical for an accurate dose calculation to the stomach as well as for assessment of the effective dose, especially in cases where the thyroid is blocked.

Today's challenges

For treatment of thyreotoxicosis, there is a need for a national (or preferentially international) consensus on an optimal protocol. The clinicians should jointly agree on parameters to characterise the result of a treatment. As it is today, some patients are treated to obtain a normal thyroid function. Others are treated to obtain a function somewhat under the normal and some patients are made totally hypothyroid. The situation gets more complicated by the fact that the thyroid function gradually decreases with time after treatment and that hypofunction easily can be treated with thyroid hormone substitution.

Even without consensus regarding the desirable effect of the treatment, it is important both for the individual patient and for the understanding of the radioiodine therapy and its future optimisation that the absorbed dose to the thyroid is quantified. The therapeutic effect of the therapy depends on the absorbed dose to the thyroid, just like in external radiation therapy. Comparing clinical observations without knowing the absorbed dose is not meaningful. If the outcome of a group of patient is to be studied the result is futile unless the absorbed dose to the thyroid is known for each individual patient.

For the individual patient it is also important not to be exposed to unnecessary radiation, but still get the absorbed dose needed to the thyroid to avoid a second treatment.

Different protocols to determine the activity of ¹³¹I to be administered for treatment of thyreotoxicosis are currently in use in Sweden, but not all of them consider the absorbed dose to the thyroid. The three major parameters needed for an absorbed dose calculation are, as discussed above; the volume of the functioning part of the thyroid, the uptake of ¹³¹I-iodide in the thyroid and the biological half-time. All three parameters are individual and vary between patients, and they are all needed for an individual dose planning.

Helene Jönsson [32] has in her thesis used detailed biokinetic data from 187 radioiodine therapies to compare the outcome with regard to absorbed dose to the thyroid assuming that different types of current protocols had been used for these specific patients. These protocols may have been constructed with very different intentions with regard to effect on the thyroid (to get normal thyroid function or to get totally hypothyroid patients or slightly hypothyroid patients).

In the 23 Swedish hospitals, performing this type of treatment, 17 different protocols are in use (2004) and only nine hospitals calculate the mean absorbed dose to the thyroid [33] by using the volume of the thyroid, the individual uptake of iodide in the thyroid and the individual biologic half-time. Nine hospitals use the volume of the thyroid and the individual uptake of iodide, but use a fixed biologic half-time. Five hospitals do not use the volume or the individual biokinetic data at all for the determination of the administered activity to the patient.

To get a national consensus on an optimal protocol in radioiodine therapy, hospitals have to have the courage to break old traditions and change their protocols into a national one.

A number of hospitals use protocols administering a fixed activity of 370 MBq ¹³¹I-iodide to all patients, not using any individual biokinetic data at all. Using this protocol, the hospital will as a mean administer 2.5 times the necessary activity and in individual patients from a slight underdosage (70% of the necessary activity) to up to 8 times (!) the activity needed [34]. Some hospitals administer 3.7 MBq per gram thyroid, but do not use the individual biologic half-time or uptake. This results in a variation of administered activity to the patient, with a range from half the needed activity to 1.6 times that activity [34].

When a protocol is used where the mean absorbed dose to the thyroid is not calculated for the individual patient, more activity than necessary is usually administered. This will unnecessarily expose the patient, the personnel and the patient's family. The excess patient exposure is of increasing concern since today younger patients are treated with ¹³¹I, which might increase the risk of cancer during the rest of the patent's life. Also, since the absorbed dose to the thyroid is unknown, there is no parameter to correlate the therapeutic effect with.

To determine the exact absorbed dose to the thyroid, volume and biokinetics have to be taken into account. Concerning the uptake and half-time measurements, one can argue that the traditional repeated measurements (at 2, 4, 24, 48 h, and even more) are too time consuming. A simplified patient-specific protocol for treating hyperthyroidism has however been proposed [35]. It is based on single uptake measurement 4–7 days after the intake of a test activity of ¹³¹I. This simplified protocol is as patient-convenient and time-effective as a protocol using a fixed administered activity. The methods for volume determination should be improved in general, since this parameter involves the largest uncertainties. Using recent SPECT/CT technology and ¹²³I, there are now improved methods to determine the functioning volume of the thyroid in a much better way than before. Volume estimation could be combined with a single uptake test – 4–7 days after the administration-using a small amount of ¹³¹I added to the ¹²³I.

The estimation of the absorbed dose to thyroid cancer remnants and metastases is still more complicated. One reason is the difficulties to define the distribution volume of the ¹³¹I-uptake and its kinetics. Therefore fixed amounts of ¹³¹I-iodide are still normally used for the treatment of thyroid cancer. To be able to find possible correlations to the registered effects, the absorbed dose to thyroid cancer remnants and metastases should be calculated.

Thyroid stunning means a temporary reduction in the ability of normal thyroid tissue or differentiated thyroid cancer to trap or retain a therapeutic activity of ¹³¹I-iodide following a prior administration of a diagnostic activity of ¹³¹I-iodide. The first observations related to thyroid stunning date back decades ago and several authors have studied the phenomenon, mainly in connection with treatment of thyroid cancer, where comparatively high test activities are used. In in vitro experiment using porcine thyroid follicle cells, Postgård et al. [36] have shown that stunning is a real phenomenon, meaning a radiation induced and dose dependent reduction of the iodine transport capacity of viable thyroid cells. The data imply a potential reduction of the therapeutic efficacy of radioiodine treatment after diagnostic use of ¹³¹I, even at low absorbed doses. Therefore, it is important to further investigate the impact of thyroid stunning with respect to the absorbed dose for the various clinical situations, especially in the treatment and follow up of thyroid cancer [37].

In a recent review article, Medvedic [38] concludes that thyroid stunning is evident in patients with well-differentiated thyroid cancer if thyroid remnants are irradiated by a few gray. The absorbed dose from the diagnostic administration of ¹³¹I prior to radioiodine therapy should be limited to 4 Gy. Sabri et al. [39] treated benign thyroid disease with two-step ¹³¹I administration and proved that stunning really exists at the second one and that it depends only on the absorbed dose at the first ¹³¹I application. The extent of stunning would be less than 1% (0.6%) for an absorbed dose of 35 Gy after the first administration, and thus negligible for practical purposes. These two examples, which show very different results, illustrate the need for further studies of the phenomenon. Possible differences in thyroid uptake of ¹³¹I-iodide during the diagnostic test and the therapeutic situation could be detected by measuring the urinary excretion during the two procedures. In the therapy situation, it is easier to analyse the urine content of ¹³¹I than to measure the high activity in the thyroid. There is obviously a need for more basic and clinically oriented studies regarding the stunning phenomenon, mainly in connection with treatment of thyroid cancer, but also in relation to treatment of thyreotoxicosis.

Since ¹²³I gives lower absorbed doses than ¹³¹I, there is a growing body of evidence that this radionuclide is to be preferred for imaging, at least in connection with whole body screening in diagnosis of distant metastases from thyroid cancer.

This paper is based on a presentation held at the symposium on Half a Century with Radioactive Iodine in Thyroid Medicine – Historical and Future Aspects in Umeå, Sweden, September 29–30, 2005, organised by one of the pioneers in thyroid medicine, professor em Lars-Gunnar Larsson and associate professor Torgny Rasmuson. The authors are grateful to Bengt Hemdal for comments to the manuscript.