A method to measure the absorbed dose of the thyroid during I-131 therapy, using a collar detector system and a SPECT acquisition

A method to measure the absorbed dose of the thyroid during I-131 therapy, using a collar detector system and a SPECT acquisition Koen van Gils and Peter Brinks Department of Medical Physics, St Antonius Hospital, Utrecht/Nieuwegein, The Netherlands Jules Lavalaye Department of Nuclear Medicine, St Antonius Hospital, Utrecht/Nieuwegein, The Netherlands Hein J. Verberne Department of Nuclear Medicine, Academic Medical Center, Amsterdam, The Netherlands Jan B. A. Habraken a) Department of Medical Physics, St Antonius Hospital, Utrecht/Nieuwegein, The Netherlands (Received 1 May 2017; revised 7 July 2017; accepted for publication 10 July 2017; published 20 August 2017) Purpose: Due to variations in biological half-life, accurate thyroid dosimetry for I-131 therapy is not trivial in clinical practice. In recent publications, systems are described to measure the uptake of I-131 in the thyroid repeatedly over time. In this work, we present a method to calculate patient specific pharmacokinetics and absorbed dose using such a collar detector system (CoTI) in combination with a SPECT acquisition and a two-compartment model fit. Methods: For three patients receiving I-131 therapy for benign thyroid conditions, the complete uptake profile is measured over a period of 15 to 25 days after administration. A SPECT measurement is performed to assess the functional volume of the thyroid and the amount of I-131 in the thyroid. The uptake profile measured in counts-per-second is converted to absolute activity in MBq using the absolute quantification of the SPECT. A two-compartment model is used as a fit to the uptake data of the thyroid and to estimate the activity in the blood-pool. The estimated absorbed dose to the thyroid is then calculated from the integral of the activity. The assessed parameters from the method (6- and 24-h uptake, thyroid volume and I-131 uptake concentration) are compared with the values as determined in clinical practice. Furthermore, the convergence of the calculated absorbed dose as a function of measurement series duration is determined to assess the required measurement duration of the uptake profile. Results: The two-compartment model fit shows a good agreement with the measured data points. Resulting dynamic uptake profiles of the three patients differ from each other. The uptake percentages differ from the pretherapy I-123 uptake measurements that are used in usual clinical practice, which shows the potential added value of the proposed method. The duration of the required measurement series appears to be patient dependent and therefore needs to be determined for each patient individually. The proposed method allows for a basic investigation of the individual dynamic uptake profile of I-131 in the thyroid and the calculation of the absorbed dose. Conclusions: The proposed measurement method is feasible and easily implementable given a system that can measure the uptake of I-131 in the thyroid repeatedly over time. The observed differences in dynamic uptake profiles and the differences in the absorbed thyroid dose as calculated with our method and the parameters of the usual clinical care support the relevance of the proposed method. In future studies, this approach may possibly be used for outcome prediction and therapeutic activity optimization. 2017 American Association of Physicists in Medicine [https://doi.org/ 10.1002/mp.12472] Key words: pharmacokinetics, radionuclide therapy, thyroid dosimetry, two-compartment model 1. INTRODUCTION Iodine-131 (I-131) is used extensively to diagnose and treat thyroid diseases such as Graves disease, euthyroid goiter, multinodular goiter, and thyroid carcinoma. Due to the selective uptake of I-131 in the thyroid, localized treatment is possible with a typically low absorbed dose to surrounding tissues and organs. Traditionally, the uptake of iodine in the thyroid is measured prior to treatment using either Iodine-123 (I-123) or I-131 to calculate an appropriate therapeutic activity. These calculations typically incorporate thyroid volume, thyroid uptake and a disease-specific constant. The Dutch Society of Nuclear Medicine (NVNG) recommends the following calculation: 1 D ¼ K V U ; (1) 5450 Med. Phys. 44 (10), October 2017 0094-2405/2017/44(10)/5450/7 2017 American Association of Physicists in Medicine 5450

5451 van Gils et al.: Absorbed thyroid dose from uptake profile 5451 where D is the activity to be administered in MBq, V is the thyroid volume in ml and U is the fractional thyroid uptake after 24 h (thyroid-activity 24 h /administered-activity 24 h ). The desired I-131 concentration in the thyroid K = 4MBq/ ml for Graves disease and euthyroid goiter, and K = 4or K = 8 MBq/ml for multinodular goiter depending on clinical assessment. While most treatment guidelines are determined empirically, recommendations of disease-specific absorbed dose requirements (in Gy) for therapeutical effect do exist. 2 For instance, the dose to achieve complete ablation, in patients with Graves disease, is in the range 200 300 Gy. 2 Even though the calculated activity to be administered is somewhat tailored to the individual patient, large variations in biological half-life of iodine exist. 3 These variations directly translate into differences in absorbed dose. Because determining the effective half-life is impractical this is typically not measured in clinical practice but roughly estimated empirically based on uptake measurements at various points in time after administration of I-123 or I-131, typically after 6 and 24 h. In this work a collar detector system (CoTI, Collar Therapy Indicator, AG Medical, Saint-Aubin, France) is used together with a SPECT acquisition in a proof of concept methodology to determine the uptake profile and the absorbed dose of the thyroid. The aim of this study is to present a methodology that can provide insight into the individual dynamic uptake profile and can more accurately calculate the total absorbed dose in the thyroid, using a simple collar measuring device and a SPECT acquisition. 2. MATERIALS AND METHODS 2.A. Patient inclusion criteria Patients undergoing I-131 treatment (administered orally via a capsule) for euthyroid goiter or hyperthyroid multinodular goiter were eligible for study participation. In addition, patients had to be admitted to the hospital for radiation safety purposes (i.e., Dutch legislation: ambient dose equivalent rate of >20 lsv per hour at 1 m distance after administration) for typically two consecutive days. The latter allows for adequate familiarization of the patient with the measurement method and device. Three subjects were included, one patient with a euthyroid goiter and two patients with a hyperthyroid multinodular goiter. All subjects provided written informed consent before study participation. The study was approved by the local institutional review board and conducted according to the principles of the International Conference on Harmonization Good Clinical Practice. 2.B. I-123 uptake measurement The usual standard clinical practice measurement for thyroid uptake in our clinic is performed with I-123 using a Siemens E.CAM equipped with a pinhole collimator. In this examination, the uptake is measured 6 and 24 h after administration of the I-123 by comparing the total thyroid counts in an anterior view to a reference activity in a polymethyl-methacrylate (PMMA) neck phantom (both at a fixed distance and corrected for background counts). In addition to the anterior view also right- and left anterior oblique views are recorded. The resulting calculated uptake percentage and the visual volume estimation based on the acquisitions are used by the nuclear medicine physician to calculate an appropriate I-131 therapeutic activity using Eq. (1). 2.C. I-131 uptake profile measurement The used measurement device (CoTI) consists of a disposable soft collar in which two separate detector modules (50 9 30 9 18 mm 3 ) are placed, one for each thyroid lobe. Each module is connected to the data acquisition unit and contains a scintillation detector (3 mm 9 3 mm active detection surface) faced to the body surface. The use of a comparable device was recently explored by F urstner et al. 4 Example photos of the CoTI are shown in Fig. 1. Energy windowing is not supported. The device is described in detail in Ref. [5] Acquisitions are initiated by the patient by pressing a large button on a second device, which is wirelessly connected to both the acquisition unit and the Internet. The latter option allows for remote monitoring of the measurements. Only measured count rates are transferred, no patient data are transmitted. In each measurement the average count rate over an interval of 20 s is recorded for both detectors. The first hour after administration measurements are performed every 15 min, starting with the first measurement immediately after administration. For the remainder of the first day after administration, a measurement is performed every hour while the second and third day (if applicable) a measurement is performed every 2 h. No measurements are performed during sleep. After discharge the patient is asked to perform measurements at home twice a day (morning and evening) for a duration of up to 3 weeks. Before the start of each measurement the patient places the collar around the neck for the duration of the measurement. In between measurements the collar is removed. 2.D. SPECT acquisition Shortly before the patient is discharged (approximately 2 3 days after I-131 administration), a SPECT acquisition is performed using a Siemens Symbia T2 SPECT/CT equipped with a medium energy collimator. While typically a high energy collimator is used for I-131 acquisitions, this is not available in our clinic. However, SPECT reconstructions of I-131 using medium energy collimators have shown good results. 6 A 15% energy window around the central photopeak and equal width upper- and lower scatter windows are acquired. A total of 64 views were acquired (32 per head) with a 128 9 128 matrix size, no zoom is applied resulting in a 4.8 mm 9 4.8 mm pixel size. No CT acquisition is performed.

5452 van Gils et al.: Absorbed thyroid dose from uptake profile 5452 FIG. 1. (a) CoTI detector module, the star-mark denotes the position of the detector. (b) The CoTI collar with the two modules placed in it. The position of the detectors are indicated by the two circles. (c) The CoTI data acquisition unit. (d) Usage of the CoTI, at the back the two connecting cables from the modules to the acquisition unit are visible. [Color figure can be viewed at wileyonlinelibrary.com] Images are reconstructed using Siemens Flash 3D OSEM algorithm (six iterations, eight subsets) with triple energy window scatter correction and a 9 mm Gaussian postfilter. The reconstructions are corrected for detector dead time, by using the SPECT-CT s dead time lookup tables. A vial containing water with a known activity of I-131 inside the PMMA neck phantom is used to quantitatively calibrate the SPECT reconstructions. The same calibration is used for all three patients. A 40% of maximum pixel value isocontour segmentation showed an accurate delineation of the activity volume as compared to the CT image of the same vial in the phantom and is therefore used for segmentation in the patient studies. In Fig. 2 example images of the acquisition, reconstruction end segmentation are shown. 2.E. Data analysis The measured uptake profile count rates of both detectors are corrected for detector dead time using a measured (I- 131) dead time lookup table (data not shown). The corrected count rates of both CoTI detectors are averaged, yielding one count rate per measurement. The two compartment model solution as described in detail in Ref. [7] is fitted to the data set using least squares. A schematic of this model is shown in Fig. 3. This solution of the thyroid activity as a function of time is: A a k t A T ðþ¼ t e ð k pþk h Þt e ð k rþk pþ k t Þt : (2) k r þ k t k h FIG. 2. Example SPECT images. (a) raw anterior projection image, (b) SPECT reconstructed coronal slice, (c) SPECT reconstructed coronal slice with segmented activity volume (indicated by dots). [Color figure can be viewed at wileyonlinelibrary.com]

5453 van Gils et al.: Absorbed thyroid dose from uptake profile 5453 thyroid volume and E is the mean energy deposition per decay, calculated using the methods described in Ref. [7]by E ¼ ðv thyroidþ 0:25 þ 18 : (4) 7:2 To assess the convergence of the fit in order to determine a suitable point of termination of measurements, the calculated thyroid absorbed dose as a function of measurement duration (and thus the number of measurements) is calculated. FIG. 3. Schematic representation of the used two-compartment model. A a is the administered I-131 activity in MBq and t is the time after administration in days. Thyroid uptake (k t ), renal clearance (k r ), physical decay (k p ), and hormone excretion (k h ) are rates with the unit day 1. The total activity in the thyroid (at the time of acquisition) and the uptake volume are calculated from the SPECT acquisition using the 40% of maximum pixel value isocontour segmentation. The SPECT acquired total activity in the thyroid (in MBq) and the two compartment model fit count rate at the time of acquisition are used to convert all measured profile count rates to total activity measurements in the thyroid. This assumes that the relative distribution of I-131 in the thyroid is constant over time. The I-131 concentration at 6 and 24 h after administration is calculated by dividing the total I-131- activity from the curve at 6 and 24 h by the thyroid volume as measured by the SPECT. The concentrations are then converted into uptake percentages and compared to the gamma camera I-123 uptake measurements of the same patient. Assuming that the total administered activity instantaneously enters the blood pool compartment, the total blood pool activity over time (A pool ) is calculated. Note that this is an approximation, as the underlying pharmacokinetics are considerably more complex. 8 The area under the curve of the resulting fit A T is equal to the total number of disintegrations of I-131 in the thyroid. Combining the total number of disintegrations with the segmented thyroid volume, the thyroid absorbed dose is calculated by D thyroid ¼ n E (3) V thyroid where D thyroid is the thyroid absorbed dose, n is the total number of disintegrations in the thyroid, V thyroid is the segmented 3. RESULTS The relevant details of the included patients are summarized in Table I. Figure 4 shows the calculated total activity in the thyroid over time for all three patients for both the first 48 h after administration (a) and up to approximately 25 days after administration (b). The area under the curve is shaded for each data set. In all three measurement series a good fit of the two compartment model is observed. The estimation of the blood pool activity (A pool ) is shown by dashed lines for the first 48 h. Convergence of the calculated dose derived from the two compartment model as a function of measurement duration is shown in Fig. 5. The calculated thyroid absorbed dose is plotted as a function of measurement duration after administration. All three measurement series show a tendency to converge to a stable calculated absorbed dose but the rate of convergence is not equal. Table II shows the administered activity and 6-h and 24-h uptake for both the usual clinical practice gamma camera measurement and the uptake profile measurement fit. Additionally, the ratio between the uptake as calculated by the uptake-profile method and the planar I-123 calculated uptake is shown. The planar I-123 volume estimation, the desired I-131 thyroid concentration at 24 h, the SPECT segmented volume, the measured I-131 thyroid concentration at 24 h and the absorbed dose are shown in Table III. A significant discrepancy between the planar estimated- and SPECT segmented volume is seen. As the fit parameters of the two compartment model are the pharmacokinetic rates of Eq. (2), it is possible to evaluate the thyroid uptake rate (k t ), renal clearance rate (k r ) and thyroid hormone excretion rate (k h ). These values are summarized in Table IV, together with the biological- and effective half-life. Note that the thyroid excretion rate and the TABLE I. Relevant details of included patients. Pat # Gender Age Condition Time between I-123 and I-131 (days) Duration of admission (days) Remarks 1 F 76 yr Hyperthyroid multinodular goiter 8 3 2 F 76 yr Euthyroid goiter 36 2 Previous hemithyroidectomy (right lobe) 3 F 68 yr Hyperthyroid multinodular goiter 42 3

5454 van Gils et al.: Absorbed thyroid dose from uptake profile 5454 FIG. 4. Total activity in the thyroid as a function of time after administration for all three patients. The two compartment model fit (A T, solid lines) is plotted. Area under the curve as defined by two compartment model fit is shaded. In (a) the thyroid activity (A T ) and the blood pool activity estimate (A pool ) during the first 48 h after administration are shown, in (b) the thyroid activity (A T ) up to approximately 25 days after administration is shown solely. The blood pool curves are not plotted in (b) since they would almost merge with the y-axis. [Color figure can be viewed at wileyonlinelibrary.com] TABLE II. Uptake parameters as assessed in usual clinical care and from the uptake-profile method. The ratio between the two calculated uptake parameters is shown in the last two columns. Pat # Administered activity (MBq) Uptake usual clinical care (I-123) Uptake uptake profile (I-131) Ratio I- 131/I-123 6 h 24 h 6 h 24 h 6 h 24 h 1 790 20.5% 34.7% 33.7% 44.6% 1,6 1,3 2 797 6.8% 17.6% 13.5% 16.1% 2,0 0,9 3 1989 13.1% 22.3% 36.9% 48.8% 2,8 2,2 FIG. 5. Convergence of thyroid absorbed dose as a function of measurement duration after administration. The dashed lines indicate a 10% interval around the last data point. [Color figure can be viewed at wileyonlinelibrary.- com] biological half-life contain the same information, as biological half-life is equal to ln(2)/k h. The effective half-life (T eff )is a combination of physical half-life (T ph ) and biological halflife (T b ): 1/T eff = 1/T ph + 1/T b. For patient 3, the observed effective half-life is equal to the physical half-life, as the biological half-life is not measurable. 4. DISCUSSION A method is presented to calculate the absorbed dose of the thyroid during I-131 therapy using a collar detector system, a single SPECT acquisition and a two-compartment model. The proposed method is feasible and easily implementable. The resulting dynamic uptake profiles of the three TABLE III. Thyroid volume as determined by nuclear medicine physician from planar I-123 acquisition, desired thyroid concentration after 24 h [K, Eq. (1)], SPECT segmented thyroid volume, actual thyroid concentration after 24 h and calculated thyroid absorbed dose. Pat # Usual clinical care Planar I-123 estimated thyroid volume (ml) Desired K (MBq/ml) SPECT thyroid volume (ml) Proposed method Actual K (MBq/ml) Absorbed dose (Gy) 1 65 4 94 3.7 100 2 30 4 46 2.8 74 3 50 8 167 5.8 199 patients differ from each other and the calculated uptake percentages differ from the pre-therapy I-123 uptake measurements, which shows the potential added value of the proposed method. Repositioning of the CoTI is expected to cause variation in measured count rates. 9 However, as observed in Fig. 4, a good fit was found in all three patients. Therefore, the impact

5455 van Gils et al.: Absorbed thyroid dose from uptake profile 5455 TABLE IV. Calculated pharmacokinetic rates with standard deviations, and biological- and effective half-life. The reported standard deviation is the standard deviation of the rates as a fit parameter in the two-compartment model. Pat # k t SD (day 1 ) k r SD (day 1 ) k h SD (day 1 ) Biological half-life (days) Effective half-life (days) 1 2.5 0.2 2.9 0.3 0.023 0.006 29.7 6.3 2 1.1 0.2 5.9 1.1 0.022 0.01 31.9 6.4 3 2.7 0.2 2.9 0.2 NM NM 8.0 NM, not measurable. of repositioning seems to be limited in clinical practice. As the I-131 capsule first has to dissolve before the iodine starts to enter the blood pool via the gastrointestinal (GI) tract, the measured count rates within 30 min after administration are not likely due to activity in the thyroid. Rather, these are probably the result of scattered or primary photons emitted from I-131 in the GI tract and blood pool. After approximately 24 h however, only a minimal residual activity in the GI tract and no blood pool activity is expected due to renal clearance and thyroid trapping, thus the subsequent measurements accurately reflect the true activity in the thyroid. When used in clinical practice, defining a cut-off point after which the measurement series can be stopped is essential. Figure 5 shows that all three measurement series converge to a stable calculated absorbed dose, but the rate of convergence is different and a general cut-off point is not likely to be found. One possible reason why convergence is reached rapidly for patient 3 is that for this patient there is no measurable biological half-life and effective half-life is determined by physical decay only. A cause for slower convergence might be a variation in positioning accuracy of the collar resulting in a wider spread of measurement values, as might have been the case in the first measuring days of patient 2. In the next version of the CoTI system adaptations will be made to secure a more uniform positioning of the detectors. The measured 6- and 24 h post administration uptake percentages from the uptake profile (Table II) differ from the gamma camera determined uptake values used in clinical care. There are several potential reasons for the found differences. Firstly, the differences may be caused by measurement inaccuracies present in both methods. It has to be noted that the usual clinical care gamma camera determined uptake value consists of a single measurement in which a variation is transmitted directly to an error in the result. In our proposed method the uptake profile is assessed by multiple measurements and a two-compartment model is used for fitting the curve. Variations in individual measurements should therefore minimally influence the result. Variations in the SPECT calibration measurement will directly influence the result, however, the 3D SPECT quantification is more reliable and causes less variation than the planar uptake measurement with pinhole collimator, which is more sensitive to geometric variations. Secondly, it has to be noted that in our clinic I-123 is used for the uptake measurement instead of I-131. It is possible that this might lead to differences. Thirdly, the patients may have changed their dietary intake of iodine between uptake measurement and therapy and thereby altered the uptake for the uptake profile measurement during therapy. For the last two reasons it might be interesting to use our proposed method in a clinical setting where I-131 is used for the uptake measurement, to assess the uptake profile during the uptake study as well as during therapy. Truly individualized therapeutic activities can be calculated by measuring the uptake curve prior to administering the therapy. Finally, by measuring the uptake profile before and after administering the therapeutic activity, possible thyroid stunning effects can be investigated. Due to differences in calculated I-123 uptake and I-131 uptake and volume assessment, the measured thyroid concentration after 24 h differs from the desired thyroid concentration after 24 h (Tables II and III). Interestingly, in our three patients large differences between the usual clinical care measurement and the measurement of our proposed method of the uptake and volume, led to relatively small differences in the measured thyroid concentration. For example patient 3 shows an underestimation of the volume of 234% and an underestimation of the 24 h-uptake of 218%, leading to a relatively reasonable deviation from the desired thyroid concentration of I-131 at 24 h of 28%. It is uncertain whether or not it is representative that the errors, as assessed in clinical care, in uptake and volume measurement are balanced in clinical outcome. To address this, a larger study is needed in which a larger group of patients is included and our proposed uptake-profile method is used to assess thyroid uptake, volume, I-131 concentration and absorbed dose. While the described uptake-profile method allows for determination of pharmacokinetic rates in a two compartment model (Table IV), additional measurements such as GI tract activity or blood- or urine sampling are necessary for a more detailed investigation of the underlying pharmacokinetics using more realistic models. For dosimetry purposes, however, the two compartment model describes the data more accurately than the usual clinical care method and can therefore be used to calculate the absorbed dose more accurately. Table IV also shows that the effective half-life of the two patients with multinodular goiter is different, underlining the importance of patient-specific measurement duration for determining the uptake profile. 5. CONCLUSION The proposed uptake-profile measurement method is feasible and easily implementable and only requires a modular and portable nonimaging system that can measure the uptake of I-131 of the thyroid repeatedly and a single SPECT measurement. The absorbed dose of the thyroid is calculated in a fairly straightforward manner and the observed differences in

5456 van Gils et al.: Absorbed thyroid dose from uptake profile 5456 the individual dynamic uptake profiles of I-131 in the thyroid support the relevance of the proposed method. In future studies this approach may be used for outcome prediction and therapeutic activity optimization. ACKNOWLEDGMENTS The authors acknowledge AG Medical for supplying the CoTI device for this study. CONFLICT OF INTEREST No conflict of interest or financial disclosures. ETHICAL APPROVAL Informed consent was obtained from all individual participants in this study. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. a) Author to whom correspondence should be addressed. Electronic mail: j.habraken@antoniusziekenhuis.nl; Telephone: +31 88 320 93 31. REFERENCES 1. Barneveld PC, van Urk P. Aanbevelingen Nucleaire Geneeskunde. Commissie kwaliteitsbevordering van de Nederlandse Vereniging voor Nucleaire Geneeskunde, Delft; 2007. 2. Stokkel MPM, Junak DH, Lassmann M, Dietlein M, Luster M. EANM procedure guidelines for therapy of benign thyroid disease. Eur J Nucl Med Mol Imaging. 2010;37:2218 2228. 3. Kramer GH, Hauck BM, Chamberlain MJ. Biological half-life of iodine in adults with intact thyroid function and in athyreotic persons. Radiat Prot Dosimetry. 2002;102:129 135. 4. F urstner M, Hentschel M, Spanjol PM, et al. Technical note: determination of individual thyroid clearance effective half-life with a common handheld electronic dosimeter. Med Phys. 2017;44:1558 1562. 5. Brinks P, van Gils K, Kranenborg E, Lavalaye J, Dickerscheid DBM, Habraken JBA. Measuring the actual I-131 thyroid uptake curve with a collar detector system: a feasibility study. Eur J Nucl Med Mol Imaging. 2017;44:935 940. 6. Kobayashi M, Wakabayashi H, Kayano D, et al. Application of a medium-energy collimator for I-131 imaging after ablation treatment of differentiated thyroid cancer. Ann Nucl Med. 2014;28:551 558. 7. H anscheid H, Canzi C, Eschner W, et al. EANM Dosimetry Committee series on standard operational procedures for pre-therapeutic dosimetry II. Dosimetry prior to radioiodine therapy of benign thyroid diseases. Eur J Nucl Med Mol Imaging. 2013;40:1126 1134. 8. Johansson L, Leide-Svegborn S, Norrgren K, Mattsson S. Biokinetics of iodide in man: refinement of current ICRP dosimetry models. Cancer Biother Radiopharm. 2003;18:445 450. 9. Kranenborg EJ, van der Boor R, Gelderblom AE, van Gils KAJ, Habraken JBA. Feasibility of thyroid uptake measurements with the CoTI device. Eur J Nucl Med Mol Imaging. 2016;43:S162.