Radiation dose estimation using preclinical imaging with ¹²⁴I-metaiodobenzylguanidine (MIBG) PET

Abstract

Purpose: A pretherapy ¹²⁴I-metaiodobenzylguanidine (MIBG) positron emission tomography (PET)∕computed tomography (CT) provides a potential method to estimate radiation dose to normal organs, as well as tumors prior to ¹³¹I-MIBG treatment of neuroblastoma or pheochromocytoma. The aim of this work was to estimate human-equivalent internal radiation dose of ¹²⁴I-MIBG using PET∕CT data in a murine xenograft model.

Methods: Athymic mice subcutaneously implanted with NB1691 cells that express high levels of human norepinephrine transporter (n=4) were imaged using small animal microPET∕CT over 96 h (approximate imaging time points: 0.5, 2, 24, 52, and 96 h) after intravenous administration of 3.07–4.84 MBq of ¹²⁴I-MIBG via tail vein. The tumors did not accumulate ¹²⁴I-MIBG to a detectable level. All four animals were considered as control and organ radiation dosimetry was performed. Volumes of interest were drawn on the coregistered CT images for thyroid, heart, lung, liver, kidney, and bladder, and transferred to PET images to obtain pharmacokinetic data. Based on tabulated organ mass distributions for both mice and adult male human, preclinical pharmacokinetic data were extrapolated to their human-equivalent values. Radiation dose estimations for different age groups were performed using the OLINDA|EXM software with modified tissue weighting factors in the recent International Commission on Radiological Protection (ICRP) Publication 103.

Results: The mean effective dose from ¹²⁴I-MIBG using weighting factors from ICRP 103 to the adult male was estimated at 0.25 mSv∕MBq. In different age groups, effective doses using values from ICRP 103 were estimated as follows: Adult female: 0.34, 15-yr-old: 0.39 mSv∕MBq, 10-yr-old: 0.58 mSv∕MBq, 5-yr-old: 1.03 mSv∕MBq, 1-yr-old: 1.92 mSv∕MBq, and newborn: 3.75 mSv∕MBq. For comparison, the reported effective dose equivalent of ¹²⁴I-NaI for adult male (25% thyroid uptake, MIRD Dose Estimate Report No. 5) was 6.5 mSv∕MBq.

Conclusions: The authors estimated human-equivalent internal radiation dose of ¹²⁴I-MIBG using preclinical imaging data. As a reference, the effective dose estimation showed that ¹²⁴I-MIBG would deliver less radiation dose than ¹²⁴I-NaI, a radiotracer already being used in patients with thyroid cancer.

INTRODUCTION

Iodine-124 is an emerging radionuclide for imaging because iodination techniques are common in developing new tracers for noninvasive nuclear molecular imaging.^{1, 2} Unlike ¹²³I which is an excellent radioiodine for imaging using a single photon emission computed tomography, ¹²⁴I can be imaged using positron emission tomography (PET) for more accurate quantification of tracer distribution. In addition, when combined with ¹³¹I therapy such as ¹³¹I-NaI for thyroid cancer or ¹³¹I-metaiodobenzylguanidine (MIBG) for neuroblastoma or pheochromocytoma, ¹²⁴I PET imaging currently is the best modality for pretherapy dosimetry because of its favorable characteristics such as whole-body tomographic capability from PET and a similar half-life (4.2 d) to ¹³¹I (8.02 d). For pretherapy dosimetry of ¹³¹I treatment, ¹³¹I can be also used as an imaging agent; however, as an imaging agent, ¹³¹I is not an ideal radiolabel because the gamma emission from ¹³¹I (photopeak of 364 keV) is not efficiently collected by a typical gamma camera with high-energy collimators and 3∕8 in. NaI(Tl) crystal.³ Roughly 50% or more of photons from ¹³¹I escape the scintillation crystal and the collimators, even if they are designed for high-energy radionuclides such as ¹³¹I, causing high fractions of collimator septa penetration and scatters that greatly degrade image quality of ¹³¹I scans.^{3, 4}

For these reasons, ¹²⁴I-NaI imaging and dosimetry for ¹³¹I-NaI treatment of thyroid cancer has been studied in recent years.^{5, 6, 7, 8, 9, 10, 11, 12, 13} There are also other ¹²⁴I-labeled radiotracers that have been studied for other types of diseases.^{14, 15, 16, 17, 18} However, ¹²⁴I-MIBG has not been systematically studied in human subjects. Although there was a report of using ¹²⁴I-MIBG in a limited number of human adult subjects (n=2) with neural crest tumors¹⁹ and animals,²⁰ children with neuroblastoma or patients with pheochromocytoma were never studied using ¹²⁴I-MIBG. A more efficiently labeled form of MIBG (no-carrier-added) is currently in clinical trials (Ultratrace, Molecular Insight Pharmaceuticals, Cambridge, MA). This technology allows more activity to be administered with less of the chemical MIBG and ensures that every molecule taken up by the human norepinephrine transporter (hNET) will carry a radioactive iodine molecule.

As ¹³¹I-MIBG therapy becomes more common and of proven efficacy for neuroblastoma and pheochromocytoma,^{21, 22, 23124}I-MIBG pretherapy imaging and dosimetry in these patients will improve efficacy and safety with individualized adjustment of activity administered. As a first step, we performed preclinical imaging with no-carrier-added ¹²⁴I-MIBG using animal models in order to assess expected radiation exposures to human subjects. Data extrapolation and radiation dose estimations for models of different age groups are presented. The data presented in this report will be critical components for the Food and Drug Administration (FDA)’s Investigational New Drug application approval.

METHODS AND MATERIALS

¹²⁴I-MIBG and Animal Models

We have labeled ¹²⁴I (purchased from IBA Molecular, Louvain-la-Neuve, Belgium) with MIBG synthesized from precursors provided by Molecular Insight Pharmaceuticals (Cambridge, MA).^{24, 25, 26, 27, 28} From the reversed phase high performance liquid chromatography spectral analysis, we observed radiochemical purity of >99%. For thyroid blocking, 0.374 mg∕100 μL of aqueous solution of sodium iodide was administered intravenously to four athymic mice (weighing approximately 25 g each). The mice were implanted with NB1691 cells that express hNET. We originally intended to investigate the tumor uptake of ¹²⁴I-MIBG in these animals. However, the subcutaneous xenograft tumors in these animal models did not accumulate ¹²⁴I-MIBG to a detectable level, possibly due to the difference in tumor microenvironments between native tumors and xenograft tumors. Thus, all four mice were considered as normal control subjects, and radiation dosimetry studies followed.

Small animal PET∕CT

A total of 3.07–4.84 MBq of ¹²⁴I-MIBG was administered to four mice via tail vein. Small animal PET combined with computed tomography (microPET∕CT) was performed immediately after ¹²⁴I-MIBG administration. We used a dedicated PET docked with multimodality CT (Inveon, Siemens Healthcare, Malvern, PA) for all of our preclinical imaging studies. Each animal was followed for over 96 h by imaging five time points (0.5, 2, 24, 52, and 96 h). Mice were under anesthesia using isoflurane at 2% concentration.

In order to obtain reasonable image qualities, we applied two PET data acquisition durations. At 0.5, 2, 24, and 52 h reference time points, the data were acquired over 3600 s, and at 96 h, 5400 s. For the 0.5 h reference time point, we used a dynamic acquisition technique to capture early pharmacokinetics of the tracer immediately after the tracer administration. Thus, the actual acquisition time was from 0 to 1 h continuously, making the midtime point as reference at 0.5 h. The durations for the dynamic acquisition were 10×10, 5×40, 1×300, and 5×600 s. The 2 h data were also divided into two 1800 s frames.

PET data acquired using a 250–650 keV energy window were reconstructed to a 128×128×159 matrix with a voxel size of 0.776 383×0.776 383×0.796 mm³ using the 2D-ordered subsets expectation maximization reconstruction algorithm. The CT images were reconstructed into a 512×512×672 matrix with an isotropic voxel size of 0.190 684×0.190 684×0.190 684 mm³ using a Feldkamp CT reconstruction algorithm modified for x-ray cone beam. The reconstruction algorithms for both PET and CT were provided by the scanner manufacturer. The parameters for microCT were 120 rotation steps over 220°, continuous acquisition, 80 kVp tube voltage, 500 μA tube current, and 175 ms exposure.

Image display and analysis were performed using the AMIDE software package.²⁹ Volumes of interest (VOIs) were drawn on the coregistered CT images for thyroid, heart, lung, liver, kidney, and bladder, and transferred to PET images to obtain the pharmacokinetic data. Once transferred to PET images, VOIs were adjusted to include apparent partial volume spill-outs for organ uptake calculation. VOIs of the thyroid were drawn with the reference to thyroid ¹²⁴I-MIBG PET images because the size of the thyroid was anatomically too small to be localized on CT.

The animal organ % of injected activity (%IA) data were converted to those for a human using the percent kilogram per gram method of Kirschner et al.³⁰ The calculated percent injected activity per gram of organ (%IA∕g) for the organs in mice was extrapolated to uptake in organs of a 73.7 kg male adult.

In this method, the human %IA∕organ is given as in Ref. 31

$${\left(\frac{\mathrm{\%}\text{IA}}{\text{organ}}\right)}_{\text{human}}=\left[{\left(\frac{\mathrm{\%}\text{IA}}{{\text{g}}_{\text{organ}}}\right)}_{\text{animal}}\times {\left({\text{kg}}_{\text{TBweight}}\right)}_{\text{animal}}\right]\times {\left(\frac{{\text{g}}_{\text{organ}}}{{\text{kg}}_{\text{TBweight}}}\right)}_{\text{human}}.$$ (1)

The calculated %IA∕organ in the human organs was fitted with exponential biokinetic functions and integrated to obtain the number of disintegrations for source organs using the OLINDA|EXM (version 1.1) software.³² The same software was used to estimate the effective dose in different age groups. The extracted organ masses and fractions for our radiation dose estimations between mouse and adult male are shown in Table 1. The values of mouse organ masses and fractions were extracted from the standard Hui et al. model.^{33, 34}

Table 1.

Fractions of the total body mass of a reference adult male human (73 700 g) and mouse (25 g).

Organ	Mouse		Human
	Mass (g)	Fraction	Mass (g)	Fraction
Thyroid	0.06	0.0024	20.7	0.0003
Heart	0.12	0.0048	316	0.0043
Lung	0.15	0.0060	1000	0.0136
Liver	1.05	0.0420	1910	0.0259
Kidney	0.27	0.0106	299	0.0041
Bladder	0.09	0.0036	47.6	0.0006

Dosimetric calculations

OLINDA|EXM software output is the radiation dose per unit of administered activity (mSv∕MBq) in each organ and the total body, as well as effective dose (ED) and effective dose equivalent.^{32, 35} Inputs to OLINDA|EXM are either residence times themselves (uCi h∕uCi, Bq h∕Bq) or %IA∕organ from which residence times are computed via exponential curve fits. The OLINDA phantoms are adjusted accordingly for different age groups from newborns to human adults. The time-activity-curves, expressed as percent injected activity per weight of tissue or percent injected activity of tissue (%IA∕organ), were generated for the organs of interest. Effective dose was estimated for different age groups using the tissue weighting factors defined in the International Commission on Radiological Protection (ICRP) Publication 103 (Ref. ³⁶) as well as the values in the ICRP Publication 60.³⁷ The current version of OLINDA|EXM (version 1.1) uses the ICRP Publication 60 tissue weighting factors.

RESULTS

Figure 1 shows selected slices of axial, coronal, and sagittal views of microPET∕CT images from ¹²⁴I-MIBG from one of the animals included in our calculations. The five time points over 96 h show typical kinetics of ¹²⁴I-MIBG uptake patterns. High uptake is found in the heart and bladder before the 2 h time point and thyroid uptake is heightened after the 24 h time point. Uptake in liver, kidneys, and lungs decreased continuously over time; however, the percent liver and lung uptake remained relatively high when compared to uptake in other organs even at the 96 h time point mostly due to their large fractions of organ masses. Table 2 lists the mean %IA∕g values in each reported organ of mice. The %IA∕g values were calculated by the fractions of activity in each organ (%IA) divided by organ masses in g (e.g., 0.12 g for heart in Table 1).

Figure 1.

Representative axial, coronal, and sagittal PET/CT views of ¹²⁴I-MIBG in a mouse. The images were acquired at 0.5, 2, 24, 52, and 96 h after intravenous injection of ¹²⁴I-MIBG. Lu=Lung; He=Heart; Li=Liver; Ki=Kidney; Bl=Bladder; and Th=Thyroid.

Table 2.

Biodistribution of ¹²⁴I-MIBG on mice at different time points (% IA∕g). Values are expressed as %IA∕g, mean±standard error of the mean.

Time (h)	Thyroid	Heart	Lung	Liver	Kidney	Bladder
0	24.00±3.89	20.88±4.07	15.44±4.86	13.06±0.94	9.75±1.05	242.51±37.66
2	41.42±5.08	19.98±3.70	14.46±2.83	11.64±1.34	9.29±1.32	124.41±30.68
24	73.47±3.10	10.75±0.75	9.42±1.13	7.36±2.22	7.39±1.47	69.44±23.12
52	46.77±9.44	7.43±1.43	7.49±1.82	4.77±0.90	5.67±0.66	79.34±33.78
96	32.04±5.72	2.81±0.34	4.58±1.56	2.94±0.83	1.27±0.36	21.24±4.85

Figure 2 illustrates extrapolated human-equivalent percent injected activity (%IA∕organ) data in heart, thyroid, bladder, kidney, liver, and lung. The values shown in Fig. 2 are averaged across all four mice. The %IA∕organ values show the expected patterns of pharmacokinetics. Since the thyroid was blocked from ¹²⁴I-MIBG by sodium iodide, the uptake in the thyroid was minimized. This process mimics the clinical situation in which patients are given potassium iodide (KI) as a thyroid-blocking agent.

Figure 2.

Mean percent injected activity and standard error of the mean for ¹²⁴I-MIBG uptake in heart, thyroid, bladder, kidney, liver, and lung. Data are expressed as the extrapolated human-equivalent uptake values (%IA∕organ).

The estimated equivalent dose to the organs and total body, and effective dose, for the reference adult male are summarized in Table 3. The mean effective dose to the adult male was estimated to be 0.25 mSv∕MBq. The highest mean equivalent dose was in the thyroid, at 2.343 mSv∕MBq. Calculated effective doses for the different age groups are shown in Table 4. Effective dose values were higher in female than in male and are higher in the younger subjects than in the older subjects as expected.

Table 3.

Estimated equivalent doses and effective dose from ¹²⁴I-MIBG in reference adult male human. Values are expressed as mean±standard error of the mean.

Target organ	Dose estimate
	mSv∕MBq	rem∕mCi
Adrenals	0.117±0.014	0.434±0.052
Brain	0.038±0.012	0.141±0.046
Breasts	0.064±0.012	0.235±0.044
Gallbladder wall	0.141±0.018	0.521±0.065
LLI wall	0.064±0.017	0.238±0.063
Small intestine	0.069±0.015	0.257±0.056
Stomach wall	0.076±0.014	0.282±0.051
ULI wall	0.074±0.014	0.275±0.050
Heart wall	0.449±0.035	1.663±0.130
Kidneys	0.316±0.038	1.170±0.141
Liver	0.472±0.112	1.750±0.414
Lungs	0.428±0.101	1.585±0.370
Muscle	0.061±0.013	0.226±0.048
Ovaries	0.067±0.017	0.249±0.063
Pancreas	0.106±0.014	0.392±0.050
Red marrow	0.062±0.012	0.230±0.045
Osteogenic cells	0.075±0.019	0.277±0.071
Skin	0.042±0.010	0.155±0.037
Spleen	0.074±0.014	0.273±0.052
Testes	0.049±0.014	0.182±0.053
Thymus	0.082±0.015	0.304±0.055
Thyroid	2.343±0.306	8.663±1.129
Urinary bladder wall	0.462±0.066	1.710±0.245
Uterus	0.084±0.019	0.310±0.069
Total body	0.080±0.012	0.295±0.046
Effective dose (ICRP 60)	0.248±0.021	0.918±0.078
Effective dose (ICRP 103)	0.252±0.021	0.932±0.078

Table 4.

Mean effective dose in different age groups for ¹²⁴I-MIBG extrapolated from animal data.

Age	Mass (g)	Effective dose (mSv∕MBq)
		ICRP 60	ICRP 103
Adult male	73 700	0.248	0.252
Adult female	57 000	0.348	0.342
15-yr-old	56 840	0.381	0.388
10-yr-old	33 200	0.568	0.578
5-yr-old	19 800	1.032	1.027
1-yr-old	9720	1.926	1.918
Newborn	3600	3.637	3.752

DISCUSSION

We have performed internal radiation dosimetry of ¹²⁴I-MIBG using microPET∕CT in mice for normal organs with extrapolation to human-equivalent radiation dose values. Our results show that ¹²⁴I-MIBG PET is a feasible imaging modality for pretherapy dosimetry and the future possibility of its use for the prediction of tumor dose as well. The tissue weight factors extrapolation method [Eq. 1] from mice to humans was based on preclinical imaging data. Therefore, the results are estimates of human absorbed dose since this method has uncertainties associated with extrapolation methods.³⁸ In our study, we also compared the effective dose estimates of ¹²⁴I-MIBG to effective doses reported for commonly known radiotracers (Table 5).³⁹ The most significant finding from this comparison was that the estimated effective dose value of ¹²⁴I-MIBG was nearly ten times smaller than that of ¹²⁴I-NaI. Such discovery is important because ¹²⁴I-NaI is already used clinically at selected research centers.^{6, 7, 9, 10, 40} The difference in effective dose values is primarily due to the thyroid blocking, which is performed for ¹²⁴I-MIBG studies, but not for ¹²⁴I-NaI studies. The ratio of estimated effective dose of ¹²⁴I-NaI to that of ¹²⁴I-MIBG (6.5∕0.25) is comparable to the reported effective dose ratios of ¹²³I-NaI to ¹²³I-MIBG (0.12∕0.019) and ¹³¹I-NaI to ¹²³I-MIBG (11.0∕0.21). The significant difference in effective doses between NaI and MIBG compounds primarily results from the different patterns of the two radiotracers. While the iodide anion will be taken up in the thyroid, the iodinated MIBG is specifically taken up by sympathetically innervated tissue (such as salivary gland) or tissue with the organic cation transporter (such as liver).⁴¹

Table 5.

Comparisons of reference adult male estimated effective dose between ¹²⁴I-MIBG and selected known radiotracers.

Tracers	Effective dose (mSv∕MBq)
124I-MIBG (est.)	0.25
124I-NaI	6.50
123I-MIBG	0.019
123I-NaI	0.12
131I-MIBG	0.21
131I-NaI	11.0
18F-FDG	0.03

The 0.25 mSv∕MBq radiation dose estimated for ¹²⁴I-MIBG is significantly smaller than that of ¹²⁴I-NaI (6.5 mSv∕MBq). However, this estimated dose is tenfold higher than that of ¹²³I-MIBG (0.019 mSv∕MBq). Hence, as a diagnostic imaging agent that is used to identify MIBG-avid tumors, ¹²⁴I-MIBG may not be an ideal choice despite the quantitative ability of PET. For this reason, ¹²⁴I-MIBG PET should be administered at low doses especially for small children, and the best indication of ¹²⁴I-MIBG would be as a pretherapy dosimetry tool before ¹³¹I-MIBG treatment for which the pretherapy radiation dose can be considered negligible by comparison. Particularly in younger pediatric patients, when ¹²⁴I-MIBG is used for pretherapy dosimetry of ¹³¹I-MIBG treatment of neuroblastoma, the activity administered should be considered carefully. For example, for children under 5 years old, the estimated effective radiation dose is 1.03 mSv∕MBq from ¹²⁴I-MIBG. In this case, if 18.5 MBq (0.5 mCi) of ¹²⁴I-MIBG is administered, the radiation exposure is estimated at 19.1 mSv, which is already approximately two thirds of effective dose from 555 MBq (15 mCi) administration of ¹⁸F-flurodeoxyglucose (FDG) for adult male, 33.3 mSv using the tabulated values listed in Table 5. An 18.5 MBq administration may appear to be an insufficient amount of activity to achieve acceptable quality PET images. However, the current PET technologies feature improved sensitivity over previous generations due to increased fields of view, 3-D acquisition. The possibility of allowing relatively long acquisition time per bed, particularly for small children without a requirement for increasing the total acquisition time, will result in accurate data for dosimetric calculations.

Several issues limited the accuracy of the dose calculation method used. In order to estimate the radiation dose using small animal data with high accuracy, we need to pay careful attention to a few specific points. One limitation of this work is the broad range of mouse organ masses from different animals. The underestimation of radiation dose in small organs can be caused by the limited spatial resolution of the microPET scanner when imaging ¹²⁴I-labeled radiotracers. However, the estimates of average whole-body dose generally are not significantly affected by partial volume errors.⁴² Identifying accurate boundaries of organs is another practical difficulty. In addition, the radiation dose estimate might change depending on how many time points were used to acquire the data, which generally affects the quality of curve-fitting procedures using biokinetic time-activity-curves.

We also considered the radiotracer excretion model in our radiation dosimetry based on the total activity changes in the whole body of mice over time. For all imaging time points, the entire body of mice was included in the PET∕CT field of view for this calculation. Radiation dose estimates including our excretion model were compared against the published ¹³¹I-MIBG (no-carrier-added) pharmacokinetic and dosimetry data in humans.⁴³ Although technologies of estimating the radiation dose are quite different (microPET data extrapolation from mouse to human versus two-view anterior-posterior scintigraphy, which lacks robust quantitative measurements of radiotracer distribution), our preclinical ¹²⁴I-MIBG data were able to predict reasonably precise radiation dose estimates for ¹³¹I-MIBG in humans within 15% of ED difference (0.34 mSv∕MBq using ¹²⁴I-MIBG mouse biodistribution data converted for human dose estimates of ¹³¹I versus 0.28 mSv∕MBq calculated using residence times of each organ reported in Coleman et al.⁴³ for no-carrier-added ¹³¹I-MIBG). Residence times predicted by the preclinical data of ¹²⁴I-MIBG were also in general agreement with those reported in Coleman et al. (Table 6).

Table 6.

Comparison of ¹³¹I-MIBG residence times (data from Coleman et al.) in humans (Ref. ⁴³) versus estimated ¹³¹I-MIBG residence times extrapolated from 124I-MIBG microPET data (this manuscript).

	131I-MIBG residence times (Coleman et al.)		Estimated 131I-MIBG residence times from 124I-MIBG microPET data
	Mean (h)	Range (h)	Mean (h)	Range (h)
Thyroid	0.45	0.27–0.72	0.53	0.38–0.72
Heart	0.80	0.48–1.18	0.88	0.80–0.96
Lung	2.97	1.69–4.89	3.27	2.50–2.86
Liver	5.37	2.57–11.92	5.03	3.37–6.75
Kidney	1.18	0.75–2.30	0.56	0.47–0.64
Bladder	2.01	1.63–2.19	1.20	0.87–1.47
Brain	0.54	0.28–0.78	−	−
Salivary gland	1.10	0.68–1.85	−	−
Spleen	0.82	0.50–1.28	−	−
Tumors	0.92	0.25–2.09	−	−
Remainder	23.25	15.21–40.39	18.86	9.26–24.54
Total body	36.54	23.86–65.18	30.33	22.46–34.39

CONCLUSION

Radiation absorbed dose estimation for internally administered radiopharmaceuticals using preclinical models is critical when translating a new radiopharmaceutical or a new indication using an existing radiopharmaceutical into clinical trials. The dose estimates for ¹²⁴I-MIBG using small animal models in our report showed that ¹²⁴I-MIBG would deliver a significantly smaller radiation dose than ¹²⁴I-NaI, a radiotracer already being used in human subjects. Our data support the safety of using ¹²⁴I-MIBG in humans as long as a relatively small amount of radioactivity is used. For example, the range of 0.5–1 mCi administration would result in effective dose estimates of 19.1–38.1 mSv in the 5-yr-old model. Also, as our results showed, for very young pediatric patients, particularly under 5 years old (effective dose estimates over 1 mSv∕MBq), ¹²⁴I-MIBG imaging should be used judiciously and mainly in conjunction with a planned ¹³¹I-MIBG therapy dose, which would then make the pretherapy imaging dose negligible by comparison.