A risk index for pediatric patients undergoing diagnostic imaging with ^99mTc-dimercaptosuccinic acid that accounts for body habitus

Abstract

Published guidelines for administered activity to pediatric patients undergoing diagnostic nuclear medicine imaging are currently obtained through expert consensus of the minimum values as a function of body weight as required to yield diagnostic quality images. We have previously shown that consideration of body habitus is also important in obtaining diagnostic quality images at the lowest administered activity. The objective of this study was to create a series of computational phantoms that realistically portray the anatomy of the pediatric patient population which can be used to develop and validate techniques to minimize radiation dose while maintaining adequate image quality. To achieve this objective, we have defined an imaging risk index that may be used in future studies to develop pediatric patient dosing guidelines. A population of 48 hybrid phantoms consisting of non-uniform B-spline surfaces and polygon meshes was generated. The representative ages included the newborn, 1 year, 5 year, 10 year and 15 year male and female. For each age, the phantoms were modeled at their 10th, 50th, and 90th height percentile each at a constant 50th weight percentile. To test the impact of kidney size, the newborn phantoms were modeled with the following three kidney volumes: −15%, average, and +15%. To illustrate the impact of different morphologies on dose optimization, we calculated the effective dose for each phantom using weight-based ^99mTc-DMSA activity administration. For a given patient weight, body habitus had a considerable effect on effective dose. Substantial variations were observed in the risk index between the 10th and 90th percentile height phantoms from the 50th percentile phantoms for a given age, with the greatest difference being 18%. There was a dependence found between kidney size and risk of radiation induced kidney cancer, with the highest risk indices observed in newborns with the smallest kidneys. Overall, the phantoms and techniques in this study can be used to provide data to refine dosing guidelines for pediatric nuclear imaging studies while taking into account the effects on both radiation dose and image quality.

1. Introduction

Current nuclear medicine dosing guidelines for pediatric patients are based on patient weight (Lassmann and Treves 2014). However, one would expect that patient habitus would be an additional contributing factor to both radiation dose and image quality. To evaluate the effect of accounting for body habitus while patient weight is kept constant, we have defined an imaging risk index (RI) that may be used to develop pediatric patient dosing that accounts for both patient weight and body habitus. Minimizing radiation dose, especially in pediatric patients, has the potential to reduce the risk of radiation-induced cancers. The objective of this study was to create a series of pediatric phantoms that realistically portray variations in pediatric anatomy so as to facilitate development and validation of techniques to minimize radiation dose while maintaining adequate image quality. The radiopharmaceutical investigated in this study was ^99mTc-dimercaptosuccinic acid (DMSA).

Pediatric patients are more vulnerable to radiation induced malignant neoplasms than adults due to the higher radiosensitivity of their tissues, body size, and their inherently longer lifespans during which stochastic radiation effects may develop (Hall 2006). Atomic bomb survivor data show lifetime cancer risk has a strong dependency on age of exposure, with a risk of about 15% per Sv at an exposure age of less than 10 years compared to 1% per Sv at age 60 (BEIR 2005).

National campaigns to limit doses to pediatric patients have been established, such as ‘Image Gently’ which was initially associated with computed tomography but later expanded to fluoroscopy and nuclear medicine. Image Gently, in collaboration with the Society of Nuclear Medicine and Molecular Imaging (SNMMI), the Society of Pediatric Radiology, and the American College of Radiology developed the North American Guidelines for common radiopharmaceuticals including DMSA (Gelfand et al 2011). The ‘Go with the Guidelines’ campaign sought to inform every nuclear medicine practice in North American regarding these guidelines. For DMSA, it also provides a footnote stating the European guidelines, which had been previously developed (Lassmann et al 2007), could also be used. In 2014, the European Association of Nuclear Medicine (EANM) and the North American group collaborated to develop harmonized guidelines (Lassmann and Treves 2014). However, the two sets are still not fully identical.

Approximately 180 000 pediatric nuclear medicine studies are conducted in the United States each year (Fahey et al 2011). Of these studies, approximately 53% are renal imaging studies (Conway 2007). Tc-99m-DMSA is typically used for evaluating pyelonephritis, due to its high sensitivity (96%) and specificity (98%), and for detecting cortical scars (Shammas et al 2013). Ultrasound, which eliminates the use of ionizing radiation, is another technique that can be used to detect kidneys scars. However, it was found that DMSA detected kidney scars in 35% of kidneys determined to be ‘normal’ under ultrasound (Temiz et al 2006). Other areas where DMSA is a more effective evaluation tool than sonography include renal dysplasia, renal agenesis, and renal ectopia (Alazraki and Braithwaite 2014). These renal anomalies are more common in males than females. This is important as gender effects on dose and radiation induced cancer risk were evaluated in the present study.

2. Materials and methods

2.1. Phantom series construction

For this study, a series of 48 phantoms was constructed. The parameters altered between the various phantoms were age, gender, height, and kidney mass. For each gender, five age groups were considered: 0 years, 1 years, 5 years, 10 years, and 15 years. For each age group, three height percentiles corresponding to the 50th percentile weight for that patient age were considered: 10th, 50th and 90th. The 10th and 90th percentile heights for each age group at the reference weight were obtained from growth charts published by the U.S. Centers for Disease Control and Prevention (http://www.cdc.gov/growthcharts). For the 10th and 90th percentile phantoms, the closest height and weight matched phantom from the UF/NCI family of hybrid computational phantoms library was selected (Geyer et al 2014). As the pediatric phantom library only encompasses ages 2 to 20 years, the newborn and 1 year-old 10th and 90th percentile phantoms, at the given 50th percentile weight, were constructed specifically for this study. Their construction was achieved by scaling the UF/NCI reference newborn and 1 year-old phantoms to match the desired height while maintaining a constant weight. The targeted heights and weights for each phantom are provided in table 1. Graphical renderings of phantoms for each height percentile and age group are shown in figure 1.

Table 1.

Targeted height and weight for pediatric patient population.

Age (years)	Height percentile	Males		Females
		Height (cm)	Weight (kg)	Height (cm)	Weight (kg)
0 year	10th	47	3.5	46.5	3.4
	50th	50	3.5	49	3.4
	90th	54	3.5	53	3.4
1 year	10th	72	10.4	70	9.5
	50th	76	10.4	74.8	9.5
	90th	79.5	10.4	77.4	9.5
5 years	10th	103	18	102	18
	50th	109	18	107.5	18
	90th	115	18	114	18
10 years	10th	130.4	32	129.6	33
	50th	138	32	138	33
	90th	147	32	147	33
15 years	10th	159.5	56	153.5	52
	50th	170	56	162	52
	90th	179.8	56	170.2	52

Figure 1.

Rendering of 10th, 50th, and 90th percentile height at constant 50th percentile weight newborn, 1 year-old, 5 years-old, 10 years-old, and 15 years-old hybrid phantoms.

Kidney reference masses reported in ICRP Publication 89 (ICRP 2002) were used for the 50th percentile height phantoms with ‘average’ kidney sizes. To investigate the dependence of kidney size on organ absorbed dose and patient risk, the kidney mass was varied uniformly by −15%, average, and +15% for the newborn male and female. The newborn age group was chosen for this organ size study as the greatest effects would be seen in this patient group due to their smaller stature and intra-organ proximity. Kidney masses for each phantom are listed in table 2.

Table 2.

Kidney masses for pediatric phantom population.

		Males Kidney mass (g)			Females Kidney mass (g)
Age (years)	Height percentile	−15%	Average	+15%	−15%	Average	+15%
0	10	19.8	23.2	26.7	19.8	23.2	26.7
	50	21.3	25.0	28.8	21.3	25.0	28.8
	90	23.6	27.8	31.9	23.6	27.8	31.9
1	10		69.7			69.7
	50		70.0			70.0
	90		72.3			72.3
5	10		93.1			93.1
	50		110.0			110.0
	90		100.1			100.1
10	10		135.0			126.8
	50		180.0			180.0
	90		155.2			155.2
15	10		204.7			215.1
	50		250.0			240.0
	90		172.6			196.0

2.2. Radiation transport

Over 700 radiation transport simulations were completed using MCNPX v2.7 (Pelowitz 2011). Each simulation directly sampled the ^99mTc emission spectrum. Decay schemes for ^99mTc were obtained from ICRP Publication 107 (ICRP 2008). The number of particles simulated ranged from 150 million to 600 million, increasing with subject age. Typically, the MCNPX tally relative errors were well below 1%, with some exceptions for highly separated source-target organ pairs, such as heart contents to gonads. In these cases, the tally relative errors were 2–3% or below. This was deemed acceptable as the contributions to total dose from such source-target combinations were orders of magnitude smaller than more significant source-target combinations, such as the kidney self-dose.

Based on the biokinetics of ^99mTc-DMSA, the following were designated as source organs: kidney cortex, kidney medulla and renal pelvis, lungs, urinary bladder contents, liver, heart contents, spleen, and rest of body. Target organs were based on those used in the calculation of the effective dose and included colon wall, lungs, stomach wall, gonads (ovaries and testes), urinary bladder wall, esophagus, liver, thyroid, brain, salivary glands, breast, skin, active marrow, bone endosteum, and remainder organs (ICRP 2007). Tissue weighting factors used for effective dose calculations were taken from ICRP Publication 103 (ICRP 2007).

2.3. Biokinetics

The effective dose calculations for ^99mTc-DMSA pediatric patients provided in the EANM pediatric dosage card (Lassmann and Treves 2014) were obtained from ICRP Publication 80 (ICRP 1998), which are also included in the recent compendium document ICRP Publication 128 (ICRP 2015). The biokinetic models in these publications were based on adults and the effective doses provided for this radiopharmaceutical have not changed between publications. To improve upon those previous efforts, age-dependent biokinetic model parameters were used in the present study. The age-specific time-integrated activity coefficients used are provided in table 3. These time-integrated activity coefficients were derived as described in Sgouros et al (2011).

Table 3.

^99mTc-DMSA time-integrated activity coefficients (h) for pediatric patients.

Source organ/tissue	Newborn	1 year-old	5 years-old	10 years-old	15 years-old
Kidney	2.99	2.79	3.04	2.98	3.28
Bladder contents	0.11	0.15	0.17	0.27	0.25
Heart contents	0.10	0.11	0.11	0.11	0.13
Liver	0.43	0.24	0.34	0.35	0.36
Spleen	0.10	0.11	0.09	0.20	0.17
Lungs	0.12	0.15	0.14	0.14	0.17
Rest of body	4.59	4.21	4.06	3.68	3.68

The time-integrated activity coefficients were derived based on a data set consisting of 15 children with normal renal function who underwent routine diagnostic imaging with ^99mTc-DMSA, as reported in Smith et al (1996).

2.4. RI calculation

Effective dose is not age- and gender-specific and cannot, therefore, be used to for diagnostic image quality optimization of the administered activity (AA) to the different morphological pediatric patient categories. We, therefore, define the RI as a ratio of the risk of radiation induced cancer from a DMSA study to the natural incidence of cancer. The estimated cancer risk is the possible radiation induced risk from a DMSA imaging study per 100 000 patients as a function of subject age (a) and gender (g) and the natural incidence of cancer is that of an unexposed population per 100 000, both of which are age (a) and gender (g) dependent.

$$\text{Risk Index}=100\times \frac{{\text{Estimated Radiation Induced Cancer Risk}}_{\mathrm{a},\mathrm{g}}}{{\text{Natural Incidence of Cancer}}_{\mathrm{a},\mathrm{g}}}$$ (1)

In order to scale the ratio to a percent, the multiplicative factor of 100 is used. For example, a RI of 0.1 would suggest the patient has a 0.1% higher chance of having a radiation induced cancer when compared to their natural probability of cancer incidence. In an optimization process, the greater the RI disparity amongst different phantoms, the greater the possibility that patient-specific optimization will lead to pediatric dose reduction. This RI will be a numerical parameter to optimize administered activities of DMSA as a function of patient gender and body habitus, coordinating with results of an ongoing image quality study. The RI is not intended to directly inform on the risk to any individual pediatric ^99mTc-DMSA study patient. Nevertheless, it is to be used in the optimization of AA guidance for specific patient populations.

To determine the estimated cancer risks to the pediatric DMSA patients, the program Radiation Risk Assessment Tool (RadRAT), developed by the National Cancer Institute (NCI), was used. This program employs risk models primarily based on those provided in the BEIR VII Report (BEIR 2005), along with seven additional models derived from Japanese bomb survivor data (de Gonzalez et al 2012). Of these additional models, the most relevant to this study is the introduction of a kidney risk model, as approximately half of ^99mTc-DMSA activity is distributed within the renal cortex (ICRP 2015).

The total absorbed dose to each target organ from the combination of all individual source organs was applied in these calculations. The organs in the risk calculations for the female phantoms were oral cavity and pharynx, esophagus, stomach, colon, liver, lung, breast, ovaries, bladder, nervous system, thyroid, kidney, and leukemia (bone marrow dose). The same organs were used for the male, with the exception of the gender specific organs: breast and ovaries. The setting used for exposure rate was acute, as it is based on one DMSA imaging study. The distribution type used was ‘fixed value’, with the value being the absorbed dose from Monte Carlo calculations for that organ. As only the mean value of risk is used in this paper to obtain a RI, the standard deviation of the Monte Carlo calculations, which were typically less than 1%, was not used. It is fully acknowledged that it is questionable to use risk models for individual patients with effective doses less than 100 mSv. Nevertheless, these models provide an approach for optimization studies in pediatric nuclear medicine that are age- and gender-specific.

These cancer risks were then used to compute the RI. The natural cancer incidence per 100 000 in an unexposed population for each age and gender was gleaned from the SEER database (Howlader et al 2014). For the study of the effects of kidney size on radiation induced kidney cancer risk, the natural occurrence of kidney/renal cancer was used and also abstracted from the SEER database.

3. Results and discussion

3.1. Effective dose

Radionuclide S-values for each age, gender, and height percentile for the source and target organs for ^99mTc-DMSA are provided in the supplementary material available at stacks.iop.org/PMB/61/2319/mmedia. The effective dose for each age group and height percentile was calculated using the North American and European AA guidelines and is provided in table 4. Effective doses calculated in this study are less than those previously documented in ICRP Publication 128 (ICRP 2015). Figure 2 compares effective doses between this study and those provided by the ICRP. These differences are most likely due to the age-specific biokinetic models and the more anatomically detailed phantoms applied in the current study. A recent study by Andersson et al (2014) calculated the effective doses resulting from 338 radiopharmaceuticals to an adult male using the ICRP computational reference phantom. They found that their calculated effective doses per AA for 79% of the radiopharmaceuticals, when compared to past ICRP calculations, gave a lower effective dose. However, DMSA was not one of the 79%. In this study, which is strictly to pediatric patients, the effective doses per AA were also lower than previous ICRP calculations.

Table 4.

Effective doses (mSv) for each age group and height percentile using the North American and European guidelines for ^99mTc-DMSA.

Age (years)	Height percentile	North American guidelines	EANM guidelines
1	10	0.60	0.60
	50	0.57	0.57
	90	0.55	0.55
5	10	0.67	0.61
	50	0.60	0.55
	90	0.63	0.57
10	10	0.88	0.73
	50	0.75	0.62
	90	0.80	0.66
15	10	1.06	0.87
	50	0.98	0.81
	90	1.08	0.89

Figure 2.

Comparison of effective dose per AA (mSv/MBq) calculated in this study versus those provided in ICRP Publication 128 for pediatric patients undergoing a DMSA imaging study.

Figure 3 shows the effective doses for 10th, 50th, and 90th percentile height pediatric patients at 50th percentile weight using the North American guidelines. These effective doses are compared to the annual natural background radiation (3.1 mSv) (NCRP 2009). For each age group, the 10th percentile height phantom had a higher effective dose than the 50th percentile height phantom. The greatest difference was observed between the 10th and 50th percentile height 10 year phantom (17.6%). The average difference between the effective doses of the 10th and 50th percentile height phantoms, across all ages, was 9.8%. The average difference between the effective doses of the 50th and 90th percentile height phantoms, across all ages, was 6.5%.

Figure 3.

Effective dose (mSv) for 10th, 50th, and 90th percentile height pediatric patients at 50th percentile weight from a DMSA imaging study compared to annual natural background radiation an individual is exposed to.

A previous study investigating body habitus effects on DMSA imaging for a 10 year-old female found that the same image quality could be achieved in a ‘tall, thin’ patient with approximately half of the AA needed for a ‘short, heavy’ patient (Sgouros et al 2011). This suggests that, since the doses for the 90th percentile patients are currently higher than their 50th percentile height counterparts for the 5 year, 10 year and 15 year old patients, doses could be reduced by decreasing the AA without sacrificing diagnostic image quality. These findings illustrate the need for separate dosing guidelines dependent on body morphometry and not body weight alone.

3.2. RI comparisons

The RI for each age and gender at the 50th percentile height is displayed in figure 4. In order to show the strong effect kidney dose has on the overall excess cancer risk compared to other target organs, these two contributions are displayed separately. This illustrates the importance of the improved kidney risk model provided in the RadRAT software, as compared to the previous BEIR VII models.

Figure 4.

RI for reference pediatric patients undergoing DMSA imaging studies due to dose to kidneys and other target organs.

The ratio of associated cancer risks for a newborn receiving the minimum AA of 18.5 MBq and the 1.85 MBq kg⁻¹ obtained from weight-based guidelines (yielding 6.8 MBq for a 50th percentile newborn) was on average 2.9. This is illustrated in figure 5, which shows the RI for reference newborns using the 18.5 MBq and 6.8 MBq AA’s compared to the index for 1 year-olds (18.5 MBq AA). The RI decreases and is close to that of the 1 year-old when the AA is 6.8 MBq. This result illustrates the need for consensus on the guideline for newborns undergoing DMSA imaging studies. The necessary AA to maintain image quality must be determined.

Figure 5.

RI for 10th, 50th, and 90th percentile height newborn males and females using 18.5 MBq and 6.8 MBq administered activities compared with RI of 1 year-olds.

The RI for each phantom in the study is provided in table 5. The percent difference in the Risk Indices for the 10th and 90th percentile phantoms when compared to the corresponding 50th percentile height phantom ranges from less than 1% to a high of 36.8%.

Table 5.

RI for each pediatric phantom and the percent difference from the 50th percentile height phantom.

Age (years)	Gender	Height percentile	Kidney size (if applicable)	Risk index	Percent difference from 50th percentile
0	Female	10	−15%	0.245	−18.43
			Average	0.225	−9.15
			+15%	0.208	−0.90
		50	−15%	0.226	−9.66
			Average	0.206
			+15%	0.196	5.28
		90	−15%	0.208	−0.9
			Average	0.190	7.86
			+15%	0.179	13.27
	Male	10	−15%	0.229	−20.95
			Average	0.203	−7.03
			+15%	0.183	3.64
		50	−15%	0.209	−10.04
			Average	0.190
			+15%	0.167	11.92
		90	−15%	0.194	−2.13
			Average	0.171	9.91
			+15%	0.154	18.82
1	Female	10		0.083	−3.65
		50		0.080
		90		0.077	3.32
	Male	10		0.067	−2.17
		50		0.064
		90		0.063	3.61
5	Female	10		0.090	−16.9
		50		0.077
		90		0.083	−7.93
	Male	10		0.083	−16.78
		50		0.071
		90		0.078	−9.34
10	Female	10		0.094	−36.78
		50		0.069
		90		0.078	−13.41
	Male	10		0.087	−28.74
		50		0.068
		90		0.077	−13.08
15	Female	10		0.084	−9.62
		50		0.077
		90		0.089	−15.12
	Male	10		0.104	−20.11
		50		0.087
		90		0.119	−37.77

3.3. Kidney mass effects for newborn

The highest RIs were observed in the shortest newborns with the smallest kidney size, and the smallest RIs were observed in the tallest newborns with the largest kidneys. However, the larger kidney sizes led to smaller effective doses due to the increased self-absorption of emitted energy within the kidney. The difference in trends between cancer risk and effective dose dependence on kidney size is due to the kidney cancer risks being calculated using the absorbed dose (mass dependent) to the kidney while effective dose incorporates all source-target combinations and weighting factors. There appears to be a correlation between kidney size and the radiation induced kidney cancer risk, as illustrated in figure 6. The same time-integrated activity coefficient (defined by age and gender) and therefore activity uptake was used for each kidney size. To determine if this correlation is true, a study using real patient data and images to determine possible uptake differences should be conducted.

Figure 6.

Kidney size effect on RI for newborn males and females.

3.4. Gender comparisons

Table 6 shows 80% of the pre-sex-averaged ‘effective’ doses per MBq were higher for the female phantoms than the male phantoms. This is most likely due to the relatively smaller gonad-to-kidney separation observed in females as compared to the males. The 90th percentile height 15 year-old male showed the greatest dose per AA difference when compared to the female counterpart due to this phantom having a much smaller kidney size. This phantom has a smaller kidney size as a result of the height scaling. In order to keep the same weight while increasing the height fairly significantly, the phantom’s abdominal pelvic region is decreased in the anterior–posterior direction, resulting in a compression the kidney and a decrease in kidney volume.

Table 6.

Pre-sex-averaged effective dose per AA for male and female pediatric phantoms.

Age (years)	Height Percentile	Male (mSv/MBq)	Female (mSv/MBq)	Ratio (M/F)
0	10	7.17E − 02	7.32E − 02	0.98
	50	6.65E − 02	6.77E − 02	0.98
	90	6.22E − 02	6.24E − 02	1.00
1	10	3.16E − 02	3.28E − 02	0.96
	50	3.03E − 02	3.13E − 02	0.97
	90	2.91E − 02	3.03E − 02	0.96
5	10	1.88E − 02	1.94E − 02	0.97
	50	1.67E − 02	1.77E − 02	0.95
	90	1.77E − 02	1.82E − 02	0.97
10	10	1.45E − 02	1.53E − 02	0.95
	50	1.27E − 02	1.27E − 02	1.01
	90	1.35E − 02	1.36E − 02	1.00
15	10	1.04E − 02	1.03E − 02	1.02
	50	9.55E − 03	9.66E − 03	0.99
	90	1.11E − 02	1.01E − 02	1.10

Typically, the RI is greater in females than in males. This is due to the incorporation of the breast and ovary risks. The RI of kidney cancer incidence for the females in each age group have a higher risk of radiation induced kidney cancer. This is expected since males are nearly twice as likely as females to develop kidney and renal pelvis cancers (http://seer.cancer.gov/) and natural incidence is the denominator of the defining equation for the RI.

4. Conclusions

Body morphometry, and not just patient weight, impact organ dose and cancer risks in pediatric nuclear medicine studies. Effective doses calculated in this study (using age-specific biokinetics and more detailed phantom modeling) are lower than those previously reported (ICRP 2015). To put these findings in perspective, the average annual background radiation dose received per person in the United States is 3.1 mSv, which is 2.9 times greater than the highest effective dose in this study from one DMSA exam (NCRP 2009). The introduction of a kidney excess cancer risk model by the NCI RADRAT code was important to this study. A correlation between kidney size in newborns with kidney cancer risk may exist; further studies using real patient data and images to determine possible uptake differences should be conducted for confirming this assertion. The RI developed in this study will be used in future studies to reduce pediatric absorbed dose while maintaining image quality across the different categories of pediatric patient morphometries as described in this work.