Validity of anthropometric measurements to characterize obesity among adult survivors of childhood cancer: a report from the St. Jude Lifetime Cohort Study

Abstract

Background

Childhood cancer survivors (CCS) are at risk for obesity. The purpose of this project was to determine which clinical measures of body composition are most accurate among CCS when compared to dual energy x-ray absorptiometry (DXA).

Methods

We evaluated agreement between body mass index (BMI), skinfolds percent body fat, and waist to height ratio (WHtR), with DXA among 1361 CCS (mean age 32.4±7.7), ≥10 years from diagnosis. We calculated sensitivity and specificity of BMI, skinfolds, and WHtR obesity classifications compared to DXA. Log-binomial regression, stratified by sex, was used to evaluate treatment-related factors for misclassification as non-obese by BMI, skinfolds, and WHtR.

Results

Mean body fat values for skinfolds were 23.3±7.7% (males) and 32.3±8.1% (females), and for DXA, 26.9±7.4% (males) and 38.4±7.7% (females). Pearson correlations between skinfolds and DXA were high: R=0.83 males, R=0.84 female. Skinfolds incorrectly classified 34.5% of obese males and 27.3% of obese females. BMI measures were the least sensitive with false negative rates of 46.4% (males) and 53.1% (females). Males exposed to abdominal/pelvic radiation were at increased risk for misclassification as non-obese by BMI (RR 1.57; 95% CI 1.25–1.95). The percentages classified as obese were highest when using DXA (63.1%, males; 84.8%, females) and lowest when using BMI (35.7%, males; 39.7%, females). Although skinfolds and WHtR underestimated the percentage classified as obese compared to DXA, the differences were not as large.

Conclusion

Findings suggest that skinfolds and WHtR are better than BMI for obesity classification in CCS. Clinicians should be aware of the high risk of misclassifying obese CCS as non-obese.

Introduction

Childhood cancer survivors (CCS) are at risk of becoming overweight and obese^1–5. This is problematic, as being overweight or obese compounds an already increased risk for cardiac morbidity and mortality in CCS^{6, 7}. Interventions designed to reduce obesity and fat mass are important for this population, as is the accuracy of defining target populations and outcome measures for interventions.

Dual energy x-ray absorptiometry (DXA) is considered the clinical standard to determine body composition⁸. Whole body fat percentage determined using DXA is highly correlated with percent body fat determined using bone densitometry⁸, the four-compartment model (underwater weighing)⁹ and magnetic resonance imaging¹⁰. However, assessment of body composition using DXA is not always available or desirable in general medical practice or in field research due to high financial costs associated with testing, or lack of accessibility. Furthermore, individuals with a history of radiation therapy may also choose to forgo unnecessary testing that would involve additional exposure. In these situations, body mass index (BMI), and skinfold and circumference measurements represent easily administered and inexpensive alternatives to assess body composition.

In clinical and research settings, BMI is a commonly used measure of body composition. However, BMI assesses weight relative to height, and does not differentiate between muscle, fat, or bone¹¹. Accordingly, BMI may not be the best measure of obesity among CCS as many survivors are of short stature¹² or have reduced bone^{13, 14} and muscle mass^{15, 16}. In fact, in a study that compared survivors of acute lymphoblastic leukemia (ALL) to age- and sex-matched controls, BMI was not found to differ between survivors and controls, even though survivors had higher body fat percentages than controls as determined by DXA⁵.

Alternative measures to BMI that have been used to characterize body composition in non-cancer populations include skinfold measurements and waist to height ratio (WHtR). In research settings, skinfold derived body fat percentages are used to calculate percent body fat and have been validated in healthy adults with correlations ranging from 0.71 to 0.82 when compared to DXA^{17, 18}. WHtR has been shown to be a good index of body adiposity in adults¹⁹ and has been shown to be a better predictor for cardiovascular events²⁰ and diabetes²¹ than BMI. In a recent study of healthy individuals by Brambilla et al, while skinfolds was the best predictor of percent body fat and percent trunk fat, WHtR was a better predictor than BMI or waist circumference alone²². Despite this, studies assessing the accuracy of skinfolds and WHtR to characterize body composition among adult survivors of childhood cancer are lacking.

The first aim of this investigation was to determine if skinfolds are able to accurately predict body fat percentage in long-term survivors of childhood cancer, when compared to DXA measure of body fat. The second aim was to determine if BMI, skinfolds, and/or WHtR are acceptable substitutes for DXA when assessing obesity in this population. Lastly, we explored patient- and treatment-specific characteristics of survivors whose anthropometric measurements misclassified their obesity status compared to DXA measurement.

Methods

Participants

Participants for this study were members of the St. Jude Lifetime (SJLIFE) cohort, a longitudinal study designed to evaluate health among adult survivors of childhood cancer as they age. Participants in SJLIFE undergo risk-based medical screening according to the Children’s Oncology Group Long-Term Follow-up Guidelines for Survivors of Childhood, Adolescent, and Young Adult Cancers²³ and participate in assessments of body composition, including skinfold measures, waist to height ratio assessments, and DXA scans. To be eligible for SJLIFE, cancer survivors must have been treated at St. Jude Children’s Research Hospital (SJCRH) and be 18 years of age or older, and at least ten years from diagnosis²³. Because SJLIFE is a retrospective cohort study with ongoing recruitment (additional survivors become eligible over time), potential participants are randomized for recruitment purposes in blocks of 50, to avoid early responder bias. Participants were randomized for inclusion in this study based on their time of recruitment into the SJLIFE cohort. Participants who completed anthropometric measures and received a DXA scan during their SJLIFE visit between October 2007 and April 2013 were eligible. Persons with amputation were excluded from analyses. Study procedures and documents were approved by the institutional review board. Informed consent was obtained from all participants.

Anthropometric measures

Height and weight were measured without shoes using a wall mounted stadiometer (SECA, Hanover, MD, USA) in centimeters, and on an electronic scale (Scale-tronix, White Plains, NY, USA) in kilograms, respectively. BMI was calculated as weight in kilograms divided by height in meters squared, and participants were classified as underweight (<18.5 kg·m⁻²), normal weight (18.6–24.9 kg·m⁻²), overweight (25.0–29.9 kg·m⁻²), or obese (≥30.0 kg·m⁻²). Skinfolds were obtained with Harpenden calipers (Baty International, West Sussex, U.K.) using established guidelines¹¹. The three site method^{24, 25} was used for both males (chest, abdomen, and thigh) and females (triceps, suprailiac, and thigh). To ensure accuracy, the average of two skinfolds from each site was calculated. A two-step calculation was employed to determine percent body fat. First, body density was calculated using sex- and race-specific regression equations^{24, 25} with age and sum of skinfolds as independent variables. Body density was then used in a second sex- and race-specific regression equation to estimate percent body fat²⁶. Males with skinfold estimated percent body fat ≥25% and females with skinfold estimated percent body fat ≥30% were classified as obese^{27, 28}. Circumference measurements were evaluated with a Gulick spring-loaded tape measure in duplicate. Waist circumference was measured to the nearest 0.1 centimeter laterally at a point midway between the xiphoid process of the sternum and the umbilicus at the minimum circumference. Male and female survivors with a WHtR ≥0.5 were classified as obese²⁹.

DXA

Whole-body DXA scans were performed using a Hologic Model QDR 4500 fan-array scanner (Bedford, MA, USA). Scans were performed and analyzed with APEX 2.3.1 software (Hologic, Inc, Bedford, MA, USA). Participants rested motionless on a radiolucent pad in the supine position with all extremities close to, but not touching the body for the duration of the scan. Whole body fat was calculated as total fat grams divided by total mass grams, and expressed as a percentage³⁰. Males with DXA estimated body fat percentages ≥25 and females with DXA estimated body fat percentages ≥30 were classified as obese^{27, 28}.

Treatment Exposure and Demographics

Information regarding cranial, abdominal, testicular and pelvic radiation, and glucocorticoid exposure, as well as data on disease type, relapse, and bone marrow transplant were abstracted from medical records. Participants treated with craniospinal radiation (N=40) were classified in both the cranial radiation and pelvic/abdominal radiation groups. Demographic information, including race, was obtained from participants via questionnaires.

Statistical Analysis

Descriptive statistics were used to characterize the study population; differences between participants and non-participants were evaluated using Chi-square statistics. Pearson’s correlation coefficients were calculated to examine linear associations between BMI, percent body fat by skinfolds, WHtR, and percent body fat measured via DXA. Unadjusted linear regression was used to evaluate how much of the variability in percent body fat by DXA was explained by BMI, percent body fat by skinfolds, and WHtR. Bland Altman methodology was used to evaluate agreement between percent body fat determined by skinfolds and percent body fat measured by DXA, and to assess potential bias of the skinfold measure across the distribution of DXA determined percent body fat. The ability of BMI, skinfolds, and WHtR to accurately classify obesity in survivors was assessed by calculating the sensitivity, specificity, false negative and false positive rates for these measures using percent body fat determined by DXA as the true measure of obesity. Log-binomial regression models stratified by sex were used to determine which demographic- and treatment-related factors were associated with being misclassified as non-obese (by BMI, skinfolds, and WHtR) when percent body fat cut points determined by DXA were used as the clinical standard. Subgroup analyses were conducted in models limited to ALL survivors. These results did not differ appreciably from the overall results so they are presented as supplemental material. All analyses were performed in SAS version 9.3 (Cary, N.C.).

Results

Participants

Among the 2268 individuals potentially eligible for this study, 1383 (61.0 %) completed a campus visit. Among the 907 non-participants, 505 (22.3%) actively or passively declined participation, 51 (2.2%) were lost to follow-up, 173 (7.6%) chose to complete questionnaires, but not to return for a campus visit, and 156 were pending a scheduled visit (6.9%). Among those who had a campus visit, 1361 completed all study evaluations and were included in these analyses. The mean age at diagnosis of participants was 7.0 years (standard deviation (SD 5.3) and the mean age of follow-up at the SJLIFE visit was 32.4 years (SD 7.7). The mean time since diagnosis was 25.4 years (SD 7.7). Compared to non-participants, a higher proportion of participants were female (50.4% vs. 44.76%, p<0.001), were leukemia survivors (58.05% vs. 47.08%, p<0.001) and were <20 years from diagnosis (58.05 vs. 47.08%, p<0.001; Table 1). In addition, participants were more likely to have been exposed to cranial radiation (45.04% vs. 33.85%, p<0.001) and glucocorticoids (66.2% vs. 52.15%, p<0.001) than non-participants.

Table 1.

Characteristics of Participants and Non-Participants

		Participants		Non-Participants		P-value
		(N = 1361)		(N = 907)
		N	%	N	%
Sex
Female		686	50.40	406	44.76	<0.001
Male		675	49.60	501	55.24
Race
White		1196	87.88	790	87.10	0.58
Other		165	12.12	117	12.90
Age At Diagnosis
0 to 4		584	42.91	368	40.57	0.78
5 to 9		341	25.06	238	26.24
10 to 14		272	19.99	180	19.85
15 to 20		160	11.76	118	13.01
Over 20		4	0.29	3	0.33
Time Since Diagnosis
10 to 19		180	13.23	72	7.94	<0.001
20 to 29		577	42.40	373	41.12
30 to 39		453	33.28	349	38.48
>= 40		151	11.09	113	12.46
Testicular/Ovarian Radiation
Yes		203	14.92	134	14.77	0.93
No		1158	85.08	773	85.23
Cranial Radiation
Yes		613	45.04	307	33.85	<0.001
No		748	54.96	600	66.15
Abdomen/Pelvic Radiation
Yes		275	20.21	180	19.85	0.83
No		1086	79.79	727	80.15
Steroids
Yes		901	66.20	473	52.15	<0.001
No		460	33.80	434	47.85
Relapse
Yes		184	13.52	117	12.90	0.67
No		1177	86.48	790	87.10
Bone Marrow Transplant
Yes		25	1.84	13	1.43	0.46
No		1336	98.16	894	98.57
Diagnosis
Leukemia		790	58.05	427	47.08	<0.001
CNS Tumor		132	9.70	85	9.37
Hodgkin's Lymphoma		121	8.89	91	10.03
Non-Hodgkin’s Lymphoma		64	4.70	43	4.74
Wilm's Tumor		42	3.09	46	5.07
Neuroblastoma		46	3.38	48	5.29
Soft tissue sarcoma		42	3.09	41	4.52
Bone Cancer		33	2.42	63	6.95
Retinoblastoma		60	4.41	38	4.19
Other		31	2.28	25	2.76

CNS, Central nervous system

Anthropometry and body composition

Anthropometric measures are reported in Table 2 (and Supplemental Table 1) by sex. Mean BMIs were in the overweight range for both males and females. Mean percent body fat determined by both skinfolds and DXA were in the obese range for females, while only mean percent body fat determined with DXA was in the obese range among males survivors. Mean WHtR was high (>0.5) for both males and females. Approximately 36% of males were classified as obese by the BMI, however, nearly two-thirds of males were classified as obese by DXA body fat and WHtR. Similarly, about 40% of females were classified as obese by BMI, but 85% were classified as obese by DXA, 62% obese by skinfolds, and 60% obese by WHtR.

Table 2.

Anthropometric measures by sex

	Males		Females
	Mean (SD)	% Obese* (95% CI)	Mean (SD)	% Obese* (95% CI)
Height (cm)	174.3 (8.5)	--	160.9 (7.5)	--
Weight (kg)	85.9 (19.3)	--	74.9 (19.8)	--
BMI (kg/m2)	28.2 (5.5)	35.7 (32.1–39.4)	28.9 (7.4)	39.7 (36.0–43.4)
% body fat by DXA	26.9 (7.4)	63.1 (59.3–66.8)	38.4 (7.7)	84.8 (81.9–87.4)
% body fat by skinfolds	23.3 (7.7)	42.5 (38.8–46.3)	32.3 (8.1)	62.1 (58.4–65.7)
WHtR	0.53 (0.08)	64.9 (61.2–68.5)	0.54 (0.10)	60.0 (56.3–63.7)

BMI, Body mass index; DXA, dual energy x-ray absorptiometry; WHtR, Waist to height ratio; CI, Confidence Interval

^*Obesity classifications are as follows: BMI ≥30.0 kg•m-2 (males and females); Body fat by skinfold and DXA males ≥25%, females ≥30%; WHtR ≥0.5 (males and females).

Agreement between BMI, skinfolds, and WHtR with DEXA

Pearson correlations between percent body fat by skinfolds and by DXA were high for both males (R=0.83) and females (R=0.84), and were slightly higher than correlations between BMI and percent body fat determined by DXA (males, R=0.74; females, R=0.81) and WHtR and percent body fat by DXA (males, R=0.76; females, R=0.77). Figures 1a–1c (and Supplemental Figures 1a–1c) shows the linear association between BMI, skinfolds, and WHtR and DXA for males and females. Percent body fat by skinfolds had the largest R² value and explained 70% of percent body fat by DXA variation among males and 71% of percent body fat variation among females. BMI explained 55% of DXA body fat variation for males and 66% of DXA body fat variation for females, while WHtR explained 58% of DXA body fat variation for males and 59% of DXA body fat variation for females. Bland-Altman plots (Figures 2a and 2b; and Supplemental Figures 2a and 2b) show that percent body fat determined by skinfolds underestimates percent body fat determined by DXA for both males (−3.59 ± 4.35) and females (−6.05 ± 4.44) and that this bias appears to be more variable, particularly among males.

Figure 1.

Linear agreement between percent body fat by (A) Body mass index (BMI) and dual energy x-ray absorptiometry (DXA), (B) skinfolds and DXA, and (C) Waist to height ratio (WHtR) and DXA

Figure 2.

Bland-Altman plots for the relationship between percent body fat determined by DXA and skinfolds

Accuracy of obesity classification by BMI, skinfolds, and WHtR

WHtR was the most sensitive measure among male survivors with a false negative rate of 12.9% (Table 3 and Supplemental Table 2). Among females, false negative rates were comparable between skinfolds (27.3%) and WHtR (29.6%). The highest false negative rates were observed for BMI at 46.4% for males and 53.1% for females. Overall, specificity was high for all models considered (>90%), with exception of WHtR which had a specificity of 71.0% among males.

Table 3.

Sensitivity and specificity of obesity classifications for BMI, skinfolds, and WHtR

Males	Obese by DXAa	Non-obese by DXAa
Obese by BMIb	53.6% (Sensitivity)	6.5% (False positive)
Non-obese by BMIb	46.4% (False negative)	93.6% (Specificity)
Obese by skinfoldsc	65.5% (Sensitivity)	3.2% (False positive)
Non-obese by skinfoldsc	34.5% (False negative)	96.8% (Specificity)
Obese by WHtRd	87.1% (Sensitivity)	29.0% (False Positive)
Non-obese by WHtRd	12.9% (False Negative)	71.0% (Specificity)
Females	Obese by DXAa	Non-obese by DXAa
Obese by BMIb	46.9% (Sensitivity)	0.0% (False positive)
Non-obese by BMIb	53.1% (False negative)	100.0% (Specificity)
Obese by skinfoldsc	72.7% (Sensitivity)	2.9% (False positive)
Non-obese by skinfoldsc	27.3% (False negative)	97.1% (Specificity)
Obese by WHtRd	70.4% (Sensitivity)	4.8% (False Positive)
Non-obese by WHtRd	29.6% (False Negative)	95.2% (Specificity)

^aMales with DXA estimated body fat percentages ≥25 and females with DXA estimated body fat percentages ≥30 were classified as obese.

^bParticipants with BMI ≥30 kg·m² were considered obese.

^cMales with skinfold estimated percent body fat ≥25 and females with skinfold estimated percent body fat ≥30 were classified as obese.

^dParticipants with WHtR ≥0.5 were considered obese.

Table 4 (and Supplemental Table 3) shows sex-specific log binomial regression models used to identify factors associated with survivors being misclassified as non-obese by BMI, skinfolds and WHtR. In models limited to males, survivors exposed to abdominal or pelvic radiation were more likely to be misclassified as non-obese when using BMI (RR: 1.57, 95% CI 1.25–1.95), skinfolds (RR: 1.50, 95% CI 1.11–1.98), or WHtR (RR: 2.35, 95% CI 1.36–3.92). Among females, exposure to abdominal or pelvic radiation was also associated with an increased risk of being misclassified as non-obese when using WHtR (RR: 1.70 95% CI 1.27–2.22). Female survivors who received cranial radiation were at a decreased risk of being misclassified as non-obese when using BMI (RR: 0.80, 95% CI 0.68–0.94), skinfolds (RR: 0.77, 95% CI 0.58–1.01), or WHtR (RR: 0.49, 95% CI=0.37–0.64). Furthermore, females of non-white ethnic descent were less likely to be misclassified as non-obese when using BMI (p=0.02) or WHtR (p=0.02) when compared to females of white descent.

Table 4.

Relative Risks for non-obese misclassification using BMI, skinfolds, and WHtR

	BMI		Skinfolds		WHtR
Effect	RR (95% CI)	P Value	RR (95% CI)	P Value	RR (95% CI)	P Value
	Male Models
Race (Non-white vs. White)	1.02 (0.72–1.34)	0.89	1.21 (0.82–1.67)	0.31	1.10 (0.48–2.12)	0.80
Cranial Radiation (Yes vs. No)	1.01 (0.82–1.24)	0.94	0.92 (0.70–1.20)	0.54	0.89 (0.53–1.50)	0.66
Abdomen/Pelvic Radiation (Yes vs. No)	1.57 (1.25–1.95)	<0.001	1.50 (1.11–1.98)	0.010	2.35 (1.36–3.92)	0.003
	Female Models
Race (Non-white vs. White)	0.76 (0.56–0.96)	0.02	0.70 (0.42–1.07)	0.11	0.63 (0.39–0.92)	0.02
Cranial Radiation (Yes vs. No)	0.80 (0.68–0.94)	0.007	0.77 (0.58–1.01)	0.06	0.49 (0.37–0.64)	<0.001
Abdomen/Pelvic Radiation (Yes vs. No)	1.13 (0.93–1.34)	0.20	1.03 (0.71–1.42)	0.89	1.70 (1.27–2.22)	0.001

All models were adjusted for current age and age at diagnosis

Discussion

The results of this study indicate that a three site skinfold method measurement of percent body fat is an acceptable measure of adiposity for survivors of childhood cancer. Although skinfolds underestimated percent body fat among CCS when compared to DXA, particularly among female survivors, it had better sensitivity than the typical clinical measure, BMI, when used to classify cancer survivors as obese or not. Similarly, WHtR was a more sensitive measure of obesity than BMI for both male and female CCS, and given the ease of which this test can be administered, would be a more appropriate measure to classify obesity than BMI in this population. Additionally, we demonstrated that male survivors treated with abdominal or pelvic radiation had an increased risk of being incorrectly classified as non-obese when BMI, skinfolds, or WHtR were utilized.

Our findings of a high correlation between skinfolds and percent body fat are comparable to those previously reported in healthy populations^{18, 31}. Skinfolds underestimated body fat percentage in males by approximately 3.5% which is similar to what has been previously reported in healthy men³². Our study underestimated percent body fat using skinfolds in women by approximately 6%, which is a larger error than previously reported in healthy populations³³. Although correlations were higher for skinfolds and DXA than BMI and DXA and WHtR and DXA, this could be because skinfolds and DXA measure the same outcome, percent body fat, while BMI and WHtR are ratios. The difference in the correlations could be due to the difference in the outcome measures alone. Overall, findings from our study indicate that skinfold measurement is an acceptable method for estimating percent body fat in long-term survivors of childhood cancer. Nevertheless, clinicians should be cautious when interpreting skinfold-derived estimates among females, and any survivor with an estimated body fat percent near the cut-off between obese and non-obese.

In the current study, we assessed the ability of skinfolds, BMI and WHtR to accurately classify obesity in survivors of childhood cancer. We observed that skinfolds were able to accurately classify more than two-thirds of survivors as obese when compared to assessment by DXA. In contrast, nearly 47% of males and 53% of females were misclassified as non-obese by BMI. This finding is consistent with a study of body composition among CCS by Blijdorp et al (2012), in which 42% of males and 65% of females were misclassified by BMI when compared to values obtained from DXA³⁴. Our study expands on these analyses, as we included additional measures of obesity. A higher number of participants were considered obese by body fat percentage standards for DXA and skinfolds than BMI standards. This is similar to previous findings among healthy individuals when using plethysmography to determine body composition³⁵. We also found that WHtR was the most sensitive measure of obesity status among male survivors, and had similar sensitivity to skinfolds among female survivors. In a group of private practice patients, only 22% of men and 48% of women were misclassified as non-obese by BMI (classified as obese by DXA)³⁶. This is different than what we found, especially for male CCS. Our results suggest that the use of BMI to determine body composition during health examinations is likely to result in a considerable number of CCS with high adiposity being misclassified as non-obese. Misclassification may result in these survivors not receiving appropriate guidance on diet and physical activity, or counseling regarding their future risk of obesity-related health outcomes. If DXA is unavailable, clinicians should consider using WHtR or skinfold measures to evaluate body composition in CCS. It should be noted, 29% of males were considered obese by WHtR but were non-obese by DXA (false positive). Clinicians should exercise caution when using WHtR in male CCS to classify obesity, as short stature and high lean muscle mass may provide misleading information.

In the current study, we observed that male survivors previously exposed to abdominal/pelvic radiation, and who were classified as obese by DXA, were more likely to be misclassified as non-obese using BMI, skinfolds, or WHtR. It has been suggested that survivors treated with abdominal radiation can suffer from soft tissue hypoplasia, giving them an inaccurate “thin” appearance, and therefore not an accurate portrayal of actual abdominal obesity³⁷. In situations where DXA is not readily available, medical professionals need to be cognizant of the limitations of interpreting BMI, skinfolds, and WHtR among male survivors exposed to abdominal/pelvic radiation. In addition, females who received cranial radiation were less likely to be misclassified as non-obese by BMI, skinfolds, or WHtR, which has been previously reported^{2, 4}.

Our ability to generalize findings from this study may have been limited by the high proportion of participants with a prior diagnosis of ALL. This is important in view of previous studies that have shown ALL survivors tend to be overweight or obese^1–4, while survivors of certain types of solid tumors are more likely to be underweight². Separate analysis of the ALL subgroup yielded similar results to the analysis of the overall group of survivors. As the primary aim of this study was to assess the performance of BMI, skinfolds, and WHtR against DXA in determining obesity, the potential contribution of conditions and health behaviors known to affect body composition, such as growth hormone deficiency, physical activity levels, and dietary intake, were not taken into consideration in these analyses.

In light of the high risk of chronic disease among survivors of childhood cancer, it is critical that survivors receive appropriate counseling on weight management from their healthcare providers. Our findings demonstrate that BMI may not be an accurate measure of adiposity among CCS. Although DXA may not always be available or affordable, skinfolds and WHtR are a feasible alternative for assessing body composition. Medical professionals should be aware of the risk of misclassifying obesity among CCS, especially those who were exposed to abdominal or pelvic radiation therapy, and consider using skinfolds or WHtR as easily administered and inexpensive methods to determine body composition.