Assessment of Segmental Arm Soft Tissue Composition in Breast Cancer-Related Lymphedema: A Pilot Study Using Dual Energy X-ray Absorptiometry and Bioimpedance Spectroscopy

Abstract

Background:

Changes in arm soft tissue composition, especially increased adipose tissue, has been found in advanced, non-pitting breast cancer-related lymphedema (BCRL). The aim of this study was to examine whether these changes were localized to any particular region of the arm and whether they occurred in lymphedema which still pitted to pressure. Secondary aims were to explore relationships between arm segment volumes, bioimpedance spectroscopy (BIS) measurements of extracellular fluid (ECF), and dual-energy X-ray absorptiometry (DXA) measurements of tissue composition.

Methods and Results:

Nine women with unilateral BCRL participated. The dominant arm was affected in 4 women, and all presented with lymphedema that pitted to pressure. Arm volume was calculated from circumferences by the truncated cone method, ECF was determined with BIS and fat and lean tissue content measured by DXA. BIS and DXA measurements for women with lymphedema were made of the whole arm and also of four 10 cm-segments measured from the ulnar styloid at the wrist. Whole arm DXA data were compared to those of 45 women of similar age and body mass index without lymphedema. All women with lymphedema had a significantly larger absolute fat mass in their affected arm compared to their unaffected arm, (median difference between arms 146.9 g). The forearm segment 10 – 20 cm proximal to the wrist had the highest median inter-limb fat difference of all four arm segments.

Conclusions:

The soft tissue composition changes associated with BCRL may occur in the presence of pitting and predominantly affect the proximal forearm.

Introduction

Arm lymphedema remains a problem that affects the quality of life of many women after the treatment of breast cancer, despite advances in surgery and radiotherapy.^1–3 The swelling characteristic of breast cancer-related lymphedema (BCRL) may affect the whole arm or be localized to a region such as the forearm.^4,5 Recent research suggests that specific regions of the arm may be affected first and herald the onset of lymphedema.⁶

As lymphedema progresses, it alters the soft tissue composition of the affected limb.⁷ Changes to extracellular fluid can exist for several months or years before swelling becomes visible.⁸ Swelling initially subsides with limb elevation. However, as the condition progresses, elevation ceases to be effective and swelling indents, or pits, with the application of pressure.⁹ Further progression of lymphedema results in increased fibrotic and adipose tissue, and pitting no longer occurs.^10,11 This results in lymphedema becoming increasingly resistance to treatment.¹²

The theories proposed for increased adiposity as a result of lymphatic dysfunction include a decreased clearance of lipids due to alterations in blood flow and lymphatic drainage¹³ and increased fat deposition as a result of a chronic inflammatory response.¹⁴ It is reasonable to surmise that the regions of the arm that develop lymphedema first would be exposed to these physiological changes for longer and may therefore demonstrate greater soft tissue compositional changes than surrounding regions. The majority of studies reporting findings of increased adipose tissue have been on women with a large volume difference between their affected and nonaffected arm, on the full arm and in late stage II non-pitting lymphedema.^8,10,11,15 Recent research has shown changes in tissue composition to be present in mild to moderate lymphedema¹⁶ and some studies suggest that these changes may begin to occur earlier than previously thought, even prior to visible swelling.¹⁷

Traditional methods of assessment of limb circumference and volume are unable to give any information on tissue composition of limbs affected by lymphedema.¹⁸ Imaging techniques such as magnetic resonance imaging (MRI), computed tomography (CT), ultrasound, and dual-energy X-ray absorptiometry (DXA) are used for this purpose.^16,19

Each method has its own advantages and disadvantages but for most practical purposes the advantages of using DXA for assessing tissue composition are relatively high accuracy and reproducibility but with much lower radiation exposure compared to CT.^18,20–22 In addition DXA instruments are widely available and the cost per patient scan is low. DXA provides body composition information as fat and fat-free mass (FFM). FFM is comprised of lean tissue mass (LTM) and bone mineral content (BMC).²¹ DXA has been previously used successfully to study soft tissue composition and volume of arms affected by BCRL.^8,16,18,23

The aim of this study was to describe the whole and segmental arm soft tissue composition in women with unilateral lymphedema to determine if changes were localized to any particular region of the arm. Secondary aims were to determine if these were reflected in segmental bioimpedance spectroscopy (BIS) or volume measures.

Materials and Methods

Ethical approval

Ethics approval was granted by the Sydney University Human Research Ethics Committee and the Sydney University Radiation Safety Committee approved the use of DXA. All participants provided written informed consent prior to participating in this study.

Participants

Nine women with BCRL were recruited from participation in other studies. They were eligible to participate if they had previously received a diagnosis of unilateral arm lymphedema from a health professional and were able to understand English. Exclusion criteria were a pacemaker or other in-built stimulator, or pregnancy, as this would contraindicate the use of BIS and DXA. The dominant arm was affected in four women and the nondominant arm in five women. Lymphedema pitted with pressure in all nine women who, using the International Society of Lymphedema classification,⁹ were classified as having early stage II lymphedema. Participant characteristics are outlined in Table 1.

Table 1.

Participant Characteristics Median (Inter-quartile Range (IQR))

	Lymphedema ( n=9)	No Lymphedema ( n=45)
Age (years)^∧	65 (63.5–70.5)	57 (55–62.5)
BMI (kg/m²)^∧	26.6 (22.5–30.8)	27.9 (24.3–31.8)
Dominant arm (R:L)	7:2	43:2
Dominant arm affected (Dom: Non-Dom)	4:5	–
Lymphedema duration (years)^∧	3.0 (1.5–8.8)	–
Inter-limb volume difference (ml)^∧^*	345 (144–411)	–
Inter-limb difference in total mass (g)^∧^*	375.9 (266–519.0)	158.0 (60.5–268.0)
Inter-limb difference in fat (g)^∧^*	146.9 (46–368.7)	−52.0 (−188.0–7.0)
Inter-limb difference in fat-free mass (lean tissue+bone mineral content) (g)^∧^*	214.7 (184.9–473.0)	237.0 (138.5–344.0)
Inter-limb difference in lean tissue (g)^∧^*	218.7 (183.2–469.5)	216.0 (138.5–320.5)
ECF (bioimpedance) ratio^∧†	1.12 (1.05–1.27)	–

∧

Median and IQR shown; ^*Absolute inter-limb difference determined by ‘Affected – unaffected’ for women with lymphedema and ‘dominant – nondominant’ for women without lymphedema; †ECF ratio calculated as ‘unaffected/affected’ impedance ratio.

Ethics approval was also granted to contact women who had previously had a DXA scans using the same instrument and DXA operator to request permission to use their scans to provide normative data. The clinic database was audited to source women with comparable characteristics but without a history of either breast cancer or lymphedema, with forty-five women consenting to the use of their DXA data. Their characteristics are presented in Table 1. The median age for women without lymphedema was 57.0 years (inter quartile range (IQR) 55.0–62.5 years), which was significantly younger than for women with lymphedema (65.0 years, IQR 63.5–70.5 years; p=0.001). The median body mass index (BMI) of women without lymphedema was not significantly different to women with lymphedema (27.9 kg/m²; IQR: 24.3–31.8 versus 26.6 kg/m²; IQR: 22.5–30.8; p=0.451).

Measurement procedures

BCRL participants

All data were collected at a single session by an experienced lymphedema practitioner. Surgical and medical history, height, weight, current medication, and arm dominance were recorded. Lymphedema stage was determined by assessing reduction of swelling with elevation, pitting with pressure and the presence of fibrotic changes.

Arm volume measurement

Arm volumes of both the affected and unaffected arm were calculated from limb circumferences. Circumference measurements were made with a Jobst nonstretch soft tape measure commencing at the mid-point of the ulnar styloid as the “0 cm” reference mark, then at 10 cm, 20 cm, 30 cm, and 40 cm from the ulnar styloid, following a protocol previously described.²⁴ The volume of each of the four arm segments was calculated from circumferences using the formula for a truncated cone. Arm segments were designated as Segment A (0–10 cm), Segment B (10–20 cm), Segment C (20–30 cm), and Segment D (30–40 cm).^25,26 The volume of the four arm segments was summated to establish total limb volume from the ulna styloid (0 cm) to 40 cm proximally along the arm.

ECF measurements

The participant rested in supine on an examination couch with their arms slightly abducted for measurement of whole arm (wrist to axilla) ECF by BIS. These measurements were performed according to a protocol previously described,²⁷ and the impedances of the affected and unaffected arms determined using an SFB7 impedance spectrometer (ImpediMed Ltd., Brisbane, Australia). The standardized protocol for the whole arm was then modified to measure individual 10 cm arm segments for both the affected and unaffected arm. The two measurement electrodes were positioned at 10 cm intervals along the dorsal surface of each arm at positions 10 cm, 20 cm, 30 cm, and 40 cm for the ulnar styloid, as previously described.²⁵ Data were processed according to the manufacturer's software (Imp RSM version 1.12.0 and Bioimp version 2.25, ImpediMed Ltd., Brisbane, Australia). From these data the inter-limb impedance ratio, corresponding to the ratio of ECF, was calculated for the whole arm and for each of the four arm segments.

DXA measurements

Participants with BCRL

The participant was positioned supine, with arms slightly abducted, on the bed of the DXA scanner (Hologic Explorer, Hologic, Massachusetts, USA, software version 12.4) according to the manufacturer's instructions for a whole body scan. Care was taken to minimize contact between the participant's arms and trunk in this position so that regions of interest (ROI) could be clearly defined for each arm. In order to define ROI for the four 10 cm arm segments (A, B, C, and D) radio-opaque markers were placed at 10 cm intervals along each arm. These commenced at the ulna styloid (0 cm), and were positioned at 10 cm intervals along the dorsal surface of each arm at positions 10 cm, 20 cm, 30 cm, and 40 cm for the ulnar styloid in the same location as the BIS measurement electrodes.

All scans were performed and analyzed by the same DXA specialist using the same equipment and version of analysis software. For purposes of comparison with data from the group of women without lymphedema, the ROI was defined as the whole arm to the axilla, including the hand. Fat, lean tissue, and bone mineral content for the affected and nonaffected arm were obtained. Delineation of the arm segment ROI, using the radio-opaque markers, was determined by the same DXA specialist and confirmed by the investigator who had performed the segmental volume and BIS measures. High intra-rater reliability using customized ROI analysis has been previously established.¹⁶ For each segment, fat mass and FFM were determined.

Participants without lymphedema

The DXA scans of consenting women without lymphedema, obtained by the same machine, were assessed by the same DXA specialist, and analyzed using the same software as for participants with BCRL. Only data for the identical whole arm ROI were available for this group. These images were re-analyzed and the region of interest defined as the whole arm, including the hand, to the axilla, as for the BCRL participants.

Data analysis

The median and inter-quartile ranges were determined for demographic data and measures of limb volume ratio, ECF ratio, and composition. As preliminary analysis showed that the data had a nonparametric distribution, a nonparametric related-samples Wilcoxon signed rank test was used to determine if there were differences in the fat and lean tissue contents between the affected and nonaffected arms in the whole arm measures of women with lymphedema. The same test was also used to analyze if there were significant differences in tissue composition between the dominant and nondominant arms of the women without lymphedema.

An independent sample Mann-Whitney U test was used to determine if there was a difference between the fat and lean tissue in the arms of women with and without lymphedema. For purposes of comparison for women without lymphedema the dominant arm was deemed “affected.” This analysis was of the inter-limb ratios of fat and lean tissue between the two groups as well as absolute whole arm fat and lean tissue (in grams). Inter-limb ratios were calculated by affected/nonaffected arm for women with lymphedema and dominant/nondominant arm for women without lymphedema. The use of inter-limb ratios allowed the unaffected arm to act as an internal control, accounting for individual variations in the body composition of the women in each group. Data were analyzed with IBM SPSS Statistics software version 19 and for all tests significance was set at p<0.05.

Results

The volume of the affected arm was significantly greater than the nonaffected arm of women with lymphedema (volume difference 345 mL; 144–411 mL, p<0.001; Table 1). Whilst the median exceeded the recently established diagnostic cut-off identified by Dylke et al.,²⁶ the IQR shows that one participant with the nondominant arm affected and two participants with dominant arm affected had arm volume differences below this threshold, even with the more liberal threshold based on 2SD.²⁶ The volume difference was mirrored by a significantly higher total limb mass for the arm affected by lymphedema with an inter-limb difference of 375.9 g (266.0 to 519.0 g, p=0.008; Table 1).

The volume of each arm segment (Table 2) was also greater in the affected arm than the unaffected arm. All women except for one with dominant arm affected and one with nondominant arm affected had a least one arm segment above established thresholds.²⁶ The distal arm (Segment A and Segment B) was most affected (Table 2).

Table 2.

Arm Segment Measures of Women with BCRL (n=9). Median (IQR) are shown

Arm Segment	A (0–10 cm)	B (10–20 cm)	C (20–30 cm)	D (30–40 cm)
Inter-limb volume difference (mL)^∧^*	51.9 (20.0–78.1)	75.7 (36.4–121.5)	80.9 (15.8–148.1)	90.3 (37.3–125.9)
Number of segments≥: <threshold #	5:4	6:3	4:5	2:7
ECF ratio^∧†	1.219 (1.084–1.593)	1.124 (1.048–1.416)	1.113 (1.048–1.584)	1.088 (0.997–1.171)
Number of segments≥: <3SD threshold¤	5:4	6:3	4:5	2:7
Number of segments≥: <2SD threshold¤	7:2	6:3	6:3	4:5
Inter-limb fat difference (g)^∧^*	9.8 (−21.9–52.2)	50.4 (30.3–99.3)	33.6 (13.6–49.4)	31.7 (8.4–126.1)
Inter-limb fat free mass difference (g)^∧^*	68.4 (−4.5–83.5)	73.83 (−3.01–105.3)	75.6 (40.2–96.8)	24.4 (5.85–83.7)

∧

Median and IQR shown; ^*Absolute inter-limb difference determined as ‘Affected – unaffected’ ; †ECF ratio calculated as ‘unaffected/affected’ impedance ratio; # Arm segment volume thresholds Dylke et al.; ²⁶ ¤ BIS segment thresholds Svensson et al.²⁸

Median whole arm ECF ratio for the whole arm was 1.12 (1.05 to 1.27; Table 1). This was low, considering the established diagnostic thresholds of 1.066 for nondominant arm affected and 1.139 for dominant arm affected.²⁹ Three of the five women with nondominant arm affected and two out of four with dominant arm affected exceeded these thresholds.

ECF ratios calculated for arm segments (Table 2) show the greatest difference in the distal arm, at Segments A and B. Analysis with BIS segment thresholds²⁸ showed that all women, except for one with the dominant arm affected and one with nondominant arm affected, had at least one arm segment above established BIS thresholds based on 3SD above normative data.²⁸ These were the same two participants identified as not having any volume segments above established volume thresholds.²⁶ When a more liberal 2SD was used, only the woman with her dominant arm affected did not have any BIS segments above threshold. The distal arm had the highest number of segments above BIS thresholds (Table 2). Further analysis for segments A and B showed that volume thresholds and BIS thresholds based on 3SD identified not only the same number, but identical segments, as affected. In segments C and D, both BIS and volume thresholds identified the same number of segments as affected (four and two, respectively) but these were only identical for two C segments and one D segment.

All of the women with lymphedema had significantly more absolute fat mass in their affected arm compared to their unaffected arm (Fig. 1A; inter-limb difference: 146.9 g (46.0–368.7 g; p=0.008; Table 1). Women in whom their non-dominant arm was affected had a greater difference in fat mass between the affected and unaffected arm (175.8 g; 38.1–466.7 g), than women in whom their dominant arm was affected (99.3 g; 46.5–314.4 g).

FIG. 1.

Inter-limb difference between fat (A) and lean tissue (B). For women with lymphedema, the inter-limb difference is affected arm - unaffected arm. Nondominant arm affected is represented by white circles, dominant arm affected by black circles. The ECF ratio measure of the woman with the highest and lowest inter-limb difference of fat and lean tissue is indicated on the graphs. For women without lymphedema represented in graphs on the right, inter-limb difference is determined by nondominant arm - dominant arm for the white boxplot, and dominant arm – nondominant arm for the black boxplot. Group mean and 95% confidence interval are shown.

In eight of the nine women with lymphedema, lean tissue mass was also greater in the affected arm, and overall, the difference was significant (218.7 g; 183.2–469.5 g; p=0.038; Table 1; Fig. 1B). Women in whom the dominant arm was affected had a greater difference in lean mass between the affected and unaffected arm (413.5 g; 246.6–720.7 g) than women in whom their nondominant arm was affected (191.5 g; - 47.2–329.3 g).

Arm segment measures revealed the largest difference in fat mass between the affected and unaffected arm was on either side of the elbow [i.e., Segments B and C (Table 2)]. Similarly when expressed as a ratio (affected arm segment/unaffected arm segment) the segment on either side of the elbow had increased fat mass (Fig. 2). These two arm segments also had the largest difference in fat-free mass between affected and unaffected limbs (Table 2).

FIG. 2.

Ratio of affected to unaffected arm segment fat (a) and fat free mass (b) in each of the four segments of the arms of women affected by lymphedema. Data for women with their dominant arm affected by lymphedema is represented by black circles and nondominant by white circles.

Comparison of the two groups by an independent sample Mann-Whitney U test showed that there was no significant difference between the median affected whole arm fat tissue for women with lymphedema and women without lymphedema (p=0.273) There was also no significant difference between median affected whole arm lean tissue measures between the group of women with lymphedema and women without lymphedema (p=0.715). When whole arm fat and lean tissue ratios (affected/nonaffected for women with lymphedema and dominant/nondominant for women without lymphedema) were compared by an independent sample Mann-Whitney U test there was a significant difference in fat ratio (p<0.0001) but not lean tissue ratio (p=0.092).

Discussion

Increased adipose tissue has been previously reported in the affected arm of women with BCRL.^8,11,16 We took this research a step further to examine segmental distribution of adipose tissue and to determine that the region around the elbow is most affected. This localized adipose tissue could be explained by findings that this region is affected early in the development of lymphedema.^4,5,30 A localized increase in extracellular fluid may initiate corresponding changes in soft tissue composition.

Women who presented with pitting lymphedema and relatively small volume changes in their affected arms were found to have soft tissue changes, indicating that increased adiposity may affect even mild lymphedema. This finding is supported by studies showing no correlation between lymphedema duration and excess adipose tissue^8,16 and postmortem adipose tissue changes even in the absence of visible swelling.¹⁷ The clinical implication is that, even in mild lymphedema, the presence of adipose tissue may hinder conservative efforts to reduce limb volume. This study is limited by the small sample size of women affected by lymphedema. However, it was intended as an exploratory study to determine methods of using DXA for arm segment measures for comparison with BIS segmental measures explored in previous research.²⁵ Despite small numbers, segmental variations were identified in soft tissue composition in BCRL. This warrants further exploration and comparison to segmental measures of women without lymphedema. Segmental DXA could detect soft tissue changes in cases of poor response to conservative treatment that might otherwise be masked by whole arm measurements.

Three women in the study did not meet the diagnostic criteria for lymphedema on the basis of established whole arm inter-limb volume thresholds.²⁶ In addition, two women were amongst the four also below the ECF diagnostic thresholds.²⁹ This was despite all women noted as having pitting lymphedema by the lymphedema practitioner who performed the measurements. A likely explanation is that pitting was present in the hand, but not the arm. At the time this research was performed, techniques for measuring hand volume were limited and assessment of hand ECF was not available. Methods outlined in recent studies^31–33 could be incorporated in future research to enable DXA, BIS, and volume comparison of the whole upper limb, including the hand.

Conclusions

These preliminary findings support the need for the early detection of BCRL to prevent chronic, irreversible tissue changes, and suggest that analysis of segmental limb composition may have an important role to play.

Footnotes

Acknowledgments

S. Czerniec was supported by a National Breast Cancer Foundation (Australia) doctoral scholarship and by the RT Hall Foundation. SLK was supported by a National Breast Cancer Foundation (Australia) career research fellowship.

Author contributions: S.A. Czerniec contributed to study conception and design, data collection, analysis, and interpretation of data and manuscript preparation. L.C. Ward contributed to study conception and design, analysis, and interpretation of data, and manuscript preparation. J Meerkin contributed to data collection and analysis. S.L. Kilbreath contributed to study conception and design, analysis, and interpretation of data and manuscript preparation.

Author Disclosure Statement

Authors S.Czerniec, J Meerkin, and S Kilbreath have no competing financial interests. Dr. Ward consults to ImpediMed Ltd. ImpediMed had no involvement in the design, undertaking, or manuscript preparation of this study.

References

Hayes

, Di Sipio

, Rye

, Lopez

, Saunders

, Pyke

, Bashford

, Battistutta

, Newman

. Prevalence and prognostic significance of secondary lymphedema following breast cancer. Lymphat Res Biol, 2011; 9:135–141.

Norman

, Localio

, Potashnik

, Simoes Torpey

, Kallan

, Weber

, Miller

, DeMichele

, Solin

. Lymphedema in breast cancer surviviors: Incidence, degree, time course, treatment, and symptoms. J Clin Oncol, 2009; 27:390–397.

DiSipio

, Rye

, Newman

, Hayes

. Incidence of unilateral arm lymphoedema after breast cancer: A systematic review and meta-analysis. Lancet Oncol, 2013; 14:500–515.

Stanton

, Mellor

, Cook

, Svensson

, Peters

, Levick

, Mortimer

. Impairment of lymph drainage in subfascial compartment of forearm in breast cancer-related lymphedema. Lymphat Res Biol, 2003; 1:121–132.

Stanton

, Modi

, Mellor

, Levick

, Mortimer

. Recent advances in breast cancer-related lymphedema of the arm: Lymphatic pump failure and predisposing factors. Lymphat Res Biol, 2009; 7:29–45.

Shah

, Vicini

. Breast cancer-related arm lymphedema: Incidence rates, diagnostic techniques, optimal management and risk reduction strategies. Intl J Radiat Oncol Biol Physics, 2011; 81:907–914.

Rockson

. Update on the biology and treatment of lymphedema. Curr Treatment Options Cardiovasc Med, 2012; 14:184–192.

Brorson

, Ohlin

, Olsson

, Karlsson

. Breast cancer-related chronic arm lymphedema is associated with excess adipose and muscle tissue. Lymphat Res Biol, 2009; 7:3–10.

The International Society of Lymphology. The diagnosis and treatment of peripheral lymphedema. 2009 Consensus document of the International Society of Lymphology. Lymphology, 2009; 42:51–60.

10.

Brorson

. Liposuction gives complete reduction of chronic large arm lymphedema after breast cancer. Acta Oncol, 2000; 39:407–418.

11.

Brorson

, Ohlin

, Olsson

, Nilsson

. Adipose tissue dominates chronic arm lymphoedema following breast cancer: An analysis using volume rendered CT images. Lymphat Res Biol, 2006; 4:199–209.

12.

Bar Ad

, Cheville

, Solin

, Dutta

, Both

, Harris

EER

. Time course of mild lymphedema after breast conservation treatment for early-stage breast cancer. Intl J Radiat Oncol Biol Physics, 2010; 76:85–90.

13.

Ryan

. Lymphatics and adipose tissue. Clin Dermatol, 1995; 13:493–498.

14.

Gaffney

, Casley-Smith

. Excess plasma proteins as a cause of chronic inflamation and lymphoedema: Biochemical estimations. J Pathol, 1981; 133:229–242.

15.

Brorson

, Svensson

. Liposuction combined with controlled compression therapy reduces arm lymphedema more effectively than controlled compression therapy. J Am Soc Plastic Reconstruct Surg, 1998; 102:1058–1067.

16.

Dylke

, Ward

, Meerkin

, Nery

, Kilbreath

. Tissue composition changes and secondary lymphedema. Lymphat Res Biol, 2013; 11:211–218.

17.

Tassenoy

, Vermeiren

, van der Veen

, Stadnick

, De Ridder

, Peeters

, Van Schuerbeek

, Lamote

, Lievens

. Demonstration of tissue alterations by ultrasonography, magnetic resonance imaging and spectroscopy and histology in breast cancer patients without lymphedema after axillary node dissection. Lymphology, 2006; 39:118–126.

18.

Gjorup

, Zerahn

, Hendel

. Assessment of volume measurement of breast cancer-related lymphedema by three methods: Circumference measurement, water displacement and dual energy x-ray absorptiometry. Lymphat Res Biol, 2010; 8:111–119.

19.

Stanton

, Badger

, Sitzia

, Stanton

, Badger

, Sitzia

. Non-invasive assessment of the lymphedematous limb. Lymphology, 2000; 33:122–135.

20.

Svendsen

, Haarbo

, Hassager

, Christiansen

. Accuracy of measurements of body composition by dual-energy x-ray absorptiometry in vivo. Am J Nutrit, 1993; 57:605–608.

21.

Leahy

, O'Neill

, Sohun

, Jakeman

. A comparison of dual energy X-ray absorptiometry and bioelectrical impedance analysis to measure total and segmental body composition in health young adults. Eur J Appl Physiol, 2012; 112:589–595.

22.

Lee

, Gallagher

. Assessment methods of human body composition. Curr Opin Clin Nutrit Metab Care, 2008; 11:566–572.

23.

Newman

, Rosenthall

, Towers

, Hodgson

, Shay

, Tidhar

, Vigano

, Kilgour

. Determining the precision of dual energy x-ray absorptiometry and bioelectrical impedance spectroscopy in the assessment of breast cancer-related lymphedema. Lymphat Res Biol, 2013; 11:104–109.

24.

Czerniec

, Ward

, Refshauge

, Beith

, Lee

, York

, Kilbreath

. Assessment of breast cancer-related arm lymphedema. Comparison of physical measurement methods and self-report. Cancer Investig, 2010; 28:54–62.

25.

Czerniec

, Ward

, Lee

, Refshauge

, Kilbreath

. Segmental measurement of breast cancer-related arm lymphoedema using perometry and bioimpedance spectroscopy. Support Care Cancer, 2011; 19:703–710.

26.

Dylke

, Yee

, Ward

, Foroughi

, Kilbreath

. Normative volume difference between the dominant and nondominant upper limbs in healthy older women. Lymphat Res Biol, 2012; 10:182–188.

27.

York

, Ward

, Czerniec

, Lee

, Refshauge

, Kilbreath

. Single frequency versus bioimpedance spectroscopy in the assessment of lymphoedema. Breast Cancer Res Treat, 2009; 117:177–182.

28.

Svensson

, Dylke

, Ward

, Kilbreath

. Segmental impedance thresholds for early detection of unilateral upper limb swelling. Lymphat Res Biol, 2014; May 16 [Epub ahead of print].

29.

Ward

. Bioelectrical impedance analysis: Proven utility in lymphedema risk assessment and therapeutic monitoring. Lymphat Res Biol, 2006; 4:51–56.

30.

Stout

, Pfalzer

, Levy

, McGarvey

, Springer

, Gerber

, Soballe

. Segmental limb change as a predictor of the onset of lymphedema in women with early breast cancer. PM R, 2011; 3:1098–1105.

31.

Dylke

, Alsobayel

, Ward

, Liu

, Webb

, Kilbreath

. Use of impedance ratios to assess hand swelling in lymphoedema. Phlebology, 2012; 27:27.

32.

Ward

, Dylke

, Kilbreath

. Measurement of hand volume by bioimedance spectroscopy. Lymphat Res Biol, 2012; 10:81–86.

33.

Lee

, Boland

, Czerniec

. Reliability and concurrent validity of the perometer for measuring hand volume in women with and without lymphedema. Lymphat Res Biol, 2011; 9:13–18.