Biomarkers in lymphedema assessment: integrating elastography and muti-frequency bioimpedance analysis

Abstract

Aim: Multi-frequency bioimpedance analysis (MFBIA) is used to measure lymphedema, but it is a biomarker that is sensitive to stiffness. Lymphedema is a condition that can be accompanied by stiffness, but no studies have considered this, so we tried to use non-invasive elastography as a biomarker for stiffness.

Methods & results: This retrospective study included 102 patients with lymphedema, divided into two groups according to the elastography strain ratio: stiff group (elastography strain ratio <0.7, n = 48) and non-stiff group (elastography strain ratio >0.7, n = 54). We estimated the volume of the affected arm based the extracellular water (ECW) volume calculated using MFBIA through a simple linear regression method. The adjusted R2 was 0.044 in the stiff group and 0.729 in the non-stiff group. Stepwise multivariate linear regression was used to investigate the significant factors for estimating the affected arm volume for each group. In the non-stiff group, the significantly associated factors were impedance at 50 kHz, weight, and height (adjusted R² = 0.724; p = 0.003). In the stiff group, significant associations were observed among impedance at 250 kHz, impedance at 1 kHz, weight, and height (adjusted R² = 0.705, p = 0.041).

Conclusion: Considering the characteristics of lymphedema, using MFBIA concurrently with elastography can be useful biomarker for estimating lymphedema.

Plain language summary

Article highlights.

Introduction

• Lymphedema initially presents as fluid accumulation in the extracellular space; as the disease progresses, fibrous matrix gradually accumulates in the extracellular space, resulting in fibrosis.

• Multi-frequency bioimpedance analysis (MFBIA) is a biomarker used to analyze body composition, including muscle, fat, and fluid volume, by measuring impedance values at different frequencies in the limb and determining the difference in conductivity of each tissue.

• Theoretically, the accuracy of estimating the volume of the affected arm based on the ECW volume calculated using the MFBIA would decrease in a group with a certain level of stiffness.

• Ultrasound elastography is an imaging technique that uses stiffness anisotropy differences to noninvasively assess the soft tissue stiffness.

Methods

• Arm volume measurement: Arm volume was estimated using arm circumference measurements at seven points, with calculations assuming a cylindrical shape.

• Ultrasound elastography measurement: The stiffness of subcutaneous tissue was assessed using ultrasound elastography, comparing values from the affected and unaffected arms to classify stiffness.

• Multi-frequency bioimpedance analysis: Using an impedance device, body composition including ECW and ICW was analyzed across multiple frequencies.

• Statistical analysis: The analysis included regression model to estimated affected arm volume based on ECW with comparisons made between lymphedema stages and another regression model was conducted to comparing groups according to elastography strain ratio.

Results

• Out of the 102 participants, they were divided into non-stiff (n = 54) and stiff (n = 48) groups based on the elastography strain ratio.

• The non-stiff group demonstrated better model fit for volume estimation with a higher adjusted R² and lower RMSE compared with the stiff group.

• Significant factors for estimating arm volume varied by group; in the non-stiff group, impedance at 50 kHz was significant, while in the stiff group, impedance at higher frequency (250 kHz) was more relevant.

Conclusion

• MFBIA has a role as a biomarker used to estimate lymphedema status in patients with lymphedema, but it is sensitive to arm stiffness. Non-invasive elastography can be used as a biomarker of lymphedema stiffness.

1. Introduction

Lymphedema is a progressive chronic disease that requires appropriate evaluation and management throughout the patient's life. Lymphedema is characterized by three stages: Stage 1 involves the early accumulation of fluid and swelling, Stage 2 is marked by the progression of fibrosis and stiffening of connective tissue, and Stage 3 features severe, chronic fibrosis leading to significant skin changes [1]. Various methods are used for evaluating lymphedema, and standardized diagnostic methods include circumferential measurement using a tape measure, water displacement volumetry, bioimpedance measurement, volumetry using a perometer, and self-reporting of symptoms [2–4]. Currently, no “gold standard” for determining volume state has been established; however, arm circumference measurement is widely used in clinical settings as an easily accessible method [3,5,6]. Multi-frequency bioimpedance analysis (MFBIA) is a biomarker used to analyze body composition, including muscle, fat, and fluid volume, by measuring impedance values at different frequencies in the limb and determining the difference in conductivity of each tissue [7–11]. This biomarker can be used to measure the extracellular water (ECW) volume in patients with early-stage lymphedema [11,12]. A meta-analysis indicates that ECW reflects volume status more effectively than intracellular water (ICW). Therefore, authors suggest that it is better to focus on ECW alone rather than using total body water (TBW), which includes ICW [11]. Lymphedema is a condition in which materials including fluids accumulate in the extracellular space. In the early stages, fluid accumulates in the extracellular space, making the quantitative assessment of lymphedema feasible using MFBIA [13,14]. The principles of MFBIA indicate that at lower frequencies, impedance is more sensitive to extracellular water, while at higher frequencies, impedance becomes more responsive to overall tissue composition, including the increased density of fibrotic tissue [8,15,16]. Lymphedema initially presents as fluid accumulation in the extracellular space; as the disease progresses, fibrous matrix gradually accumulates in the extracellular space, resulting in fibrosis. Therefore, if MFBIA, which is greatly affected by tissue characteristics, is used for the quantitative measurement of lymphedema, it should only be applied in the early stages before fibrosis progression [3,8,17,18].

Theoretically, the accuracy of estimating the volume of the affected arm based on the ECW volume calculated using the MFBIA would decrease in a group with a certain level of stiffness [19]. To the best of our knowledge, no previous studies have established the criteria for defining a specific level of stiffness. Ultrasound elastography is an imaging technique that uses stiffness anisotropy differences to noninvasively assess the soft tissue stiffness. Recently, it has also been used to assess lymphedema [4,20–24]. Therefore, this study aimed to measure and classify the tissue stiffness of a lymphedematous limb using ultrasound elastography to investigate whether this classification affects the accuracy of estimating the volume of affected arm using MFBIA.

2. Methods

2.1. Study design

This study was a retrospective analysis based on routine clinical data collected from patients in a real-world clinical setting. Due to the retrospective nature of the study, some aspects such as prospective measurement protocols were not implemented. Nonetheless, the clinical setting was meticulously structured to collect as comprehensive data as possible within the practical constraints of routine care.

2.2. Participants

This was a retrospective study of 102 patients with unilateral lymphedema who visited a lymphedema clinic between 1 January 2019, and 31 January 2022, after undergoing breast cancer surgery. All patients were women. Data on age, height (cm), weight (kg), body mass index (BMI), time from surgery (day), surgical method used (breast conservation or total mastectomy), axillary dissection status, side of lesion, lymphedema stage, elastography strain ratio, arm volume ratio, and MFBIA results (ECW volume (kg), ICW volume (kg), and impedance at each frequency (Ω)) were collected.

Approval for ethical conduct of the research was granted by the Institutional Review Board (IRB) of Catholic Kwandong University Hospital (IS23RISI0004) on February 23, 2023. As this study was conducted using retrospective medical records, the IRB waived the requirement for informed consent.

2.3. Arm volume measurement

The arm circumference measurement method was used to estimate the arm volumes in the affected and unaffected sides. The patients were placed on a sitting position with the forearm supinated, and measurements of both arms were taken at seven points (+15 cm above the elbow level, +10 cm above the elbow level and elbow crease level, -7 cm below the elbow level, and -15 cm below the elbow level, wrist level, and metacarpophalangeal joint level) using a tapeline to measure arm circumference (Figure 1). Data was collected by a single nurse in the clinic, who ensured that tension was applied to the tape measure without compressing the patient's arm, allowing for no space between the tape and the skin. The data collected during the clinical process was analyzed retrospectively, and the nurse who conducted the measurements remained the same throughout the data collection period. In this study, we measured the circumference of the affected arm at seven different points, which is more detailed than the typical four or six-point measurements used in other clinics [25,26]. This enhanced precision in circumference measurement was implemented to better capture the distribution and extent of lymphedema, ensuring a more accurate volume estimation. This meticulous setup differentiates our clinic from standard practices and contributes to the reliability of the collected data. Using the circumference measurements obtained in all seven points, the volume of each arm was calculated by assuming that the arm was cylindrical using the following formula [27]:

$$\text{Arm volume}=\frac{\left(C{1}^2+C{2}^2+C{3}^2+C{4}^2+C{5}^2+C{6}^2+C{7}^2\right)}{\pi }$$

Figure 1.

Circumference measured at seven points; each circumference defined as C1–C7. (A) Probe located below elbow using solid gel pad as ultrasound medium, (B) Ultrasound image, (C) Strain image of ultrasound elastography.

2.4. Ultrasound elastography measurement

To evaluate the lymphedema status, the stiffness of the subcutaneous tissue in both arms of the patient was measured using ultrasound elastography based on the differences in stiffness anisotropy [24,28–31]. An ultrasound (Logiq E9, GE Healthcare, US) and a 9-Hz linear probe were used. The patients were placed on a sitting position with the forearm supinated. To ensure inter-rater reliability and minimize the variability in elastography measurements, we used a solid pad gel instead of the conventional gel-type medium [32]. The solid pad gel allows for better control of pressure applied during the ultrasound examination, reducing the potential for measurement bias. This approach is not widely used in clinical practice but has been reported in select studies to improve measurement consistency. The use of this method highlights the rigorous setup of our clinic and supports the credibility of the data obtained. Measurements were taken on the ventral side of the arm (+10 cm above the elbow level) and forearm (-7 cm below the elbow level), and the average of the two values was used. The elastography strain value was defined as the ratio of the dermal and subcutaneous values (Figure 1).

In this study, the stiffness of the affected arm was classified by comparing the stiffness of the affected arm with that of the unaffected arm based on the ratio of elastic strain values measured in each arm.

$$\begin{array}{c}\text{Elastography strain ratio}=\\ \frac{\text{elasography strain of affected arm}}{\text{elastography strain of unaffected arm}}\end{array}$$

To account for measurement errors, a stiffness of 30% higher than that of the unaffected arm was defined as “stiff”; the participants were divided into two groups based on their elastography strain ratios: “stiff” group (elastography strain ratio <0.7; n = 48) and “non-stiff” group (elastography strain ratio >0.7; n = 54).

2.5. Multi-frequency bioimpedance analysis

MFBIA is a method used for analyzing body composition, such as muscle, fat, ECW, and ICW, by measuring impedance at various frequencies in the limbs. Different frequencies are used to distinguish between tissues and based on their specific electrical conductivity characteristics, as low frequency currents predominantly pass through extracellular water, while higher frequency currents penetrate both intracellular and extracellular compartments. By using these frequency-specific measurements, MFBIA can provide detailed information about body composition [7,9,10,27].

In clinical practice, only the impedance values are reported by the device, while specific resistance and reactance values are not typically accessible. The analysis was therefore based on the data available in routine clinical settings, and the absence of direct resistance values is a limitation inherent to the use of standard clinical equipment rather than research-specific instrumentation. An ACCUNIQ BC720 (Selvas Healthcare, Daejeon, Republic of Korea) machine was used, and measurements were taken using the tetrapolar 8-point tactile-electrode impedance method. The patients were examined while standing, with the anterior and posterior aspects of both feet in contact with the electrodes, and holding the handles with the palms and thumbs in contact with the electrodes [9]. We collected the impedance at frequencies of 1 kHz, 5 kHz, 10 kHz, 50 kHz, 500 kHz, and 1 MHz; the ECW and ICW volumes were also calculated using the machine.

2.6. Statistical analysis

In this study, we first analyzed the patients according to the stages of lymphedema and then divided them into stiff and non-stiff groups based on the elastography strain ratio, comparing the differences between the two groups. The Kolmogorov-Smirnov test was used to test the normality of continuous variables. Because all continuous variables showed normal distribution, we used an independent t-test to compare the differences between the two groups. Categorical variables were compared using the chi-square test or Fisher's exact test. Simple linear regression analyses were performed to estimate the affected arm volume based on the ECW volume calculated for each group using the MFBIA. The adjusted R² and root mean squared error (RMSE) were obtained to evaluate the regression performance metrics. Additionally, regression analysis was conducted after removing outliers to address potential leverage effects. As an additional analysis, we performed stepwise multivariate linear regression using weight, height, age, and impedance at each frequency to determine the significant factors in estimating the affected arm volume for each group. Statistical analysis was performed using SPSS software (version 22.0; IBM Corp., NY, USA), and a p-value of less than 0.05 was considered significant.

3. Results

A total of 102 patients were included in the study and divided into non-stiff group (n = 54) and stiff group (n = 48) based on the elastography strain ratio results. The mean ages were 58.69 years in the non-stiff group and 58.81 years in the stiff group; the BMI values were 25.83 and 25.38 in the non-stiff and stiff group, respectively, with no significant differences between the two groups. The general characteristics of all participants are presented in Table 1.

Table 1.

General Characteristics.

Group	Total (n = 102)	Non-stiff (n = 54)	Stiff (n = 48)	p-value
Age	58.74 ± 11.54	58.69 ± 11.27	58.81 ± 11.97	0.956
Weight	63.47 ± 7.98	64.56 ± 8.47	62.25 ± 7.30	0.145
Height	157.44 ± 5.35	158.16 ± 5.52	156.64 ± 5.10	0.155
ICW, unaffected side	1.39 ± 0.40	1.41 ± 0.36	1.37 ± 0.46	0.659
ECW, unaffected side	0.85 ± 0.17	0.87 ± 0.16	0.83 ± 0.19	0.297
ICW, affected side	1.52 ± 0.35	1.53 ± 0.33	1.51 ± 0.39	0.797
ECW, affected side	0.91 ± 0.16	0.93 ± 0.17	0.91 ± 0.15	0.481
BMI	25.61 ± 3.08	25.83 ± 3.31	25.38 ± 2.81	0.469
Time from surgery	1496.24 ± 1534.26	1526.41 ± 1570.94	1462.31 ± 1507.77	0.834
Surgical method				1.000
Breast conservation	80 (78.43%)	42 (77.78%)	38 (79.17%)
Total mastectomy	22 (21.57%)	12 (22.22%)	10 (20.83%)
Axillary dissection				0.264
No	6 (5.88%)	5 (9.26%)	1 (2.08%)
Yes	96 (94.12%)	49 (90.74%)	47 (97.92%)
Side of lesion				0.817
Left	66 (64.71%)	36 (66.67%)	30 (62.50%)
Right	36 (35.29%)	18 (33.33%)	18 (37.50%)
Stage of lymphedemac				<0.001a
Stage 1	32 (31.37%)	25 (46.30%)	7 (14.58%)
Stage 2	62 (60.78%)	29 (53.70%)	33 (68.75%)
Stage 3	8 (7.84%)	0 (0.00%)	8 (16.67%)
Elastography strain ratiob		2.00 ± 1.05	0.45 ± 0.15	<0.001a
Arm volume of unaffected side	1163.46 ± 183.31	1182.29 ± 180.28	1142.29 ± 186.28	0.273
Arm volume of affected side	1381.76 ± 266.59	1405.21 ± 277.08	1355.39 ± 254.60	0.349
Arm volume ratiob	1.18 ± 0.11	1.19 ± 0.12	1.19 ± 0.11	0.982
X1 kHz unaffected side	344.7 ± 38.2	349.7 ± 33.5	339.2 ± 42.6	0.168
X5 kHz unaffected side	272.9 ± 46.3	273.7 ± 39.7	272.0 ± 53.2	0.857
X50 kHz unaffected side	251.7 ± 51.6	259.4 ± 46.0	243.1 ± 56.4	0.113
X250 kHz unaffected side	214.8 ± 40.5	203.3 ± 29.5	215.0 ± 35.7	0.074
X550 kHz unaffected side	184.8 ± 34.0	183.6 ± 34.6	186.1 ± 33.7	0.710
X1 MHz unaffected side	168.5 ± 27.1	167.9 ± 28.1	169.1 ± 26.2	0.822
X1 kHz affected side	328.6 ± 30.0	331.8 ± 31.7	325.0 ± 27.9	0.253
X5 kHz affected side	238.3 ± 39.8	236.8 ± 39.4	240.0 ± 40.7	0.687
X50 kHz affected side	202.6 ± 40.2	204.7 ± 39.2	200.2 ± 41.6	0.578
X250 kHz affected side	186.7 ± 32.6	182.1 ± 27.3	191.8 ± 37.3	0.140
X550 kHz affected side	163.5 ± 28.1	163.4 ± 27.1	163.6 ± 29.5	0.968
X1 MHz affected side	156.0 ± 20.2	156.6 ± 19.2	155.3 ± 21.6	0.746

^ap < 0.05.

^bAffected side/unaffected side.

^cStage 1: early stage of lymphedema, swelling occurs as fluid accumulates; Stage 2: fibrosis progresses, and the connective tissue becomes stiff; Stage 3: fibrosis becomes severe and chronic.

Significant differences were observed in lymphedema stage and elastography strain ratio between the non-stiff and stiff groups.

BMI: Body mass index; ECW: Extracellular water; ICW: Intracellular water; X: Impedance at frequency.

When a simple linear regression analysis according to lymphedema stage was performed to estimate the affected arm volume using the ECW volume of the total group (n = 102), the adjusted R² was 0.343 and RMSE was 217.107. In the Stage 1 group (n = 32), the adjusted R² was 0.623 and RMSE was 148.755. Meanwhile, in the Stage 2 group (n = 62), the adjusted R² was 0.256 and RMSE was 218.519, while in the Stage 3 group (n = 8), the adjusted R² was 0.604 and RMSE was 245.339 (Figure 2). In the scatter plot of the simple regression model, some outliers that could have a leverage effect were observed. Therefore, we also present the scatter plot after removing the outliers. The adjusted R² were 0.403 and RMSE was 195.016 in total group (n = 100) after removing outlier data. And the adjusted R² were 0.339 and RMSE was 186.967 in Stage 2 group (n = 60) after removing outlier data (Figure 3). Another simple linear regression analysis according to the elastography strain ratio was performed to estimate the affected arm volume using the ECW volume. In the non-stiff group (n = 54), the adjusted R² was 0.729 and RMSE was 145.713. Meanwhile, in the stiff group (n = 48), the adjusted R² was 0.044 and RMSE was 251.578 in stiff group (n = 48) (Figure 4). We also presented the scatter plot after removing outliers. In the stiff group (n = 46) after removing outlier data, the adjusted R² were 0.070 and RMSE was 206.179 (Figure 5).

Figure 2.

Scatterplot of regression model for estimating arm volume by stage classification. (A) Total group, (B) Stage 1, (C) Stage 2, (D) Stage 3.

Figure 3.

Scatterplot of regression model for estimating arm volume by stage classification excluding outliers. (A) Total group, (B) Stage 1, (C) Stage 2, (D) Stage 3.

Figure 4.

Scatterplot of regression model for estimating arm volume by elastography classification. (A) Non-stiff group, (B) Stiff group.

Figure 5.

Scatterplot of regression model for estimating arm volume by elastography classification excluding outliers. (A) Non-stiff group, (B) Stiff group.

The simple regression showed a significant difference, indicating that there are frequencies sensitive to tissue condition [15]. To identify which frequencies are sensitive according to the status of lymphedema, we performed a multivariate regression analysis. The multivariate regression was conducted using a forward stepwise approach, analyzing the variables starting from the most significant ones. Multicollinearity was examined by checking the variance inflation factor values. Table 2 shows the results of the stepwise multivariate linear regression analysis conducted to identify the significant factors for estimating the affected arm volume in each of the two groups divided by the elastography strain ratio. For the non-stiff group, the significant factors for estimating the affected arm volume were impedance at 50 kHz, weight, and height, with an adjusted R² of 0.724 (p = 0.003) and RMSE of 149.825. For the stiff group, the significant factors for estimating the affected arm volume were impedance values at 250 kHz and 1 kHz, height, and weight, with an adjusted R² of 0.705 (p = 0.041) and RMSE of 144.565.

Table 2.

Multiple linear regression model for estimating arm volume.

	Non-stiff group (Adjusted R2 = 0.724, p = 0.003a)			Stiff group (Adjusted R2 = 0.705, p = 0.041a)
	B (95% CI)	p-value	VIF	B (95% CI)	p-value	VIF
Weight	23.560 (17.891, 29.228)	<0.001a	1.348	25.160 (18.345, 31.975)	<0.001a	1.370
1 kHz				2.306 (0.096, 4.515)	0.041a	2.115
50 kHz	-2.127 (-3.428, -0.825)	0.002a	1.528
250 kHz				-2.925 (-4.669, -1.180)	0.002a	2.349
Height	-14.755 (-24.352, -5.158)	0.003a	1.644	-13.079 (-22.702, -3.456)	0.009a	1.333
Constant	2653.090 (1387.013, 3919.167)	<0.001a		1649.706 (277.145, 3022.267)	0.020a

^ap < 0.05.

Independent variable (impedance at 550 kHz) that has high VIF (>10) were excluded from regression model.

Significant factors for estimating affected arm volume: non-stiff group-impedance at 50 kHz, weight, height; stiff group- impedance at 250 kHz, 1 kHz, height, weight.

CI: Confidence interval; VIF: Variance inflation factor.

4. Discussion

This retrospective study was conducted to determine whether stiffness biomarker using an ultrasound elastography device affected the arm volume estimation obtained using MFBIA. Stiffness was measured using ultrasound elastography, and the group whose arm stiffness was 30% higher than that of the unaffected arm was defined as the stiff group. The results of this study showed differences in the estimation of lymphedema volume using bioimpedance in the two groups classified according to the ultrasound elastography-based classification.

The results of the simple linear regression analysis by lymphedema stage to estimate affected arm volume based on ECW volume showed an adjusted R² of 0.343 and an RMSE of 217.107 for the total group (n = 102), an adjusted R² of 0.256 and an RMSE of 218.519 for the Stage 2 group (n = 62), and an adjusted R² of 0.604 and and RMSE of 245.339 for the Stage 3 group (n = 8). Notably, the Stage 1 group (n = 32) showed significantly better model fit, with an adjusted R² of 0.623 and an RMSE of 148.755. Among the 62 patients in Stage 2, elastography classification divided them into 29 non-stiff and 33 stiff patients. The lower model performance observed in Stage 2, as reflected in the lower adjusted R² and higher RMSE, indicates reduced accuracy in this group. Similarly, the total group also showed lower accuracy compared with the Stage 1 group, which had the best fit. The Stage 3 group had very few patients (n = 8), and the large RMSE suggests high variability in this group. The results of the simple linear regression analysis by elastography strain ratio conducted to estimate the affected arm volume using ECW volume showed an adjusted R² of 0.343 and an RMSE of 217.107 for the total group, and an adjusted R² of 0.044 and an RMSE of 251.578 for the stiff group. In contrast, the non-stiff group showed significantly better model fit, with an adjusted R² of 0.729 and an RMSE of 145.713. These findings are consistent with those reported in previous studies and clinical guidelines, which recommend using MFBIA only in the early stages of lymphedema [3]. They also suggest the limitations of MFBIA in more advanced-stage lymphedema. In healthy individuals, the extracellular space is very narrow compared with that in patients with lymphedema, and the change in volume is markedly determined by the amount of ECW. Therefore, ECW measurement by MFBIA can sensitively reflect the changes in arm volume. Similarly, in patients with early lymphedema without fibrosis, swelling is mainly due to fluid accumulation; therefore, the volume of the arm is closely related to the ECW volume biomarker using MFBIA [1,8]. However, in the stiff group, fibrotic changes occurred in the extracellular space as lymphedema progressed; this finding suggests that substances other than fluids occupy the extracellular space, causing the volume of the arm to be less correlated with ECW volume measured by MFBIA [5]. Although the theoretical background on the impact of stiffness in evaluating lymphedema with MFBIA has been suggested [19], no previous studies have presented a method for measuring stiffness and its criteria. The results of this study showed that when the stiffness measured by ultrasound elastography in patients with lymphedema differed by less than 30% from that of the unaffected arm, estimating the volume of the affected arm using the MFBIA-based ECW volume could be relatively accurate. However, when the stiffness was more than 30% greater than that of the unaffected arm, the accuracy of estimating the volume of the affected arm using the ECW volume was markedly reduced. This confirms the value of elastography as a biomarker for measuring lymphedema stiffness.

To further investigate the relationship between bioimpedance and arm volume in patients with lymphedema, we examined which variables of impedance measured at various frequencies using MFBIA were significant in estimating the volume of the affected arm in each group. We conducted a forward stepwise multivariate linear regression analysis to identify these variables; in the non-stiff group, the significant variables for estimating the affected arm volume were impedance at 50 kHz, weight, and height. In the stiff group, the significant variables for estimating the affected arm volume were impedance values at 250 kHz and 1 kHz, weight, and height. The frequency values that significantly affected the arm volume estimation differed between the groups. In previous studies, impedance values of 50 kHz and 5 kHz were mainly used to calculate the ECW volume using bioimpedance in healthy individuals [8,10,11,15,33,34]. When estimating the ECW volume using MFBIA, the weight may vary depending on the statistical contribution; in general, impedance at low frequencies is the main variable that reflects the ECW volume. Consistent with the results of multiple studies, the impedance at 50 kHz has been the most significant variable related to the ECW volume [10,11,15]. In addition, among various frequencies, impedance at 50 kHz was the most significant variable in estimating the arm volume in the non-stiff group. This is believed to be because the non-stiff group had conditions similar to those of healthy individuals or even had more fluid accumulation with less stiffness, making the impedance at 50 kHz, which represents the ECW volume, a good biomarker of arm volume.

In our study, the non-stiff group, classified using ultrasound elastography, showed early stages of lymphedema and relatively fluid-rich arm, which could be more sensitively measured at low frequencies. However, as lymphedema progresses, the substances other than fluid may occupy the volume of the extracellular space, accompanied by fibrotic changes [1]. Among the various tissue types, fibrotic tissue, which has lower electrical conductivity than fluid, can affect the impedance at higher frequencies [9,16,35]. Consistent with this theoretical background, our study results showed that the impedance at 50 kHz was not a significant variable in estimating the affected arm volume in the stiff group, whereas the impedance at a higher frequency of 250 kHz was a significant variable.

In bioimpedance analysis, which measures the body composition, the impedance values are proportional to height²/resistance in the calculation of volume, as they are influenced by the tissues resistance value and length [7–10,15]. A previous study that performed lymphedema analysis using BIA found a correlation between BIA and lymphedema stage, but no correlation was found between BIA and lymphedema severity according to volume [36]. Considering that the frequency reflecting the arm volume differed depending on the degree of stiffness in our study, it is assumed that previous studies did not consider or adjust for stiffness [37–39]. Based on the results of this study and the theoretical background that the frequency reflecting the volume of the arm may change with the progression of lymphedema, we suggest that stiffness, a characteristic of lymphedema, should also be considered when assessing lymphedema volume using BIA.

Lymphedema is a progressive chronic disease that requires continuous evaluation and management throughout the patient's life. Therefore, it is important to understand the nature of lymphedema progression and develop evaluation biomarker that can provide accurate measurements. In the clinical evaluation of lymphedema, classification is typically performed based on physical examination and staging. However, as lymphedema progresses, even within Stage 2, patients may present with fluid-rich conditions in the early phase (as seen in Stage 1), whereas in the later phase, prior to Stage 3, fibrotic changes may become more pronounced. MFBIA measurement is a reliable and noninvasive biomarker that is less time consuming, cost effective, and convenient. However, owing to the nature of lymphedema progression, which involves changes in tissue stiffness, it can only be applied to a limited number of cases. To more accurately assess these changes, the application of ultrasound elastography, alongside MFBIA, is crucial for evaluating the status of lymphedema. In this study, when predicting ECW-based volume, the performance metrics of the simple regression model differed across stages, particularly between Stage 1 and Stages 2 and 3. Furthermore, in the elastography-based simple regression model, the differences between groups were more distinct. Therefore, elastography-based classification, which accounts for tissue changes, may provide a more effective method of classification than physical examination alone. This suggests that elastography could serve as a clinically useful biomarker for classifying lymphedema based on tissue status. This study provides initial evidence that noninvasive ultrasound elastography can be applied when the degree of stiffness is considered during MFBIA for estimating the arm volume in patients with lymphedema.

This study has some limitations. First, the sample size was relatively small, which prevented cross-validation from being performed in the analysis of the regression model. Second, this study was conducted retrospectively using data collected from routine clinical assessments to calculate arm volume. In clinical practice, lymphedema is typically monitored by measuring circumferences at various points along the arm to track changes over time. Although this approach is widely used, it does not provide segmental length data required for more accurate volume estimation methods, such as the frustum or truncated cone model. Instead, we employed the cylindrical model, which is a feasible option for estimating volume based on available circumference measurements. This choice reflects the practical constraints of using routinely collected clinical data rather than prospective research setting with comprehensive measurement protocols. Additionally, for tissue status assessment, we used the impedance values provided by the MFBIA device. While this method is suitable for clinical practice, we were unable to access resistance and reactance values, which would have offered a more comprehensive analysis of tissue composition. The equipment used in our clinic reports only aggregate impedance values, and obtaining more detailed bioimpedance parameters would have required different, research-specific devices. This limitation is inherent to the use of standard clinical equipment, and although it constrains the granularity of our analysis, it accurately reflects the real-world clinical application of these tools. Third, the criteria for dividing the stiff group (those whose arm was 30% stiffer than the unaffected arm) were based on a rule of thumb. This study was first study that introduced non-invasive ultrasound elastography-based classification for lymphedema. This classification can be applied when the degree of stiffness is considered during MFBIA for estimating the arm volume in patients with lymphedema. In the future, larger, well-designed prospective studies should be conducted to investigate the linear relationship between stiffness measured by elastography and BIA as well as the evidence-based criteria for standard references.

5. Conclusion

MFBIA has a role as a biomarker used to estimate lymphedema status in patients with lymphedema, but it is sensitive to arm stiffness. Non-invasive elastography can be used as a biomarker of lymphedema stiffness.

Author contributions

Conceptualization: DY Kim, data curation: DY Kim, D Kim; H-W Han; N Jeon, H Jeon, formal analysis: DY Kim, supervision: DY Kim, S-W Park; B-S Lee; writing: DY Kim, H Jeon, writing (review and editing): DY Kim, D Kim, H-W Han, N Jeon, H Jeon, S-W Park, B-S Lee.

Financial disclosure

The authors have no financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, stock ownership or options and expert testimony.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

The authors state that they have obtained appropriate institutional review board approval (Institutional Review Board (IRB) of Catholic Kwandong University Hospital (IS23RISI0004)) and/or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.