Harvey N. Mayrovitz, PhD, Alexander Mikulka, MS, and Don Woody. Lymphatic Research Biology 2018

Main findings

• As tissue dielectric constant (TDC) measurements are increasingly being used in clinical practice toassess lymphoedema it is essential that the measurement reliability be established.
• A central question in this regard relates to the minimal detectable change (MDC) that, with a given level of confidence, may be ascribed to this technology.
• This study addressed this issue by performing test-retest measurements using two different TDC measuring devices from which intraclass correlations coefficients (ICC) and MDC could be estimated thereby providing a guide to aid interpretation of measured TDC changes or differences in a clinical environment.
• Sample N= 40 adult volunteers without lymphoedema
• Two different TDC measuring devices were used, both manufactured by Delfin Technologies (Kuopio, Finland). One was the MoistureMeterD using a 2.5mm effective depth measuring probe, and the other was the more recently available compact version referred to as the MoistureMeter Compact using a 2.0mm effective depth. TDC measurements were obtained by placing the probe surface perpendicularly on the subject’s skin with firm but gentle contact pressure.
• Measurements were done in triplicate sequentially and bilaterally at three standardized locations: (1) the anterior forearm 5cm distal to the antecubital fossa, (2) the center of the hand palmar thenar eminence, and (3) the hand dorsum mid-web between the thumb and index finger that was not over bone. The triplicate measurements at each site were done by alternating between sides sequentially, starting with the dominant side. The order of the measurements was hand dorsum to hand palmar to forearm. After placing the probe on the skin, a measurement takes about 5 seconds. Anatomical sites were chosen to be inclusive of sites used related to lymphedema and other TDC related measurements.
• The main findings with respect to measured TDC values are as follows.

(1) For both measurers and for both probes TDC values were significantly different among anatomical sites (p<0.001) with the forearm having the least TDC value and the hand palm showing the greatest TDC value.

(2) TDC values recorded by the compact probe were statistically different from that recorded by the multiprobe probe at all sites and for both measurers. However, the main difference between probes was recorded at the forearm where the compact probe TDC values exceeded the multiprobe by an average of 16.3% for measurer 1 (M1) and by 16.6% for measurer 2 (M2). At the hand sites the difference between probe values was less than 5% at the hand dorsum and less that 2% at hand palm.

(3) There was an overall statistically significant difference (p<0.001) in TDC values measured by M1 and M2 at each site with percentage differences ranging from 2.0% to 4.3% for the multiprobe and 2.2% to 5% for the compact probe.

• The fact that absolute TDC values recorded at the forearm were greater when measured using the compact probe was confirmed through the present measurements, with the compact yielding a value of about 4.2 TDC units higher at the forearm. This difference is likely since measuring deeper will include more low-water content subcutaneous fat to be included in the measured volume. The effect would be a lower TDC reading for the 2.5mm probe.
• Contrastingly, average TDC values at the hand differed by less than 1.0 TDC unit. The smaller probe-related difference might reflect structural differences inherent among anatomical sites.
• The MDC values provide the main reliability outcome of the present study. Because some research and clinical applications may utilize absolute TDC differences while others may use inter-side ratios, MDC is needed separately for each application. The results for the absolute value analysis demonstrate that MDCs are dependent on the anatomical site being evaluated and on the probe being used. For all sites measured, the compact probe had the least MDC value that ranged from 2 TDC units at the forearm to 4 TDC units at the hand dorsum.
• The standard compact device has an included pressor sensor, not present in the multiprobe device, which might allow for a more uniform application pressure which in turn may have yielded a greater intraclass correlation coefficient value. In addition, because the compact probe measures to a lesser depth than the multiprobe, minor differences in contact pressure during the measurement might have less effect on the measured TDC value. This follows since a lesser depth measurement probe includes proportionately moreof the homogeneous high-water content dermal region as opposed to possibly including different proportions of dermis and low-water content hypodermis as would be the case for the deeper measurement probe.

• The minimal detectable change for the probes were:

• These MDC values may be used for lymphedema assessments in two separate ways. When comparing a prior measured interarm ratio to a measurement taken later or after treatment, a change in ratio needs to be greater than the MDC values specified in the table above to be considered a possibly real change. For example, if a presurgery interarm ratio (at-risk/contralateral) of 1.065 was measured at the forearm, using the multiprobe, then anything less than a subsequently measured ratio less than 1.065+0.080 or 1.145 should not be considered as a real change at the 95% confidence level. If the compact probe were being used, its smaller MDC leads to a threshold of 1.065+0.067=1.132. Similar calculations could be done for other anatomical sites of interest.
• MDC values determined for interarm ratios may also help specify threshold ratios to detect lymphedema. Such thresholds have been calculated as those values that exceed normal reference ratio averages by at least 3 SDs. For forearm these range between 1.262, and 1.2938 and for hand dorsum are 1.23 for a 2 SD threshold or 1.32 for a 3 SD threshold.6 Threshold values based on the present more limited data set and calculated using the 3 SD threshold using site-dependent and probe-dependent MSD values added ranged between about 1.23 and 1.27, values similar to those previously reported.