Abstract
Background This study assesses associations between bioimpedance spectroscopy (BIS) and magnetic
resonance lymphangiography (MRL) in the staging and assessment of lymphedema.
Methods Adults who received MRL and BIS between 2020 and 2022 were included. We collected
fluid, fat, and lymphedema severity ratings, and measured fluid stripe thickness,
subcutaneous fat width, and lymphatic diameter on MRL. BIS lymphedema index (L-Dex)
scores were collected from patient charts. We assessed sensitivity and specificity
of L-Dex scores to detect MRL-identified lymphedema, and examined associations between
L-Dex scores and MRL imaging measures.
Results Forty-eight limbs across 40 patients were included. L-Dex scores had 72.5% sensitivity
and 87.5% specificity for detecting MRL-defined lymphedema, with a 96.7% estimated
positive predictive value and 38.9% negative predictive value. L-Dex scores were associated
with MRL fluid and fat content scores (p ≤ 0.05), and lymphedema severity (p = 0.01), with better discrimination between fluid than fat content levels on pairwise
analysis, and poor discrimination between adjacent severity levels. L-Dex scores were
correlated with distal and proximal limb fluid stripe thickness (distal: rho = 0.57,
p < 0.01; proximal: rho = 0.58, p < 0.01), partially correlated with distal subcutaneous fat thickness when accounting
for body mass index (rho = 0.34, p = 0.02), and were not correlated with lymphatic diameter (p = 0.25).
Conclusion L-Dex scores have high sensitivity, specificity, and positive predictive value for
the identification of MRL-detected lymphedema. L-Dex has difficulty distinguishing
between adjacent severity levels of lymphedema and a high false negative rate, explained
in part by reduced discrimination between levels of fat accumulation.
Keywords
MR lymphangiography - bioimpedance spectroscopy - lymphedema - diagnostic imaging
- lymphedema staging
