Lymphatic ultrasound detects more vessels than indocyanine green lymphography in lymphedematous limbs.

Abstract

Indocyanine green (ICG) lymphography is widely used for preoperative assessment in lymphaticovenous anastomosis; however, its limited imaging depth and inability to detect vessels with impaired flow may lead to underestimation of surgical targets. Lymphatic ultrasound is a noninvasive, contrast-free modality capable of visualizing deeper lymphatic vessels and providing structural assessment. This retrospective study compared lymphatic ultrasound and ICG lymphography for detecting lymphatic vessels in patients with lower limb lymphedema. Thirteen female patients (26 limbs; 104 areas) underwent both ICG lymphography and lymphatic ultrasound preoperatively. Each limb was divided into four regions (lateral thigh, medial thigh, medial calf, and lateral calf). The number of lymphatic vessels identified by each modality was recorded per region, and analyses were stratified by dermal backflow (DB) status on ICG lymphography. Overall, lymphatic ultrasound identified significantly more lymphatic vessels than ICG lymphography (mean ± SD: 1.9 ± 1.5 vs. 0.6 ± 0.7 vessels per area, p < 0.01). In DB-positive areas, ultrasound detected more vessels (1.9 ± 1.5) than ICG (0.3 ± 0.5, p < 0.01). In DB-negative areas also, ultrasound identified more vessels (2.0 ± 1.5) than ICG (1.1 ± 0.9, p < 0.01), except in the DB-negative lateral calf, where ICG detected more vessels. Across all the International Society of Lymphology stages, ultrasound identified more vessels; in stage 3, no vessels were detected using ICG, whereas ultrasound detected an average of 2.4 vessels per area. Lymphatic ultrasound detects more lymphatic vessels than ICG lymphography, particularly in DB-positive regions and advanced stages. Negative ICG findings did not necessarily indicate the absence of lymphatic vessels. The combined use of ultrasound and ICG lymphography may broaden lymphaticovenous anastomosis eligibility and enhance preoperative planning.