Clinicoradiologic predictors of sclerotherapy response in low-flow vascular malformations

To evaluate the clinical effectiveness of sclerotherapy agents in low-flow vascular malformations (LFVM) and identify clinical/imaging features to predict response.

A retrospective analysis of hospital records of symptomatic LFVM patients who underwent phlebosclerotherapy from January 2015 to April 2018 was done. Patients were subdivided into venous malformations (VM) and lymphatic malformations (LM). Out of 246 cases, 223 patients (132 males, 91 females; age range, 2-52 years) had VM and 23 (13 males, 10 females; age range, 3 months to 45 years) had LM. The clinical response was graded as excellent (>60%), good (30%-60%), and poor (<30%). More than 30% was considered as acceptable response. The χ² test was performed for correlation between clinical response and clinical, sonographic, magnetic resonance imaging, phlebographic parameters followed by multilinear regression.

Cavitary (43%) and spongy (37.7%) were the most common phlebographic patterns seen among VM and a cavitary pattern (87%) was most frequent in LM. Sodium tetradecyl sulphate and bleomycin were most commonly used sclerosants in VM and LM, respectively. The mean number of sessions was 4.35 (range, 1-23) in VM and 2.64 (range, 2-7) in LM. Among VM, 114 patients (51.1%) had excellent response to treatment (>60%) and 75.8% patients had an acceptable response (>30%). All patients with LM had an acceptable response (excellent response in 86.9%). Clinical disfigurement, discoloration, diffuse involvement, dysplastic venous morphology on phlebogram, and late and indirect draining vein correlated with poor response to sclerotherapy in VM (P = .003, P = .036, P = .007, P = .008, P = .003, and P = .035, respectively). Cystic components on ultrasound examination and direct draining vein were seen more often in excellent responders (P = .004 and P = .007) in addition to absence of disfigurement, discoloration, and diffuse involvement (P = .032, P = .003, and P = .002). Mod els comprising clinical disfigurement, dysplastic veins, and late draining vein had the greatest predictive value for poor response (R² = 0.256). Also, the best model for predicting excellent response comprised presence of direct draining vein and absence of skin discoloration (R² = 0.109). Eleven instances of minor complications occurred among a total of 1032 sessions, seven with sodium tetradecyl sulphate and four with polidocanol.

Acceptable response to sclerotherapy was achieved in majority of LFVM with extremely low complication rates. Clinicoradiologic features, especially phlebographic findings, correlated with response to sclerotherapy.