MRI for Response Assessment of Extensive Lymphatic Malformations in Children Treated With Sirolimus

BACKGROUND. Extensive lymphatic malformations (LMs) may cause substantial morbidity. The mammalian target of rapamycin (mTOR) inhibitor sirolimus shows promise for treating vascular anomalies, although response assessment is not standardized. OBJECTIVE. The purpose of this study was to retrospectively characterize changes seen on MRI of children with extensive LMs treated with sirolimus. METHODS. Twenty-five children treated with sirolimus for extensive LMs were included. Baseline MRI was defined as the MRI examination performed closest to therapy initiation; follow-up MRI was defined as the most recent MRI examination performed while the patient was receiving therapy. Two pediatric radiologists independently determined MRI lesion volume by tracing lesion contours on all slices (normalized to patient body surface area expressed in square meters) and determined signal by placing an ROI on the dominant portion of the lesions (normalized to CSF signal) on baseline and follow-up T2-weighted MRI sequences. Interreader agreement was determined, and values were averaged for further analysis. Volume and signal changes were compared with patient, lesion, and treatment characteristics. RESULTS. The mean (± SD) interval between initiation of sirolimus treatment and follow-up MRI was 22.1 ± 13.8 months. The mean lesion volume index on baseline and follow-up MRI was 728 ± 970 and 345 ± 501 mL/m², respectively (p < .001). Ninety-two percent of children showed a decrease in lesion volume index that was greater than 10% (mean volume change, -46.4% ± 28.2%). Volume change was inversely correlated with age (r = -0.466; p = .02). The mean volume change was -64.7% ± 25.4% in children younger than 2 years old versus -32.0% ± 21.6% in children 2 years old or older (p = .008). The mean volume change was -58.1% ± 24.0% for craniocervical lesions versus -35.5% ± 28.2% for lesions involving the trunk and/or extremities (p = .03). Mean lesion signal ratio on baseline and follow-up MRI was 0.81 ± 0.29 and 0.59 ± 0.26, respectively (p < .001). Mean signal ratio change was -23.8% ± 22.7%. Volume and signal changes were moderately correlated (r = 0.469; p = .02). Volume and signal changes were not associated with sex, lesion subtype, serum concentration of sirolimus, or the interval between sirolimus initiation and follow-up MRI (p > .05). Interreader agreement for volume index change was excellent (intraclass correlation coefficient, 0.983), and that for signal ratio change was moderate to good (intraclass correlation coefficient, 0.764). CONCLUSION. Sirolimus treatment of extensive LMs in children is associated with significant reductions in volume and signal on T2-weighted MRI. The decrease in volume is greater in younger children and craniocervical lesions. CLINICAL IMPACT. The results may facilitate development of standardized MRI-based criteria for assessing the response of vascular malformations to pharmacotherapy.