Magnetic resonance imaging patterns and perfusion changes of posterior reversible encephalopathy syndrome in children with clinical outcome correlation.

Abstract

Background: Posterior reversible encephalopathy syndrome (PRES) in children has a propensity towards atypical features on magnetic resonance (MR) imaging, with limited literature on perfusion changes and clinicoradiological correlation.

Objective: We aimed to comprehensively study MR imaging patterns of pediatric PRES, including cerebral blood flow variations on arterial spin labeling, and looked for any MR biomarkers of poor clinical outcome.

Materials and methods: In this retrospective observational study conducted in a tertiary hospital setting, MR records over a 4-year period (May 2019 to May 2023) were systematically searched along with their clinical details. Patients with an age less than 18 years and a clinicoradiological constellation consistent with PRES were included. MR scans were analyzed by two neuroradiologists with 8 years' and 10 years' experience. Association was sought with poor clinical outcome (defined as modified Rankin Scale score at discharge of > 2).

Results: A total of 45 patients (29 boys) were included in the study, with a mean age (± standard deviation) of 11.19 (± 4.53) years. On MR imaging, 95.6% of patients (n = 43) showed atypical features and/or atypical areas of involvement. The superior frontal sulcus (n = 18) was the most predominant MR pattern, and cerebellar involvement was not uncommon (n = 15). Unilateral involvement (n = 3), isolated central pattern (n = 1), and spinal cord involvement (PRES-SCI: n = 1) were also encountered. Brainstem involvement (n = 4) showed a characteristic "V-sign" of anterior medullary hyperintensity. Patchy restricted diffusion (46.6%), punctate hemorrhages (37.7%), and leptomeningeal contrast enhancement (36%) were not uncommon. Arterial spin labeling sequence (available in 24 patients) showed increased cerebral blood flow in the involved areas in 79.2% of patients. Univariate analysis showed a significant association of the presence of hemorrhage (P = 0.003), involvement of brainstem (P = 0.007), deep white matter (P = 0.008), and thalamus (P = 0.026) with poor clinical outcome. Multivariate regression analysis found that hemorrhage on MRI (P = 0.011, odds ratio 8) was an independent factor associated with poor clinical outcome.

Conclusions: The conventionally described atypical features in PRES are common in children and therefore may no longer be considered exceptions. Raised perfusion on arterial spin labeling sequence was seen in the majority of cases. Hemorrhage on MRI was an independent predictor of poor clinical outcome in pediatric PRES.