Reproducibility of hindbrain herniation severity grading on MRI scan in fetuses undergoing prenatal open neural tube...

Hindbrain herniation is part of Chiari II deformation and serves as an inclusion criterion for prenatal open neural tube defect (ONTD) repair. Its assessment on fetal MRI remains challenging, due to limited spatial resolution, fetal motion and difficulty obtaining a true mid-sagittal plane.

We aimed to evaluate the reproducibility of the Hindbrain herniation severity classification, based on repeated consensus readings by two evaluators.

We retrospectively reviewed fetuses who underwent prenatal ONTD repair at a single center (2011-2024), with pre- and 6 week post-operative T2-weighted fetal brain MRI scans. Hindbrain herniation was graded as: 0 = normal; 1 = effacement of the 4th ventricle and/or cisterna magna without Hindbrain herniation; 2 = Hindbrain herniation above the C1 posterior arch; 3 = Hindbrain herniation bellow C1 posterior arch. Two experienced fetal imaging specialists independently assessed all scans, resolved discrepancies by consensus, and repeated the evaluation two months later, blinded to initial grading. Reproducibility was assessed with weighted Cohen's Kappa overall and separately for pre- and post-operative scans.

121 fetuses were included. Postoperative MRI was available in 113 (93.4%). Overall reproducibility was excellent (κ = 0.87). Agreement was lower for pre-operative (κ = 0.57) than post-operative (κ = 0.74) scans (p < 0.01).

Hindbrain herniation grading on fetal MRI shows excellent overall reliability. The moderate agreement in preoperative scans reflects the difficulty of assessing Hindbrain herniation before prenatal repair and highlights the need for more standardized imaging criteria, when accurate assessment is necessary for surgical eligibility and parental counseling.