Concordance and Association of Kidokoro MRI Scores at 32 and 40 Weeks Postmenstrual Age with Neurodevelopmental...

Early identification of neurodevelopmental risk in very preterm infants is critical for timely intervention. The Kidokoro MRI scoring system offers a semiquantitative approach to evaluating brain abnormalities. This study aimed to assess the agreement of Kidokoro scores obtained in the same infants at 32 and 40 weeks postmenstrual age, and their association to cognitive, language, and motor outcomes at 24 months corrected age.

A cohort of 187 very preterm infants (median 28⁺⁵ weeks gestational age; 55% boys) underwent structural MRI at 2 time points: early (median 32⁺² weeks) and term-equivalent age (TEA; median 40⁺⁵ weeks), using 3T scanners. Scans were scored independently using a modified version of the Kidokoro system. Bland-Altman analysis assessed agreement between early and TEA scores. Neurodevelopmental outcomes were assessed at 24 months using the Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III). Multivariable linear regression models evaluated the predictive value of MRI scores, adjusting for sex, gestational age, birth weight centile, and socioeconomic status.

Early MRI scores were consistently higher than TEA scores (mean difference: 1.9; 95% CI, 1.6-2.2), with WM scores showing the largest discrepancy (mean difference: 1.68; 95% CI, 1.46-1.90), primarily due to the myelination delay item. Excluding this item reduced the difference to 0.32 (95% CI, -0.01 to 0.64). Early and TEA Kidokoro Global Brain Abnormality Score (GBAS) were negatively associated with Bayley-III cognitive and motor, but not language scores. Higher GBAS were associated with lower cognitive (early: β = -0.67 [95% CI, -1.23 to -0.11], partial R ² = 0.03; TEA: β = -0.60 [95% CI, -1.15 to -0.06], partial R ² = 0.03) and motor scores (early: β = -1.04 [95% CI, -1.67 to -0.41], partial R ² = 0.06; TEA: β = -1.18 [95% CI, -1.78 to -0.58], partial R ² = 0.08). Reduction in GBAS from early to TEA correlated with higher motor scores (β = 1.02 [95% CI, 0.03-2.02], partial R ² = 0.02). WM abnormalities were linked to poorer motor outcomes at both time points.

Kidokoro scores at early and term-equivalent time points are associated with neurodevelopmental outcomes at 24 months. Global abnormalities and WM scored at either time points, or their longitudinal changes are significantly linked to motor outcomes. Refinement of the scoring system for the early assessment may enhance clinical utility.