Primary tooth development is shaped by prenatal experience and has consequences for childhood caries and lifelong oral health. However, evidence linking prenatal conditions to dental development has relied largely on postnatal proxies, such as parent-recalled eruption timing, that conflate prenatal programming with postnatal exposures. Here, we directly phenotype developing dentition in vivo using routinely acquired neonatal brain MRI. Using T2-weighted imaging from the HEALthy Brain and Child Development Study, we trained a 3D U-Net on 100 semi-automatic labels and applied the model to 1,433 quality-controlled neonatal scans. Automated post-processing extracted quantitative features of tooth volume, mineralization, and arch geometry. These features were used to predict postmenstrual age at MRI and to derive a bias-corrected tooth age gap (TAG), indexing relative dental maturity at birth. The segmentation model achieved mean cross-validated Dice = 0.94. Dental features predicted postmenstrual age with R2 = 0.30 and mean absolute error = 6.7 days, outperforming standard anthropometric measures (R2 = 0.22). In adjusted models, TAG varied by infant sex and was associated with gestational age at delivery, maternal pre-pregnancy BMI, and prenatal tobacco use. In infants with longitudinal dental follow-up, greater dental maturity at birth predicted earlier first tooth emergence and more teeth at one year. Automated segmentation and age prediction generalized to an independent cohort. These findings establish neonatal dental MRI phenotyping as an objective, scalable index of dental maturation and a potential readout of prenatal influences on a developmental system relevant to lifelong oral health.
Meng, Y., O'Connor, T. G., Bidlack, F., Simmons, S. A., Xiao, J., Rasmussen, J.
Advertisement
Stats
- Recommendations n/a n/a positive of 0 vote(s)
- Views 9
- Comments 0