Pedicle-dural sac congruence and its relationship to pedicle dysplasia in adolescent idiopathic scoliosis: A quantitative analysis using an image recognition-assisted gaussian process curve model

Citations

WEB OF SCIENCE

0

초록

Purpose To quantitatively evaluate pedicle–dural sac congruence and its relationship to pedicle dysplasia in adolescent idiopathic scoliosis (AIS). Methods We analyzed 64 cord-contact vertebrae (CCVs) from 52 patients with severe AIS. Using image recognition-assisted Gaussian Process (GP) curve modeling, we quantified the geometric congruence using three metrics between the pedicle wall and the dural sac on axial MRI and CT images: Procrustes disparity, Dynamic Time Warping (DTW), and Hausdorff distance. Results The concave pedicles presented significantly higher congruence with the dural sac compared to the convex pedicles across all metrics (P < 0.001). This high conformity was strongly correlated with the severity of dysplastic narrowing of the pedicle (all P < 0.001). Inter- and intra-observer agreement showed good-to-excellent reliability for all three congruency metrics, with statistical significance (all P < 0.001). The Bland-Altman plot showed a good consistency for the Procrustes disparity between MR and CT imaging modalities. Conclusions We quantitatively validated the dysplastic concave pedicle in AIS does not merely lie adjacent to the dural sac, but conforms to its geometric contour. This congruence strongly correlates with the severity of pedicle narrowing. Therefore, our results support the hypothesis that dysplastic pedicle narrowing in AIS represents a morphological adaptation, occurring in association with spatial competition with the deviated dural sac within the rotated vertebral canal.

키워드

Adolescent idiopathic scoliosis; Congruency; Pedicle dysplastic narrowing; Dural sac; Spinal cord; Cord-contacted vertebra
제목
Pedicle-dural sac congruence and its relationship to pedicle dysplasia in adolescent idiopathic scoliosis: A quantitative analysis using an image recognition-assisted gaussian process curve model
저자
Kim, Hong Jin; Yang, Jae hyuk; LEE, Hyung Rae; Lim, Su Bin; Suh, Seung Woo
DOI
10.1007/s00234-026-04243-x
발행일
2026-10
유형
Article; Early Access
저널명
Neuroradiology