Estimation of a statistical geometric model for the cervical vertebrae of children aged 10-18 years

Abstract

Child neck injuries in motor vehicle crashes (MVCs) result in high morbidity and mortality rates. Estimating a statistical cervical vertebrae geometric model and quantifying the variations of the size and shape with age are very important for investigating the dynamic response and injury risk to a child's cervical spine, as well as for providing a geometric basis for developing child anthropomorphic test devices (ATDs) and finite element models (FEMs) of different ages. In this study, spatial geometric points were automatically extracted from the cervical vertebrae computed tomography (CT) scans of 30 children aged 10 to 18 years old (YO), and a statistical geometric model was estimated for the cervical vertebrae as a function of age and neck circumference/neck length according to the method of principal component analysis and regression (PCA&R). Based on this statistical model, geometric point sets representing cervical vertebrae geometries at different ages and percentiles were generated and formed to envelope surfaces. Meanwhile, the size changes of the cervical vertebrae with child growth from 10 to 18 YO were quantified. In general, the anteroposterior length (APL), transverse process width (TPW), vertebral body height (VBH), and vertebral body depth (VBD) of the cervical vertebrae increase with age; the VBH and VBD increase faster than the APL and TPW. Compared with other vertebrae, the APL of C7 is larger, and the rate of increase of C1 with age is evidently slower. The TPWs of C1 and C7 are greater than those of C2 to C6. C7 has higher average values for the VBH and VBD than C3 to C6.

Keywords: Cervical vertebra; Principal component analysis and regression; Quantification; Size/shape; Statistical geometric model.