Department of Dentistry, section of Orthodontics and Craniofacial Biology, Radboud University Medical Center, Philips van Leydenlaan 25, 6525 EX, Nijmegen, The Netherlands.
Radboudumc 3DLab, Radboud University Medical Center, Geert Grooteplein Zuid 10, 6525 GA, Nijmegen, The Netherlands.
Clin Oral Investig. 2019 Oct;23(10):3705-3712. doi: 10.1007/s00784-018-2791-5. Epub 2019 Jan 11.
The aim of this study was to develop an accurate and intuitive semi-automatic segmentation technique to calculate an average maxillary arch and palatal growth profile for healthy newborns in their first year of life.
Seventy babies born between 1985 and 1988 were included in this study. Each child had five impressions made in the first year after birth that were digitalized. A semi-automatic segmentation tool was developed and used to assess the maxillary dimensions. Finally, random effect models were built to describe the growth and build a simulation population of 10,000 newborns. The segmentation was tested for inter- and intra-observer variability.
The Pearson correlation coefficient for each of the variables was between 0.94 and 1.00, indicating high inter-observer agreement. The paired sample t test showed that, except for the tuberosity distance, there were small, but significant differences in the landmark placements between observers. Intra-observer repeatability was high, with Pearson correlation coefficients ranging from 0.87 to 1.00 for all measurements, and the mean differences were not significant. A third or second degree growth curve could be successfully made for each parameter.
These findings indicated this method could be used for objective clinical evaluation of maxillary growth.
The resulting growth models can be used for growth studies in healthy newborns and for growth and treatment outcome studies in children with cleft lip and palate or other craniofacial anomalies.
本研究旨在开发一种准确且直观的半自动分割技术，以计算健康新生儿出生后第一年的上颌弓和腭生长曲线。
本研究纳入了 1985 年至 1988 年间出生的 70 名婴儿。每个婴儿在出生后的第一年都进行了五次数字化印模。开发了一种半自动分割工具来评估上颌尺寸。最后，建立随机效应模型来描述生长并构建 10000 名新生儿的模拟人群。对分割进行了观察者内和观察者间变异性的测试。
每个变量的 Pearson 相关系数在 0.94 到 1.00 之间，表明观察者间具有高度的一致性。配对样本 t 检验显示，除了结节距离外，观察者之间的标志点位置存在微小但显著的差异。观察者内的重复性很高，所有测量的 Pearson 相关系数在 0.87 到 1.00 之间，平均差异无统计学意义。可以为每个参数成功绘制三阶或二阶生长曲线。
这些发现表明，该方法可用于对上颌生长进行客观的临床评估。
生成的生长模型可用于健康新生儿的生长研究，以及唇腭裂或其他颅面畸形儿童的生长和治疗效果研究。
