Estimating head circumference from pediatric imaging studies an improved method

• Published in: Academic radiology Vol. 14; no. 9; p. 1102
• Main Authors: Vorperian, Houri K, Durtschi, Reid B, Wang, Shubing, Chung, Moo K, Ziegert, Andrew J, Gentry, Lindell R
• Format: Journal Article
• Language: English
• Published: United States 01.09.2007
• Subjects: Adolescent; Adult; Anthropometry - methods; Body Size; Child; Child, Preschool; Computer Simulation; Female; Head - anatomy & histology; Humans; Image Interpretation, Computer-Assisted - methods; Imaging, Three-Dimensional - methods; Infant; Infant, Newborn; Magnetic Resonance Imaging - methods; Male; Models, Biological; Pediatrics - methods; Reproducibility of Results; Sensitivity and Specificity
• ISSN: 1076-6332
• DOI: 10.1016/j.acra.2007.05.012

Head circumference (HC) is an important developmental measure used both clinically and in research. This paper advances a method to estimate HC from imaging studies when a direct HC-tape measurement cannot be secured. Unlike former approaches, the model takes into account the fact that growth is nonlinear, and that HC growth rates are sexually dimorphic. A model was first established based on published data to represent the normative HC growth curves for males and females. Then, using magnetic resonance (MR) studies of 90 subjects (birth to 18 years), a linear method to estimate HC was adapted to take into account the nonlinear and sex-specific HC normative growth curves. The accuracy of this model was tested prospectively by comparing the estimated HC with HC measurements from twelve computed tomography (CT) studies using the perimeter tracing of oblique slices that correspond to the plane at which a clinical HC-tape measurement is secured. Prospective comparison of estimated HC to HC tracings using a paired t-test validates that the model provides an accurate estimation of the measured HC (t=-.845, p=0.416 overall; t=.54, p=.615 for females and t=-2.34, p=.066 for males). HC can be calculated indirectly from imaging studies. The model is highly predictive of HC-tape measurements and provides the physician or scientist with a very reliable method to secure HC when it is not feasible to secure the HC-tape measurement.