Hello everyone, my name is Nathan, and today I'll be presenting our work involving the comparison of 2D head photography to caliper and optical scanning for assessing craniometrics. So capturing reliable craniometric data that accurately and objectively describes a newborn head shape is important for the diagnosis, assessment, and follow-up of congenital head malformations. Current tools used to assess craniometrics have their own advantages and disadvantages. Calipers, which are readily accessible, have a high degree of user variability, while 3D laser imaging and CT scanning, which are both highly accurate, are not readily available in neurosurgical outpatient settings or require sedation and exposure of infants to radiation. And so there really isn't a great method currently that captures both the accuracy of laser imaging and CT scanning with the accessibility of calipers. In this study, we describe and validate the use of an image analysis software that calculates craniometrics from top-down orthogonal photographs. As proof of principle, we compare this approach to caliper and 3D laser scanning measurements in infants. This tool starts with obtaining a top-down photograph of the infant's head, which can be captured by any standard digital camera. Next, the software tool will segment and outline the infant's head. Using this outline, the software generates imaging masks that are then used in the quantification of conventional craniometrics. And so the final output is a list of automatically generated craniometrics that can be used to objectively assess the infant's head. To validate the use of this tool to quantify craniometrics, we compared it to caliper and 3D laser scans using interclass correlations that describe how closely two datasets resemble each other. Visually, we can represent this data using Blandt-Altman plots, with limits of agreement that show how close measurements by two methods tend to be, and a mean difference that shows any skew in the data from one method of measurement to the other. When comparing this photograph-based tool with conventional caliper measurements, we find a significant interclass correlation for the measurement of cephalic index, however no correlation in the assessment of cranial wall asymmetry index. In addition, we see a wide range in our LOAs, or limits of agreement, for both CI and CVAI, as well as little to no linear correlation between the two measurement modalities. We then compared this tool to craniometrics obtained from 3D laser scans, and found a very strong agreement in the cephalic index and a significant correlation with CVAI measurements. Additionally, our limits of agreement for both cephalic index and CVAI are much closer together, and we also see a strong linear correlation for both CI and CVAI when comparing photograph-based methods versus this optical scanning method. So in conclusion, photograph-based measurements are not well correlated with caliper measurements, possibly due to user variability in using calipers. However, craniometrics from photographs were highly correlated with those from 3D laser scans, indicating this method can provide a simple but objective and reliable way of measuring craniometrics. Widespread use of photograph-based measurements can be used for routine follow-ups and the assessment of outcomes for surgeries involving the cranial vault. And with that, I would like to acknowledge the entire Department of Neurosurgery at Connecticut Children's Medical Center and the Scientific Committee for the AANS-CNS section on Pediatric Neurosurgery.

2D head photography

caliper

optical scanning

craniometrics

image analysis software