Predicting Waist Circumference From a Single Computed Tomography Image Using a Mobile App (Measure It): Development and Evaluation Study

Abstract

Abstract Background Despite the existing evidence that waist circumference (WC) provides independent and additive information to BMI when predicting morbidity and mortality, this measurement is not routinely obtained in clinical practice. Using computed tomography (CT) scan images, mobile health (mHealth) has the potential to make this abdominal obesity parameter easily available even in retrospective studies. Objective This study aimed to develop a mobile app as a tool for facilitating the measurement of WC based on a cross-sectional CT image. Methods The development process included three stages: determination of the principles of WC measurement from CT images, app prototype design, and validation. We performed a preliminary validity study in which we compared WC measurements obtained both by the conventional method using a tape measurement in a standing position and by the mobile app using the last abdominal CT slice not showing the iliac bone. Pearson correlation, student t tests, and Q-Q and Bland-Altman plots were used for statistical analysis. Moreover, to perform a diagnostic test evaluation, we also analyzed the accuracy of the app in detecting abdominal obesity. Results We developed a prototype of the app Measure It, which is capable of estimating WC from a single cross-sectional CT image. We used an estimation based on an ellipse formula adjusted to the gender of the patient. The validity study included 20 patients (10 men and 10 women). There was a good correlation between both measurements (Pearson R=0.906). The student t test showed no significant differences between the two measurements (P=.98). Both the Q-Q dispersion plot and Bland-Altman analysis graphs showed good overlap with some dispersion of extreme values. The diagnostic test evaluation showed an accuracy of 83% when using the mobile app to detect abdominal obesity. Conclusions This app is a simple and accessible mHealth tool to routinely measure WC as a valuable obesity indicator in clinical and research practice. A usability and validity evaluation among medical teams will be the next step before its use in clinical trials and multicentric studies.