Laser scanning as a tool for assessment of HIV-related facial lipoatrophy: evaluation of accuracy and reproducibility.
BACKGROUND: Facial lipoatrophy (LA) is a common complication of highly active antiretroviral therapy (HAART). Research into causes and treatment of facial LA is hindered by the lack of an objective measurement tool. OBJECTIVE: To evaluate the accuracy and reproducibility of three-dimensional laser scanning (LS) for estimating cheek volume changes. METHODS: Paired laser scans were performed and the images superimposed using commercial software. The volume difference between images was computed within a circle of radius 25 mm placed in a standardized position over the cheek area. Accuracy was tested by scanning before and after known volumes of plasticine (0.5--5 mL) were applied to the cheek area of a mannequin to simulate volume change. Reproducibility was tested by repeated scanning of the mannequin with and without 2 mL of plasticine, and repeated scanning of 10 healthy subjects over the course of 1 week. RESULTS: The mean difference between actual and estimated volume change was small across the range of volumes tested [mean difference 0.08 mL; 95% confidence interval (CI)-0.36 to 0.20 mL). The coefficient of variation for repeated measurements of 2-mL volume change was 5.8%. The intraclass correlation coefficient for scan-to-scan variability was 0.812 (95% CI 0.515--0.947) and for day-to-day variability it was 0.764 (95% CI 0.332--0.935). Conclusions LS is an accurate and reproducible method for estimating cheek volume changes. It may be useful as an objective tool for assessment of facial LA in clinical research studies.