An analysis of pose in 3D stereophotogrammetry of the breast

Abstract

Background: Volume of the breast can be objectively measured by three-dimensional (3D) photographs. This study describes the analysis of three different positions of the patient in image acquisition, in order to find the best pose for reproducible 3D photographs of the breasts.

Method: Twenty-four patients were included between February and September 2014 in a consecutive way. Data were collected prospectively. 3D photographs were acquired using a stereophotogrammetry system. Images were taken twice in three different positions (arms behind the back, arms placed on the hips and arms horizontally placed). Surface based matching was applied and the absolute mean distance between the surfaces of both 3D models of the same position was calculated. This difference measure represents the similarity of the photographs.

Results: Univariate ANOVA showed a significant difference in distance between the three positions (sum of squares 1.12, p < 0.001). The horizontal position presented the lowest absolute mean distance (0.45 mm). Additional post hoc multiple comparisons analysis revealed a statistically significant difference between the distances of horizontal and back position (-0.22 mm, p < 0.001) and of back and hip position (0.13 mm, p = 0.009), with better results of the horizontal and hip positions, respectively.

Conclusion: Standardization of 3D acquisition in pre- and postoperative breast imaging could improve imaging reproducibility. Based on the results of this study, we recommend a pose with the arms in a horizontal position.

Keywords: 3D stereophotogrammetry; Breast surgery; Diagnostic accuracy; Three-dimensional imaging.