Assessing the Accuracy of a 3-Dimensional Surface Imaging System in Breast Volume Estimation

Abstract

Background 

Preoperative prediction of breast volume can guide patient expectations and aid surgical planning in breast reconstruction. Here, we evaluate the accuracy of a portable surface imager (Crisalix S.A., Lausanne, Switzerland) in predicting breast volume compared with anthropomorphic estimates and intraoperative specimen weights.

Methods 

Twenty-five patients (41 breasts) undergoing mastectomy were scanned preoperatively with the Crisalix surface imager, and 1 of 3 attending plastic surgeons provided an anthropomorphic volume estimate. Intraoperative mastectomy weights were used as the gold standard. Volume conversions were performed assuming a density of 0.958 g/cm³.

Results 

The Pearson correlation coefficient between imager estimates and intraoperative volumes was 0.812. The corresponding value for anthropomorphic estimates and intraoperative volumes was 0.848. The mean difference between imager and intraoperative volumes was −233.5 cm³, whereas the mean difference between anthropomorphic estimates and intraoperative volumes was −102.7 cm³. Stratifying by breast volume, both surface imager and anthropomorphic estimates closely matched intraoperative volumes for breast volumes 600 cm³ and less, but the 2 techniques tended to underestimate true volumes for breasts larger than 600 cm³. Stratification by plastic surgeon providing the estimate and breast surgeon performing the mastectomy did not eliminate this underestimation at larger breast volumes.

Conclusions 

For breast volumes 600 cm³ and less, the accuracy of the Crisalix surface imager closely matches anthropomorphic estimates given by experienced plastic surgeons and true volumes as measured from intraoperative specimen weights. Surface imaging may potentially be useful as an adjunct in surgical planning and guiding patient expectations for patients with smaller breast sizes.