Validation of Subjective Estimates of Female Breast Volume and Comparison with Objective Methods

Abstract

To date it remains unclear how well we are doing in subjectively judging breast volume. 2 experienced female examiners reviewed images of 88 female breasts and estimated breast volumes. The "true" breast volumes were previously objectively obtained by breast analysis tool (BAT) software. Tests were repeated, and Bland/Altman statistics were utilised. Mean breast volume was at 419±274 cc statistically significantly larger by 59, respectively, 75 cc than estimated breast volume by the 2 examiners. The comparison of the first and second repeated test of subjective estimation of breast volumes revealed a correlation coefficient R of about 0.92 (p<0.001) while the same comparison for the BAT measurements provided a close to perfect correlation coefficient of R>0.99 (p<0.001). With increasing volumes the repeated estimations became less reliable, not so the BAT measurements. Small breast volumes were estimated as too large and large breast volumes were estimated as too small by the examiners and the connection was nearly perfectly linear. The correlation coefficient R between estimations and measurements was lower at 0.86 than in the repeated tests and estimations showed some considerable deviations in individual cases. Subjective breast assessments underestimated breast volumes by around 70 cc, with some deviations. Breast volume determination was less reproducible by subjective estimation than by objective measurement with BAT software. With increasing breast volumes the error of the subjective estimations in the repeated tests as well as in comparison to the BAT data increased, not so the error of the BAT data. Small breast volumes were overestimated and large volumes underestimated with some considerable individual deviations. There was a good reliability of the objective measurements without much systematic error, however only limited validity of the subjective method.