In the United States, approximately 600,000 hysterectomies are performed annually. Preoperatively, uterine size is estimated through bimanual examination and ultrasound. Previous studies have compared estimated uterine volumes with histopathology. However, the bimanual exam measures uterine length. The objective is to assess the accuracy of bimanual exam (BE) and ultrasound (US) in determining uterine length.METHODS:
A retrospective review of 207 hysterectomies was performed. Uterine size from BE became clinical length (CL); uterine ultrasound length (UL) was recorded; uterine pathologic length (PL) was the “gold standard.” Scatterplots were generated for CL vs PL and UL vs PL. Linear regression assessed CL and UL as PL predictors. Paired t tests detected systematic bias of CL and UL as PL predictors. Root mean squared error (RMSE) determined the better predictor of PL.RESULTS:
Linear regression and paired t tests revealed CL and UL overestimate PL. These data suggest a guess of PL results in an average error of 5.45 cm; aided by CL this decreased to 4.75 cm; aided instead by UL this decreased to 4.07 cm. RMSE revealed the prediction error by BE or US is 4.00 cm, which is large considering average PL was less than 13 cm. Although UL is somewhat superior to CL, neither is associated with as much as 50% of total variance in PL scores.CONCLUSION:
Neither BE or US are accurate predictors of uterine length. Accurate estimation of uterine size is important in determining route of hysterectomy. Additional research needs to be conducted regarding accuracy of measurement modalities in predicting uterine size.