Comparison of 3- and 2-dimensional sonographic techniques for counting ovarian follicles

The purpose of this study was to compare 3-dimensional (3D) and 2-dimensional (2D) ovarian follicle counts and 3D counts using stored volumes between experienced and inexperienced operators.

Follicles larger than 5 mm were counted on 1 randomly selected ovary. Two-dimensional follicle counts were compared with stored 3D volumes by the same experienced operator (registered diagnostic medical sonographer [RDMS]). Counts using 3D stored data were compared between the experienced operator and inexperienced operator (principal investigator [PI]). The mean difference in follicle counts was computed, and a 1-sample Student t test was performed to test the hypothesis that the mean of the differences was 0. Comparison of the 2 methods and observers by Bland-Altman plots was used to determine any systematic differences based on the total number of follicles per selected ovary.

Mean differences differed from 0 (P < .005) for all 3 comparisons: 2D RDMS versus 3D RDMS, 2D RDMS versus 3D PI, and 3D RDMS versus 3D PI. For the comparison of 2D versus 3D counts done by the RDMS, 5 ovaries (10%) had a difference of more than 5 follicles counted; for the 2D RDMS versus 3D PI, 11 ovaries (22%) had a difference of more than 5 follicles; for the 3D RDMS versus 3D PI, 8 ovaries (16%) had a difference of more than 5 follicles. Mean differences in counts ranged 0.29 to 1.04 for ovaries with 10 or fewer follicles compared with 3.94 to 9.00 for ovaries with more than 10 follicles.

Follicle counts using 3D volumes were similar to 2D counts, and 3D follicle counts done by an inexperienced operator were similar to counts done by an experienced sonographer.