3D quantitative analysis of normal clitoral anatomy in nulliparous women by MRI

Introduction and hypothesis We present a 3D computational approach for automated clitoral measurements. We hypothesized that computationally derived measurements would be comparable and less variable than reported manual measures. Methods In this retrospective study, MRIs of 22 nulliparous women age 20-49 years with normal vaginal and clitoral anatomy were collected. Manual segmentations were performed to reconstruct 3D models of the whole clitoris (glans, body, crura, and bulbs) and vagina. The length, width, and volume of the clitoral structures and the distance between the vagina and clitoral structures were calculated. Computed clitoral morphometrics (length, width) were compared to median [range] values from a previously published cadaver study (N = 22) using the median test and Moses extreme reaction test. Calculated distances were compared to mean (+/- SD) reported by a 2D MRI study (N = 20) using independent t-test and Levene's test. Results Overall, computed clitoral morphometrics were similar to manual cadaver measurements, where the majority of length and width measures had similar to 1-2 mm difference and had less variability (smaller range). All calculated distances were significantly smaller and had smaller SDs than manual 2D MRI values, with two-fold differences in the means and SDs. Large variation was observed in clitoral volumetric measures in our cohort. Conclusions The proposed 3D computational method improves the standardization and consistency of clitoral measurements compared to traditional manual approaches. The use of this approach in radiographic studies will give better insight into how clitoral anatomy relates to sexual function and how both are impacted by gynecologic surgery, where outcomes can assist treatment planning.