Dynamic MRI of pelvic floor morphology in young, healthy Asian nulligravid women: an exploratory study across rest, Valsalva, and voluntary contraction

Pelvic floor support involves passive connective tissue and active levator ani mechanisms. Although previous studies have quantitatively assessed pelvic floor morphology during voluntary contraction, the state-related behavior of multiple pelvic landmarks and the uterovaginal axis remains incompletely characterized in young, healthy, nulligravid women. This exploratory prospective imaging study included 16 healthy Asian nulligravid women aged 18–30 years. Pelvic 3.0T dynamic MRI was obtained at rest, during instructed maximal Valsalva, and during instructed maximal voluntary contraction. Midsagittal coordinate landmarks, angular measures, length and area parameters, and conventional parameters were measured in Mimics using a Pelvic Inclination Correction System-based coordinate approach. Two radiologists performed independent measurements; interobserver reliability was assessed using ICC(2,2) and Bland-Altman analysis, and the mean of the two readers’ values was used for subsequent analyses. State-related differences were evaluated using repeated-measures ANOVA with Greenhouse–Geisser correction when required. Benjamini–Hochberg correction was applied across 33 endpoint-level omnibus tests, and Friedman and paired Wilcoxon analyses were performed as sensitivity analyses. ICC(2,2) point estimates ranged from 0.858 to 0.999 across the three states, although several 95% confidence intervals were wide and extended below 0.75, particularly during Valsalva and contraction. After Benjamini–Hochberg correction, 32 of 33 endpoints showed exploratory evidence of state-related differences; nonparametric sensitivity analyses were not fully concordant. The largest differences were generally observed between instructed maximal Valsalva and instructed maximal voluntary pelvic floor contraction. For contraction minus Valsalva, the mean differences were 18.50 mm (95% CI, 13.07–23.93) for ARJ-Y, − 30.44° (95% CI, − 38.29° to − 22.59°) for ULP angle, − 12.69 mm (95% CI, − 15.30 to − 10.07) for M-line displacement, and − 2.41 cm² (95% CI, − 3.64 to − 1.18) for LHA. Mean landmark trajectories suggested possible segment-dependent differences in displacement along the uterovaginal axis during contraction. In this small cohort of young, healthy Asian nulligravid women, dynamic MRI suggested distinct maneuver-associated pelvic floor morphological configurations between instructed maximal Valsalva and instructed maximal voluntary pelvic floor contraction. These preliminary exploratory physiological observations are hypothesis-generating and require validation in larger, more diverse cohorts using three-dimensional imaging and concurrent functional measurements.