OBJECTIVE: The objective of this study was to assess the reliability and reproducibility of existing and new computed tomography (CT)-pelvimetry measurements.METHODS: A retrospective cohort study of 63 women with a mean age of 33.9±5.2 (SD) years (range: 19-49 years) was conducted. Classical pelvimetry measurements were collected including the obstetric conjugate (OC), median transverse diameter (MTD), and interspinous diameter (ISD). Additionally, we used multiplanar reconstruction (MPR) mode to define two oblique planes: inlet pelvic plane (IPP) and mid-pelvic plane (MPP) and measure new pelvic parameters, including anteroposterior (APD), transverse diameters and circumference of both IPP and MPP (inletAPD, inletMTD, inletCIRC and midAPD, ISD, midCIRC, respectively). The reproducibility (intra- and inter-observer) of our results were assessed. Multivariate analyses using principal component analysis and clustering methods were conducted to analyze the association between pelvimetry measurements and identify patient sub-groups.RESULTS: All linear measurements (OC, inletAPD, MTD, inletMTD, midAPD, and ISD) showed statistically "almost perfect" intra- and inter-observer correlation coefficients (range: 0.924-0.980). Circumferences (inletCIRC and midCIRC) showed statistically "almost perfect" intra- (range: 0.847-0.857) and inter-observer correlation coefficients (range: 0.923-0.957). The measurement of 6 pelvimetric parameters allowed determining three groups of pelvis size.CONCLUSIONS: New pelvic measurements have excellent reproducibility and are similar to the classical measurements, based on the MPR analysis of CT planes adjusted to the inner bony pelvis.