Ultrasonography measurement
After emptying the bladder, all participants underwent transvaginal ultrasonography in the dorsal lithotomy position avoiding menstrual period. Ultrasonographic examination was performed using an ultrasonography machine (VolusonTM E6 BT15, GE Healthcare Austria) and a volumetric intravaginal probe (RIC 5-9-D, frequency 5–9 MHz, 3D measurement angle 75°, PD: 0.6). Among all the subjects, the standard setting of each parameter was kept consistent. First, 2D ultrasonography was used to observe the morphology and position of the uterine body, endometrial thickness, and whether there is any abnormity in the uterine cavity. Then, the 3D energy mode was switched on, and all 3D data-sets acquisitions were performed in a standardized manner (sagittal view of the central part of the uterus; manually drawing a straight or curved line through the middle of the endometrium; maximum sweep angle of 180°; the approximate angle between the uterine axis and the ultrasound beam, 90°; ultrasound probe held steadily by the investigator, holding breath and refrain from moving by the study women during 3D volumes acquisition). Upon acquisition the 3D volume, a satisfactory coronal plane of uterus was obtained by adjusting the slice through the multiplanar views. Uteri with internal fundal indentation depth≥5mm (namely arcuate/septate uteri) were excluded. Eligible 3D datasets of each uterus were recorded and stored in hard disk of ultrasonography machine for further manipulation and analysis offline. The techniques of post processing functions for image optimization include surface rendering and volume contrast imaging(VCI). Rendering is applicable to visualizing the external serosal contour of the uterus and identifying the internal os in the coronal plane with the function of Fixed ROI(fixed region of interest) in 2D sagittal view. OmniView combined with VCI improves the image contrast between endometrial cavity and myometrium and can also allow the identification of the internal os in the coronal plane.
All measurements were taken after obtaining a satisfactory coronal view with visible bilateral horns of uterus and the internal os. A standardized set of seven measurements and their methodology was established. Detailed criteria and definitions of each morphological parameter is presented in Figure 1. The width of the uterine cavity was measured between the two cornual ostia of endometrial cavity. The length of the uterine cavity was the distance from the fundus to the anatomical internal os. The width of the internal os was the width of inferior segment of the uterine cavity running through the anatomical internal os. The surface area of uterine cavity was determined by manually outlining the entire surface of the endometrial cavity, including all the external limits and the anatomical internal os. The cornual angle was calculated by connecting the three points: the point of cornual ostia and two points on the fundus and lateral edge respectively 5mm distant from this point. The lateral indentation angle was measured using the three points including the apex of lateral indentation and two points on the lateral edge of uterine cavity 5mm distant from this point. The lateral indentation depth was defined as the distance between the line connecting the anatomical internal os and the point of cornual ostia and the parallel one across the apex of lateral indentation. The measurement method of this study was partially based on the work of Ludwin et al (16).
All patients were undergoing 3D ultrasonography measurement by two deputy directors of ultrasound doctors, who were blinded to each other’s measurements and there was no statistically significant difference in the measurements between both ultrasonography specialists. The coefficient of variation was <4.6% for uterine cavity area and <3.6% for uterine cavity width.