Suitability of antral follicle counts and computer-assisted analysis of ultrasonographic and magnetic resonance images for estimating follicular reserve in porcine, ovine and bovine ovaries ex situ

This study was conducted to determine if correlations exist between the numbers of microscopic follicles comprising ovarian follicular reserve (OFR) and antral follicle counts (AFCs), and to assess the usefulness of computerized analyses of ovarian ultrasonograms and magnetic resonance (MR) images for estimating OFR in excised porcine, ovine and bovine ovaries. As a pre-requisite to these analyses, we characterized and compared ovarian cortical histomorphology and follicle populations in the three species varying in prolificacy and overall reproductive longevity, and hence the total number of microscopic and antral follicles. Ultrasonographic and MR images were obtained at the scanner settings optimized to provide opposing contrasts between antral follicles and the ovarian stroma. Commercially available ImageProPlus® analytical software was used to calculate numerical pixel values (NPVs) and pixel heterogeneity (standard deviation of the pixel values) along the computer-generated lines (4-6) placed in the area corresponding to the ovarian cortex. The numbers of primordial (r = 0.38, P < 0.01) and intermediate follicles (r = 0.37, P < 0.01) were correlated with the numbers of antral follicles in bovine ovarian sections. The numbers of primordial (r = 0.28, P < 0.05), intermediate (r = 0.31, P < 0.01) and primary follicles (r = 0.27, P < 0.05) correlated directly with mean NPVs of the ultrasonographic ovarian images in cattle. There was a negative correlation between primary follicle numbers and NPVs of MR images (3D FAST-SPOILED GRADIENT ECHO) of the porcine ovarian cortex (r = -0.31, P < 0.05). To summarize, the numbers of primordial and intermediate follicles could only be estimated from AFCs in cows. Using ultrasound NPVs, the numbers of primordial, intermediate and primary follicles could be directly estimated in bovine ovaries and the quantitative image attributes of MR images were useful for quantifying porcine primary follicles. The bovine ovarian model is compatible with human situation and hence future studies should be undertaken to ascertain the usefulness of AFCs and ultrasonographic image analyses for estimating OFR in women.

Level set segmentation of bovine corpora lutea in ex situ ovarian ultrasound images

BACKGROUND

The objective of this study was to investigate the viability of level set image segmentation methods for the detection of corpora lutea (corpus luteum, CL) boundaries in ultrasonographic ovarian images. It was hypothesized that bovine CL boundaries could be located within 1-2 mm by a level set image segmentation methodology.

METHODS

Level set methods embed a 2D contour in a 3D surface and evolve that surface over time according to an image-dependent speed function. A speed function suitable for segmentation of CL's in ovarian ultrasound images was developed. An initial contour was manually placed and contour evolution was allowed to proceed until the rate of change of the area was sufficiently small. The method was tested on ovarian ultrasonographic images (n = 8) obtained ex situ. A expert in ovarian ultrasound interpretation delineated CL boundaries manually to serve as a "ground truth". Accuracy of the level set segmentation algorithm was determined by comparing semi-automatically determined contours with ground truth contours using the mean absolute difference (MAD), root mean squared difference (RMSD), Hausdorff distance (HD), sensitivity, and specificity metrics.

RESULTS AND DISCUSSION

The mean MAD was 0.87 mm (sigma = 0.36 mm), RMSD was 1.1 mm (sigma = 0.47 mm), and HD was 3.4 mm (sigma = 2.0 mm) indicating that, on average, boundaries were accurate within 1-2 mm, however, deviations in excess of 3 mm from the ground truth were observed indicating under- or over-expansion of the contour. Mean sensitivity and specificity were 0.814 (sigma = 0.171) and 0.990 (sigma = 0.00786), respectively, indicating that CLs were consistently undersegmented but rarely did the contour interior include pixels that were judged by the human expert not to be part of the CL. It was observed that in localities where gradient magnitudes within the CL were strong due to high contrast speckle, contour expansion stopped too early.

CONCLUSION

The hypothesis that level set segmentation can be accurate to within 1-2 mm on average was supported, although there can be some greater deviation. The method was robust to boundary leakage as evidenced by the high specificity. It was concluded that the technique is promising and that a suitable data set of human ovarian images should be obtained to conduct further studies.

Promise of new imaging technologies for assessing ovarian function

Advancements in imaging technologies over the last two decades have ushered a quiet revolution in research approaches to the study of ovarian structure and function. The most significant changes in our understanding of the ovary have resulted from the use of ultrasonography which has enabled sequential analyses in live animals. Computer-assisted image analysis and mathematical modeling of the dynamic changes within the ovary has permitted exciting new avenues of research with readily quantifiable endpoints. Spectral, color-flow and power Doppler imaging now facilitate physiologic interpretations of vascular dynamics over time. Similarly, magnetic resonance imaging (MRI) is emerging as a research tool in ovarian imaging. New technologies, such as three-dimensional ultrasonography and MRI, ultrasound-based biomicroscopy and synchrotron-based techniques each have the potential to enhance our real-time picture of ovarian function to the near-cellular level. Collectively, information available in ultrasonography, MRI, computer-assisted image analysis and mathematical modeling heralds a new era in our understanding of the basic processes of female and male reproduction.

Magnetic resonance image attributes of the bovine corpus luteum during development and regression

To determine whether magnetic resonance (MR) image attributes of the corpus luteum (CL) reflect its physiologic status at different phases of the bovine ovarian cycle, we analyzed the numerical pixel values (NPVs), relaxation rates, proton densities (PDs), and apparent diffusion coefficients (ADCs) from T(1)-, T(2)-, and diffusion-weighted in vitro images and maps of the CL acquired at defined phases of luteal function. Ovaries were removed and serum samples taken on days 3, 6, and 10, or >/=17 (day 0 = ovulation), representing metestrus (n = 10), early diestrus (n = 7), mid-diestrus (n = 9), and proestrus (n = 7), respectively. Regions of interest (ROIs) in each quadrant of the CL (which occupied at least 20% of the quadrant) were selected for analysis. Three MR image slices were analyzed: the slice with the greatest cross-sectional area of the CL, and the slices acquired immediately before and after that slice. The mean NPVs of the CL in T(1)-weighted images increased at each phase from metestrus (1,104 +/- 40 msec) to early diestrus (1,119 +/- 77 msec), to late diestrus (1,206 +/- 43 msec) and proestrus (1446 +/- 80 msec; P < 0.001). The mean NPVs in T(1)-weighted images were higher in regressing CL (proestrous) than in any other phase (P < 0.002). Grayscale heterogeneity of CL in T(1)-weighted images tended to increase during regression (P < 0.07). Regressing CL (proestrus) exhibited higher T(1)-weighted mean NPVs (P < 0.01) and tended to have greater heterogeneity (P < 0.06) than growing (metestrus) and mature (diestrus) CL, even though similar progesterone concentrations were observed. The increased brightness and heterogeneity of regressing CL in T(1)-weighted images appeared to be correlated with increased connective tissue and triglyceride content and decreased vascularity. It is anticipated that diagnostic markers for luteal viability and atresia in the in vitro bovine model will be applied to in vivo studies in women.