| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
a Departments of Obstetrics, Gynecology and Reproductive Sciences and
b Medical Imaging,
c College of Medicine, and Veterinary Medical Biosciences, Western College of Veterinary Medicine, University of Saskatchewan, Saskatoon, Saskatchewan, Canada S7N 0W8
ABSTRACT
We analyzed image characteristics in T1-, T2-, and diffusion-weighted in vitro magnetic resonance (MR) images acquired at predefined stages of the ovarian cycle in 36 heifers to test the hypothesis that MR image attributes of the follicle wall reflect the physiologic status of ovarian follicles (viable, atretic, dominant, subordinate). Numerical pixel values (NPV), standard deviation of pixel values (heterogeneity), and area under the curve were used to assess images of follicle walls. Pixel values of the wall were used to calculate a regression line from which intercept, slope, and coefficient of determination were calculated. In T1 images, NPV of dominant follicles were less likely to fit a regression line at the preovulatory phase than at any other phase (P < 0.1). Preovulatory dominant follicles had lower area under the curve in diffusion-weighted images than early and late static dominant follicles of the anovulatory wave (P < 0.02). Subordinate follicles in the presence of a preovulatory dominant follicle had lower mean NPV in T1- and T2-weighted images and lower intercepts in T1-weighted images than subordinate follicles of the anovulatory wave (P < 0.02). Early atresia of dominant follicles was identified at the late static phase by greater area, mean NPV, and slope in T2-weighted images (P < 0.02). Preovulatory dominant follicles had poor fit of NPV to a regression line in T1-weighted images and lower area under the curve in diffusion images. Atretic follicles had brighter walls with more acute transitions from follicular fluid to stroma in T2-weighted images and more heterogeneous walls in diffusion images. The MR image attributes of the follicle wall reflected the physiologic status of dominant and largest subordinate follicles.
follicle, follicular development
INTRODUCTION
The follicle wall consists of the stratum granulosum, basement membrane, and the vascularized theca interna. Cells of the follicle wall function in response to gonadotropins to produce secretory products like mucopolysaccharides as well as steroid and protein hormones. The follicle antrum is a reservoir for products of the follicle wall and the plasma transudate that filters through the wall [1]. Predictors of ovarian follicular viability or atresia have been identified in ultrasonographic images of bovine and human follicles under both normal and induced ovulation conditions [2]. Previous research using ultrasonographic imaging and computer-assisted image analysis has shown pronounced differences in image attributes of the follicle wall among phases of development in bovine ovaries [3]. To our knowledge, however, magnetic resonance (MR) image attributes of the ovarian follicle wall have not been quantified, nor have they been examined systematically at specific phases of follicular development and regression.
Magnetic resonance images are produced from radiofrequency (RF) signals emitted by protons (hydrogen nuclei) in the body while in a large magnetic field. Image contrast is produced based on the differing proton density (PD) and differing signal decay rate (also called relaxation rates) of different tissues. The two main mechanisms of signal decay, T1 (or spin-lattice) relaxation and T2 (or spin-spin) relaxation, were manipulated to produce varied image contrasts in the present work. Image contrast was dominated by either T1 (to give T1-weighted images), T2 (to give T2-weighted images), or PD (to give PD-weighted images). Diffusion-weighted images were produced that contained contrast based on the self-diffusion (Brownian motion) of water within tissues and fluids with further manipulation of the magnetic fields during imaging.
Generally, tissues with high concentrations of solutes, high viscosity, or low concentrations of fluid will have short relaxation rates, because more pathways are available for the energy exchange required for relaxation [4]. The more water-like a fluid, the longer the T1 and T2 relaxation rates. Brighter areas in an MR image (areas of higher pixel values) will have higher signal intensity and correspond to a shorter T1 relaxation rate for T1-weighted images and a longer T2 relaxation rate for T2-weighted images. Tissue with a high PD will appear bright in both T1- and T2-weighted images.
Magnetic resonance imaging has been identified as a powerful tool for studying ovarian physiology [4–6]. Ovarian structures, including follicles and corpora lutea, can be visualized with unmatched anatomic resolution [4]. Generally, the follicle antrum appears dark in T1-weighted images and bright in T2-weighted images because of the fluid content [4, 7]. When the follicle wall can be differentiated from the stroma, it appears bright in T1-weighted images and dark in T2-weighted images [7]. Magnetic resonance imaging using fast imaging with steady-state precession and maximum-intensity projection reconstruction can display the distribution of follicles and their spatial relationship to luteal structures in a three-dimensional frame [4]. In a companion paper, MR image attributes of the antrum (follicular fluid component) of bovine ovarian follicles at specific phases of development and regression exhibited differences reflective of the physiologic status of the follicles [6].
The bovine model provides an excellent paradigm for studying ovarian function [8]. Bovine follicular kinetics are well documented from ultrasonographic studies, and image-analysis studies based on ultrasonography have been directly applicable to human research and clinical medicine [8]. In cattle, follicular development occurs in two or three waves per estrous cycle [9–14]. As a cohort of follicles from a follicular wave grows, one follicle is physiologically selected for preferential growth and functional dominance. The rest of the recruited cohort, which do not grow to the same degree, become atretic and regress. The dominant follicle of the first wave is anovulatory under normal conditions, because production of progesterone by the corpus luteum inhibits the LH surge and ovulation. A new follicular wave emerges as the anovulatory dominant follicle regresses, and a new dominant follicle is selected. The corpus luteum regresses at approximately Day 17, allowing for a surge of LH at approximately Day 20 to induce ovulation of the dominant follicle of the second or third wave. Anovulatory follicles can be identified with ultrasonography at the growing (increasing in diameter), early static, late static (no longer increasing in diameter), and regressing phases (decreasing in diameter), whereas ovulatory follicles have only a growing phase [12, 14]. Ultrasonographically classified phases of development and regression have been shown to reflect follicular function (steroid and protein hormone production) and health (viability and atresia) [3, 15–18].
The present study and its companion studies were designed to provide information regarding quantitative attributes of MR images of follicles with known status using in vitro MR imaging. The objective of the present study was to characterize the follicle wall of dominant and subordinate follicles in T1-, T2-, and diffusion-weighted images acquired at defined phases of the bovine ovarian cycle. Our hypothesis was that the walls of dominant and largest subordinate follicles would exhibit quantitative differences in MR image attributes reflective of their physiologic status (viability or atresia, dominant or subordinate).
MATERIALS AND METHODS
The assignment of animals to various groups, ultrasonographic methods, and ovariectomies were conducted as described in the study by Singh et al. [3], which was based on ultrasonographic image analysis. The study was conducted in accordance with guidelines for the ethical use of animals in research outlined by the Canadian Council of Animal Care. The follicles analyzed in the present study were analogous to the follicles described in a companion paper [6].
Experimental Animals: Grouping and Ultrasonography
Thirty-five sexually mature, nulliparous heifers were used in two replicates (replicate 1 = 16 heifers, replicate 2 = 19 heifers). The development of ovarian follicles was monitored by transrectal ultrasonography using a 7.5-MHz, linear-array transducer (Aloka SSD 500; ISM, Inc., Edmonton, AB, Canada). Transrectal ultrasound examinations were performed daily to monitor the development of follicles 4 mm or greater in diameter commencing at least 2 days before the ovulation preceding the beginning of the estrous cycle under study, and these examinations continued until the day of ovariectomy [3]. Retrospective analysis of hand-drawn diagrams (maps) showing topographical location and diameter of individual follicles and corpora lutea was used to determine the day of wave emergence (Day 0). The day of wave emergence was defined as the day on which the dominant follicle was first detected at a diameter of 4–5 mm [3, 8, 12, 14]. The dominant follicle was identified as the largest follicle of a wave; remaining follicles were identified as subordinate follicles [3, 12, 14]. The heifers were designated randomly for ovariectomy on Day 3 of wave one (D3W1; n = 10), Day 6 of wave 1 (D6W1; n = 9), Day 1 of wave 2 (D1W2; n = 9), or in the immediate preovulatory period at least 17 days after ovulation (D
17; n = 8) [3]. On D1W2, the largest of the preselection follicles was selected to be the dominant follicle of wave 2. Heifers in the D17 group were ovariectomized 1 day after detection of proestrus [3]. Proestrus was defined as the day when any three of four estrus-like characteristics were detected (high uterine tone, edematous echotexture, intrauterine fluid collection, and mucous discharge) [3, 9]. Days of ovariectomy were presumed to represent the growing (D3W1), early static (D6W1), late static (D1W2), and regressing phases (D17) of dominant follicles of wave 1 as well as the preselection (D1W2) phase and the postselection growing phase (D17) of the preovulatory dominant follicle. Subordinate follicles were also available for analysis on the selected days of ovariectomy.
Ovariectomy
Ovaries were removed from the heifers with a single incision through the dorsolateral aspect of the vaginal wall [19]. Surgery was conducted with cows in the standing position and under caudal epidural anesthesia using 2% (w/v) lidocaine HCl with 0.001% (w/v) adrenaline. Clenbuterol (0.6 µg kg-1 body weight; Boehringer Ingelheim Ltd., London, ON, Canada) was given i.v. 10 min before colpotomy to induce relaxation of the ovarian ligament. The ovarian ligament was compressed with a lidocaine-soaked gauze immediately before the ovaries were excised using a chain ecraseur looped around the ovarian ligament. After ovariectomy, ovaries were placed immediately in warm (37°C) physiologic saline, transported to the MR suite in an insulated container, and imaged within 60 min.
Hormonal Confirmation of Endocrine Status
Blood samples were collected on the day of ovariectomy, and serum was stored at -20°C. Follicular fluid was aspirated after MR imaging was completed. Follicular fluid samples were collected from the dominant and subordinate follicles; fluid from subordinate follicles was pooled within animals. All follicles other than the morphologically dominant follicle made up the pooled subordinate follicular fluid fraction. Follicular fluid was also stored at -20°C. Hormonal data are reported in a companion paper [6].
MR Imaging
Imaging was performed on four ovaries (the ovaries from two heifers; n = 31) on or two ovaries (the ovaries of one heifer; n = 4) at a time in a small plastic dish, with a divider separating the ovaries of the two heifers. The dish contained only the ovaries and no surrounding material. Imaging was done using a 1.5-Tesla SP Magnetom MR imager (Siemens, Erlangen, Germany) with a Helmholtz RF receiver coil (diameter, 150 mm). Image contrast was manipulated through the variation of the data acquisition parameters of repetition time (TR) and echo time (TE) using standard methods [20]. Images were acquired using T1 and PD sequences (TR/TE = 480/15, 1000/15, 2000/15, and 4000/15 msec), 16-echo T2 sequences (TR/TE = 2000/20–245 msec), as well as unweighted and weighted diffusion sequences (b = 0 and 17 776 sec/cm2, in the z-direction). The T1 sequences at TR/TE = 2000/15 and 4000/15 msec and diffusion sequences were added for the second replicate (19 of the 35 heifers used in this study) to obtain more accurate T1 maps and to add diffusion maps to the analysis (D3W1, n = 5; D6W1, n = 5; D1W2, n = 5; D17, n = 4). The T1 and T2 protocols were the same as the standard protocols found to provide opposing contrasts for the follicles and corpus luteum versus the stroma [4].
The T1, T2, and PD data were acquired with a field of view of 120 x 120 mm for the first replicate and 100 x 100 mm for the second replicate. Both replicates had an acquisition matrix size of 160 x 256 and a section thickness of 2 mm. Weighted and unweighted diffusion data were acquired with a field of view of 100 x 100 mm, an acquisition matrix size of 128 x 256, and a section thickness of 2 mm. All MR data were Fourier-transformed into 256 x 256 pixel images, with a pixel resolution of 0.47 or 0.39 mm depending on the field of view.
Image Analysis of the Follicle Wall
A subset of MR images was selected from the image sets with the best follicle-to-stroma image contrast; T1- (TR/TE = 480/15 msec), T2- (TE/TR = 2000/50 msec), and diffusion-weighted images were used for analysis of the follicle wall. Images that contained the greatest cross-sectional area of the follicle in question were selected. The selected images were linearly converted from 12- to 8-bit images and saved in a database. Images were analyzed by assessing numerical pixel values (NPV; 0 = black and 255 = white, with 254 shades of gray in between) along a line traversing the follicle wall from the antrum to the stroma using Synergyne 1 version 2.8 (R.A. Pierson, Saskatoon, SK, Canada) on an Ultra 2 SunSparc graphics workstation (Sun Microsystems, Palo Alto, CA). The follicle walls were analyzed at the 3- and 12-o'clock positions, and the values were averaged for further analysis. The opposite pole was used if one of the aforementioned positions bordered the outside of the ovary. Values reported are for the section of the line corresponding to the follicle wall from the antrum-wall interface to the wall-stroma interface. Interfaces were identified as acute changes in pixel intensity.
Wall thickness was measured as the length of the line across the follicle wall. Mean NPV, standard deviation, and area under the curve were calculated from the profile of pixel values along the line. Standard deviations were used as a measure of pixel heterogeneity, or pattern of signal intensities in the wall. A higher standard deviation indicated a more heterogeneous wall. Area under the curve or profile was the integral of pixel intensities across the follicle wall. The NPV across the follicle wall were used to calculate a regression line (antrum-follicle wall interface as the zero-point) from which the intercept, slope, area under the curve, and coefficient of determination were calculated [2, 3, 21]. The intercept was the numerical pixel value corresponding to the antrum-wall interface on the linear regression line, and the slope was an indicator of relative change in NPV of successive pixels from the antrum to stroma along the follicle wall [3]. The coefficient of determination was an indicator of how well the fitted regression line fit the profile (curve) of pixel values along the line though the follicle wall.
Statistical Analyses
All quantitative end points were compared between replicates 1 and 2 using Student t-test. No statistically significant differences were found; therefore, data were combined. Two-factor ANOVA was used to compare MR image attributes and hormone concentrations of follicles of different type (dominant or largest subordinate) and phase of development (growing, early static, late static, regressing, preselection, and preovulatory) using SPSS version 9.0 (Statistical Product and Service Solutions, Chicago, IL). Protected least-significant-difference (LSD) tests were performed to determine differences between specific groups when significance (P 
0.05) was shown for follicle type, follicle phase, or type-by-phase interaction [22]. Significance values for comparisons between specific groups reported in the text were taken from the protected LSD. Image attribute end points were analyzed among dominant and largest subordinate follicles. Only the largest of the subordinate follicles was analyzed, because ultrasonographic end points were found to be similar between the largest and second-largest subordinate follicles [3]. Pearson correlation coefficients between hormone data and MR image attributes were calculated by matching the values of individual follicles without categorization by type or phase. Only significant correlation coefficients greater than 0.25 are reported. Results are reported as the mean ± SEM.
RESULTS
Follicle Status
The follicle diameter profiles, which confirmed that the anatomic follicular status of the dominant and largest subordinate follicles was appropriate for the days of ovariectomy, are reported in the companion paper [6].
Appearance of the Follicle Wall in MR Images
The follicle wall is shown in T1-, T2-, and diffusion-weighted images (Fig. 1). The follicle wall was differentiated from the stroma in all follicles analyzed in the present study. In T1-weighted images, the follicle wall appeared as a bright band around the dark follicle antrum (Fig. 1a). In T2- and diffusion-weighted images, the follicle wall was observed as a dark band around the bright follicle antrum (Fig. 1, b and c).
|
Line Analysis of the Follicle Wall
The line analysis printouts of the profile of pixel values along the wall and the values calculated from the regression are shown in Figure 1, d–f. The MR image attributes of the wall are shown graphically with the first four groups, representing growing and early static dominant and largest subordinate follicles and late static and regressing dominant follicles of wave 1, followed by the preselection and preovulatory dominant follicles of the ovulatory wave in the last two groups with their corresponding largest subordinate follicles (Figs. 2–4). In T1-weighted images, the walls of preovulatory dominant follicles tended to have the lowest coefficient of determination values (P < 0.1; Fig. 2c). The walls of preselection follicles at D1W2 had higher mean NPV and intercepts in T1-weighted images than preovulatory dominant follicles (D17; P < 0.004) and the growing and early static follicles of the first wave (D3W1 and D6W1; P < 0.02). The subordinate follicle in the presence of a preovulatory dominant follicle (D17) had lower NPV than subordinate follicles of the anovulatory wave at D3W1 (55.7 ± 3.8 vs. 69.3 ± 2.9, P < 0.02) and lower intercepts than the subordinate follicles of the anovulatory wave at D3W1 and D6W1 (51.1 ± 3.5 vs. 63.5 ± 2.8 and 63.3 ± 5.5, respectively; P < 0.04).
|
In T2-weighted images (Fig. 3), area under the profile of pixel values across the wall of the dominant anovulatory follicle was higher in the late static follicles (D1W2) than in the growing and early static follicles of wave 1 (D3W1 and D6W1) and the preovulatory follicle (D17; 930 ± 75 vs. 680 ± 30 and 719 ± 30 and 760 ± 80, respectively; P < 0.02). Mean NPV and intercepts for the wall profile of the anovulatory dominant follicle were higher in the late static phases (D1W2) than in growing and early static phases (D3W1 and D6W1; NPV: 134.9 ± 7.1 vs. 103.0 ± 5.5 and 109.1 ± 7.9, respectively, P < 0.03; 181.6 ± 8.3 vs. 141.5 ± 4.1 and 145.7 ± 7.9, respectively, P < 0.01). The wall of the anovulatory follicle had an increasingly negative slope as the follicle regressed in the late static and regressing phases (D1W2 and D17 of wave 1; P < 0.05). Area, NPV, slopes, and intercepts of the preselection follicle walls (D1W2; ovulatory wave) were higher than the values for the preovulatory dominant follicle walls (D17; P < 0.04). The subordinate follicle in the presence of a preovulatory dominant follicle had lower NPV than the subordinate follicles from the anovulatory wave (P < 0.02). Images of subordinate follicles exhibited higher T2 wall NPV, heterogeneity, and intercept than the dominant follicles at D3W1 and D6W1 (P < 0.04).
|
In diffusion-weighted images (Fig. 4), preovulatory dominant follicles (D17) had lower area under the curve of pixel values across the follicle wall than early and late static dominant follicles (D6W1 and D1W2; 522 ± 33 vs. 676 ± 20 and 673 ± 37, respectively; P < 0.007). Preselection follicles at D1W2 had greater pixel heterogeneity, slopes, and intercepts than growing (D3W1), early static (D6W1), regressing (D17, wave 1), and preovulatory dominant follicles (D17, ovulatory wave) and greater area under the curve than observed in growing (D3W1), regressing (D17; wave 1), and preovulatory follicles (D17, ovulatory wave; P < 0.04). Pixel heterogeneity of dominant anovulatory follicles tended to increase from growing and early static phases to the late static phase (P = 0.06) and then decreased in the regressing phase (P < 0.02).
|
Correlations
The intercepts of the regression line of pixel values across the wall in T1-, T2-, and diffusion-weighted images were positively correlated with serum progesterone concentration (r = 0.42, 0.61, and 0.48, respectively; P < 0.04). Mean NPV, pixel heterogeneity, and intercept of the regression line across the wall in T2-weighted images were negatively correlated to follicular fluid estradiol concentration (r = -0.33, -0.30, and -0.33, respectively; P < 0.02). In contrast, mean NPV, pixel heterogeneity, and intercept of the regression line across the wall in T2-weighted images were positively correlated with serum progesterone concentration (r = 0.50, 0.57, and 0.61, respectively; P < 0.006). The slope of the regression line in T2-weighted images positively correlated with follicular fluid estradiol concentration (r = 0.36, P = 0.007) and negatively correlated with serum progesterone concentration (r = -0.40, P = 0.03). Pixel heterogeneity across the wall in diffusion-weighted images was positively correlated with serum progesterone concentration (r = 0.66, P = 0.002).
DISCUSSION
The results of this study supported the hypothesis that the walls of dominant and largest subordinate follicles exhibit quantitative differences in MR image attributes reflective of their physiologic status (viability or atresia, dominant or subordinate). Pixel values taken across the walls of healthy preovulatory follicles in T1-weighted images were less likely to fit the calculated regression line. In diffusion-weighted images, viable follicles also had lower area under the curve or profile of pixel values along the wall. Subordinate follicles, in the presence of the preovulatory dominant follicle, exhibited lower T1- and T2-weighted quantitative signal intensity (NPV) and lower T1-weighted intercepts than subordinate follicles of the anovulatory wave. Early atresia of dominant follicles of the anovulatory wave was identified by higher values for some T2- and diffusion-weighted image attributes of the follicle wall.
Similar to the results observed examining ultrasonographic image attributes of bovine ovarian follicles in previous studies, differences in T2-weighted MR image attributes of dominant follicle walls were observed between early and late static phases when loss of functional dominance occurs [3]. In ultrasound images, heterogeneity of pixel values comprising the follicle wall and slope of the regression line of pixel values at the fluid-follicle interface increased between early and late static phases [3]. In T2-weighted MR images, the walls of late static follicles were brighter and had a greater area under the curve and slope of the regression line than growing and early static follicles. In diffusion-weighted images, follicle walls at the late static phase appeared more heterogeneous than those of early static follicles. The MR image attributes that reflected the early atretic change in dominant follicles, from early to late static phases (T2 area, mean NPV, and slope), were significantly correlated with serum progesterone and follicular fluid estradiol concentrations [6].
Increased heterogeneity of the follicle wall in diffusion-weighted images might be attributable to the disorganization of the granulosa cells as they lose the ability to produce steroid hormones, become atretic, and are sloughed into the antrum during the late static phase [3, 23]. The brighter walls in T2-weighted images of late static follicles suggest a less cellular, more water-like composition and might be attributed to the atresia-related disorganization at the antrum-follicle wall interface. Similarly, in ultrasound images, late static and regressing dominant follicles characteristically had brighter walls than healthy follicles [3]. The increase in area under the profile of pixel values across the follicle wall was due to brighter walls at the late static phase. A greater change in signal intensity from the antrum-wall interface to the wall-stroma interface was observed in late static follicles and regressing follicles than was observed in growing and early static follicles. In ultrasound images, regressing follicles had steeper slope and higher intercept values, which were attributed to thinner walls of regressing follicles [3].
At D3W1 and D6W1, atretic subordinate follicles had brighter walls and higher intercepts in T2-weighted images than their dominant counterparts. Regressing subordinate follicles also had greater area under curve of pixel values across the follicle wall in T2-weighted images at D3W1 than dominant follicles at D3W1. Similarly, in ultrasound images, the follicle wall was brighter in subordinate follicles than in the corresponding dominant follicles [3]. The walls of atretic subordinate follicles showed similar MR image attributes reflective of atresia (higher T2-weighted NPV, area, and intercept) than observed in the walls of the late static dominant follicles. Differences in MR image attributes observed between the dominant and largest subordinate follicles most likely reflected the atretic changes occurring in subordinate follicles at D3W1 and D6W1. Histologically, at D3W1 and D6W1, subordinate follicles have thinner walls and decreased mitotic activity of the granulosa and theca interna than their dominant counterparts [12]. In the present study, the fluid of subordinate follicles had higher levels of progesterone and lower estradiols concentration than those observed in the fluid of the dominant follicles at D3W1 and D6W1, indicating differences in hormone production between dominant and subordinate follicles that were reflected in the MR image attributes of the follicle wall [6].
The preovulatory, healthy growing status of the dominant follicle can be recognized by the MR image attributes of the largest subordinate counterpart. Subordinate follicles in the presence of a preovulatory dominant follicle appeared darker in T1- and T2-weighted images and had lower T1 intercepts than subordinate follicles of the anovulatory wave. In the companion study, the follicular fluid of subordinate follicles in the presence of a preovulatory dominant follicle exhibited different MR image attributes than the fluid from subordinate follicles of the anovulatory wave [6]. Subordinate follicles in the presence of a preovulatory follicle also have different ultrasound image attributes than the subordinate follicles of the anovulatory wave [3]. The unique MR image attributes of the subordinate follicles in the preovulatory state were not significantly different from those of the preovulatory dominant follicle; therefore, differences most likely reflected a change in the follicles caused by a systemic mechanism. The preovulatory dominant follicle produces high levels of estradiol that induce an LH surge, which may have a systemic effect on both ovaries, including the subordinate follicles. Alternatively, subordinate follicles from the same wave as the viable dominant follicle might have had a more inhospitable environment than that experienced by the subordinate follicles from the anovulatory wave and, thus, appeared more atretic. In this case, similarities in MR image attributes observed between the preovulatory dominant and regressing subordinate follicles could only be explained by preovulatory changes in the dominant follicle.
Preovulatory dominant follicles characteristically had lower coefficient of determination values in T1-weighted images than dominant follicles at any other phase. That is, pixel values along the wall of viable preovulatory dominant follicles in T1-weighted images were the least likely to fit the regression line. The preovulatory growing dominant follicles also had lower area under the curve of pixel intensities across the follicle wall than growing and early static dominant follicles of the anovulatory wave. The unique MR image attributes of the preovulatory dominant follicle wall can probably be attributed to the unique histological appearance, in that it has the thickest wall, the lowest cell density in the granulosa and theca layers, and the highest vascularity and edema in the theca interna compared to other follicles [23]. In ultrasound images, regression lines across the wall of the preovulatory dominant follicle characteristically had the lowest slope [3].
Dominant follicles with NPV across the walls that did not fit a linear regression and whose subordinate counterparts appeared as dark as the dominant follicle in T1-weighted images were viable preovulatory dominant follicles. In contrast, follicles that had high signal intensity and steep changes in signal level across the wall in T2-weighted images were atretic. Preovulatory dominant follicles could be identified by examining the coefficient of determination value of the wall in T1-weighted images and the area under the curve of pixel values across the follicle wall. In addition, the brightness of the walls of the largest subordinate follicles in T1- and T2-weighted MR images could be used to determine whether the corresponding dominant follicle from the same wave was a healthy preovulatory dominant follicle. The walls of regressing dominant and subordinate follicles appeared brighter and tended to be more heterogeneous.
In summary, we were able to quantify the MR imaging appearance of ovarian follicle walls at specific phases of development and regression. Changes in MR image attributes coincided with changes in steroid hormone production and ovulatory potential of the follicles [6]. Image attributes of the follicle wall in T1-, T2-, and diffusion-weighted MR images could be used to differentiate healthy from atretic follicles, and analysis of MR images revealed both follicle-type and phase-specific differences in the follicle wall. The MR images offered additional physiological information about follicles not available with ultrasonography in addition to superior contrast resolution of ovarian and other reproductive structures. However, further study is needed to elucidate the exact physiological significance of the relaxation rate data. The results of this study on excised bovine ovaries provides further support for continued research into acquiring high-resolution MR images of human ovarian follicles for computer-assisted analysis using a noninvasive, in vivo approach. Results of this study must be confirmed in vivo and under superstimulation conditions, but analysis of MR image attributes of ovarian follicle walls shows promise for use as a diagnostic tool to assess the physiologic status of ovarian follicles in women as they develop and to predict which follicles will respond to ovulation-inducing agents.
ACKNOWLEDGMENTS
We thank Drs. Julio Tegli and Rob McCorkell for their surgical assistance in acquiring the ovaries, John Deptuch for computer-programming assistance, and Susan Cook for her help with the radioimmunoassays.
FOOTNOTES
First decision: 22 January 2001.
1 Supported by grants from the Canadian Institutes for Health Research (R.A.P./G.E.S.) and the Natural Sciences and Engineering Research Council of Canada (G.P.A.). 
2 Correspondence: R.A. Pierson, Department of Obstetrics and Reproductive Sciences, College of Medicine, University of Saskatchewan, Royal University Hospital, 103 Hospital Drive, Saskatoon, SK, Canada S7N 0W8. FAX: 306 966 8796; pierson{at}erato.usask.ca
Accepted: May 21, 2001.
Received: December 2, 2000.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and largest subordinate follicles (
) at the four examined phases of development and regression are shown. Values under the x-axis indicate the number of follicles analyzed for each follicle type. When overall significance was found by phase, group, or in group by phase, the bars with no common letters indicate significantly different values (P < 0.05)
