Luteal and follicular count in bitches: assessment by means of magnetic resonance imaging.

This study was done to determine whether preovulatory follicles or corpora lutea (physiological structures, PS) can be counted in the ovaries of bitches by means of magnetic resonance imaging (MRI). In Experiment 1, the ovaries from 15 German shepherd bitches (five in the follicular phase, one in the periovulatory period, five during the first 38 days of diestrus and four between Day 48 of diestrus and full-term gestation) were embedded in gelatin to form three phantoms with 10 ovaries each. Each phantom was exposed to MRI, using a 1mm slice thickness, a 1mm slice interval, a voxel size of 1mm cubic and a variety of pulse sequences, whereafter the ovaries were dissected and the numbers of follicles, corpora lutea and cysts counted. T2-weighted images were superior to T1-weighted images. Each of three operators counted the numbers of PS and cysts on T2-weighted images obtained in the coronal, transverse and sagital planes of each ovary, which, for the 30 ovaries, provided 270 operator by ovary by plane estimations and 90 operator by ovary estimations for each type of structure. Images of cysts were hyperintense, those of early corpora lutea and follicles similar and moderate and those of late corpora lutea hypo-intense and not clearly discernable from ovarian stroma. Estimations of PS were too low in 68%, correct in 12% and too high in 20% of estimations (n=270). Estimations of PS were correct in three operator by ovary combinations, out by 1 in 22 and out by more than 1 in 65. No operator estimated PS correctly in any bitch. In Experiment 2 MRI was done on three deeply sedated bitches in the periovulatory phase in an attempt to obtain images of the ovaries in order to count the follicles. The acquisition time of 5-7 min rendered images of poor quality from live bitches and none of their ovaries could be seen. MRI is not suitable for counting follicles or corpora lutea in the ovaries of bitches.