Header

UZH-Logo

Maintenance Infos

Examination of cyclic changes in bovine luteal echotexture using computer-assisted statistical pattern recognition techniques


Herzog, Kathrin; Kiossis, Evangelos; Bollwein, Heiner (2008). Examination of cyclic changes in bovine luteal echotexture using computer-assisted statistical pattern recognition techniques. Animal Reproduction Science, 106(3-4):289-297.

Abstract

B-mode sonography is a well-established diagnostic tool for determination of cycle stage in gynaecology. The aim of this study was to determine whether computer-assisted texture analysis of B- mode sonographic images of bovine luteal glands provides further information about the animal's plasma progesterone concentration and cycle stage. Four Simmenthal cows were examined during two consecutive estrous cycles with an ultrasound device equipped with a 7.5MHz microconvex probe. During each examination three B-mode images of the corpus luteum (CL) were digitized and analyzed off-line using a computer-assisted texture analysis program. Size, echogeneity, and echotexture of the CL were characterized by the following texture parameters: area of cross-sectional planes of the CL (A), mean gray level (MGL), correlation (CORR), run percentage (RPERC), and long-run emphasis (LREM). Plasma progesterone levels (P4) were also determined. All parameters showed characteristic changes during the estrous cycle (P<0.05). Variance component estimates for the effect of Day of estrous cycle on A, MGL, CORR, RPERC, and LREM were 56.6%, 64.6%, 77.6%, 89.9%, and 86.0%, respectively, and 20.6%, 24.5%, 7.2%, 0.0%, and 14.0% for the influence of the individual cow. The factor estrous cycle within cows was responsible for 22.8%, 10.9%, 15.2%, 10.1%, and 0.0% of the variability of A, MGL, CORR, RPERC and LREM values, respectively. Cyclic changes were similar in A and P4. In contrast to P4, which decreased already between Days -5 and -3 (Day 0=ovulation), A stayed at constant high values until Day -3. Mean MGL values were higher (P<0.05) on Days 7, 9, and 13 compared to Days 3 and -3. Mean CORR values were constantly high (P>0.05) during the first days after ovulation and decreased continuously (P<0.05) between Days 5 and 13. Thereafter, mean CORR values remained low (P<0.05) until the next ovulation, except on Day -3 (P<0.05). Mean RPERC rose between Days 1 and 9 from low to high values (P<0.0001) remained at these high values (P>0.05) between Days 9 and 15, and decreased (P<0.05) afterwards to baseline values on Day -1. Mean LREM inclined steeply (P<0.0001) from minimum to maximum between Days 1 and 5. From Days 7 to -3, mean LREM remained (P>0.05) at a constant level close below the maximum value, and decreased to baseline values on Day -1. The results of this study show that statistical pattern recognition techniques provide new information about the luteal glands, thus facilitating a more accurate differentiation between different cycle stages in cows.

Abstract

B-mode sonography is a well-established diagnostic tool for determination of cycle stage in gynaecology. The aim of this study was to determine whether computer-assisted texture analysis of B- mode sonographic images of bovine luteal glands provides further information about the animal's plasma progesterone concentration and cycle stage. Four Simmenthal cows were examined during two consecutive estrous cycles with an ultrasound device equipped with a 7.5MHz microconvex probe. During each examination three B-mode images of the corpus luteum (CL) were digitized and analyzed off-line using a computer-assisted texture analysis program. Size, echogeneity, and echotexture of the CL were characterized by the following texture parameters: area of cross-sectional planes of the CL (A), mean gray level (MGL), correlation (CORR), run percentage (RPERC), and long-run emphasis (LREM). Plasma progesterone levels (P4) were also determined. All parameters showed characteristic changes during the estrous cycle (P<0.05). Variance component estimates for the effect of Day of estrous cycle on A, MGL, CORR, RPERC, and LREM were 56.6%, 64.6%, 77.6%, 89.9%, and 86.0%, respectively, and 20.6%, 24.5%, 7.2%, 0.0%, and 14.0% for the influence of the individual cow. The factor estrous cycle within cows was responsible for 22.8%, 10.9%, 15.2%, 10.1%, and 0.0% of the variability of A, MGL, CORR, RPERC and LREM values, respectively. Cyclic changes were similar in A and P4. In contrast to P4, which decreased already between Days -5 and -3 (Day 0=ovulation), A stayed at constant high values until Day -3. Mean MGL values were higher (P<0.05) on Days 7, 9, and 13 compared to Days 3 and -3. Mean CORR values were constantly high (P>0.05) during the first days after ovulation and decreased continuously (P<0.05) between Days 5 and 13. Thereafter, mean CORR values remained low (P<0.05) until the next ovulation, except on Day -3 (P<0.05). Mean RPERC rose between Days 1 and 9 from low to high values (P<0.0001) remained at these high values (P>0.05) between Days 9 and 15, and decreased (P<0.05) afterwards to baseline values on Day -1. Mean LREM inclined steeply (P<0.0001) from minimum to maximum between Days 1 and 5. From Days 7 to -3, mean LREM remained (P>0.05) at a constant level close below the maximum value, and decreased to baseline values on Day -1. The results of this study show that statistical pattern recognition techniques provide new information about the luteal glands, thus facilitating a more accurate differentiation between different cycle stages in cows.

Statistics

Citations

Dimensions.ai Metrics
13 citations in Web of Science®
13 citations in Scopus®
Google Scholar™

Altmetrics

Downloads

0 downloads since deposited on 23 Aug 2018
0 downloads since 12 months

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:05 Vetsuisse Faculty > Veterinary Clinic > Department of Farm Animals
Dewey Decimal Classification:570 Life sciences; biology
630 Agriculture
Scopus Subject Areas:Health Sciences > Food Animals
Life Sciences > Animal Science and Zoology
Life Sciences > Endocrinology
Language:English
Date:July 2008
Deposited On:23 Aug 2018 17:56
Last Modified:31 Jul 2020 01:52
Publisher:Elsevier
ISSN:0378-4320
OA Status:Closed
Publisher DOI:https://doi.org/10.1016/j.anireprosci.2007.05.004
PubMed ID:17573209

Download

Closed Access: Download allowed only for UZH members