Theriogenology 64 (2005) 408–415

www.journals.elsevierhealth.com/periodicals/the

Relationship between semen quality and

pixel–intensity of testicular ultrasonograms after
scrotal insulation in beef bulls
Andres A. Arteaga, Albert D. Barth, Leonardo F.C. Brito*
Department of Large Animal Clinical Sciences, Western College of Veterinary Medicine,
University of Saskatchewan, 52 Campus Drive, Saskatoon, SK, Canada S7N 5B4
Received 16 December 2004; accepted 17 December 2004

Abstract

The objective of this study was to determine the relationship between semen quality and testicular
pixel–intensity derived from image analysis of ultrasonograms after scrotal insulation in bulls. In
addition, the ability to predict semen quality based on testicular pixel–intensity was evaluated.
Sixteen beef bulls were selected on the basis of satisfactory semen quality and normal testicular
ultrasonogram appearance. Bulls were allocated into two groups for scrotal insulation for 4 days
(group 1) or 8 days (group 2). Semen was collected and evaluated twice weekly and testicular
ultrasonograms were evaluated once weekly for 8 weeks after removal of scrotal insulation. In
general, the percentages of motile and morphologically normal spermatozoa decreased below pre-
insulation levels from 1 to 5 weeks after scrotal insulation removal. Overall, group 1 had greater
(P < 0.01) percentages of motile and normal spermatozoa than group 2. Mean testicular pixel–
intensity (PI), and the number of pixels corresponding to the intensity that occurs most frequently
(NP) decreased in the first 2–3 weeks after scrotal insulation, coincidently with the decrease in sperm
motility and normal morphology. When the entire data set was evaluated, there was no association
between testicular PI or NP with semen quality observed at the same week of ultrasound examina-
tions. However, regression models indicated that testicular PI and NP accounted for 13–25% of the
variation in sperm motility and morphology in ejaculates collected 2–4 weeks after ultrasound exam.
Testicular PI and NP had moderate sensitivity and negative predictive values (64.5–82.6%), but low
specificity and positive predictive values (33.3–61.2%) as predictors of satisfactory semen quality
(60% motile spermatozoa and 70% morphologically normal spermatozoa) for ejaculates

* Corresponding author. Tel.: +1 306 966 7169; fax: +1 306 966 7159.
E-mail address: lfcbrito@lycos.com (Leonardo F.C. Brito).

0093-691X/$ – see front matter # 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.theriogenology.2004.12.008
 A.A. Arteaga et al. / Theriogenology 64 (2005) 408–415 409

collected 2–4 weeks after ultrasound exam. In conclusion, the deleterious effects of scrotal insulation
on semen quality were dependent on the length of the period of insulation and were associated with
changes in testicular ultrasonogram pixel–intensity. Testicular ultrasonogram pixel–intensity had a
better association with future semen quality than with present semen quality and was a better
predictor of unsatisfactory semen quality than satisfactory semen quality.
# 2004 Elsevier Inc. All rights reserved.

Keywords: Bull; Scrotal insulation; Semen; Testicular ultrasound

1. Introduction

Normal spermatogenesis in scrotal mammals depends upon maintenance of optimum

testicular temperature [1]. In bulls, testicular temperature is maintained 4–5 8C below body
temperature [2], and an increase in testicular temperature has detrimental effects on semen
quality [3]. The effect of increased testicular temperature on sperm motility and
morphology has been studied using scrotal insulation as a model [2–6]. The wide range of
effects of heat on sperm quality and testicular tissue make it difficult to propose a single
etiological factor. However, since blood flow in the testis does not increase at all, or not
enough to match the increased metabolic rate of the heated testis tissue, tissue hypoxia is
likely the main deleterious effect of testis heating [7]. The effect of scrotal insulation may
be reversible [3], but when damage is too severe, it may be irreversible [6].
Ultrasound examination of the testes has proven to be a valuable, noninvasive technique
for evaluation of testicular pathology in bulls [8]. Image analysis of ultrasonograms can
provide substantial information regarding the function and structure of tissues [9,10]. An
ultrasonogram is composed of an array of picture elements (pixels), with each pixel
representing a determined tissue density that is displayed in a range of shades of gray
(ranging from white to black). High-resolution ultrasonograms are necessary for accurate
detection of extremely small variations in tissue density that can be neither appreciated nor
measured by the human eye [9,10].
The objective of the present study was to evaluate the relationship between semen
quality and pixel–intensity analysis of testicular ultrasonograms after scrotal insulation in
bulls. In addition, the ability to predict semen quality based on testicular pixel–intensity
was also evaluated.

2. Materials and methods

Sixteen beef bulls (Hereford, Angus, and Charolais; 18–36-month old) were randomly
allocated into two treatment groups with scrotal insulation applied for 4 days (group 1) or
8 days (group 2). Bags for scrotal insulation were made with single layers of denim,
thinsulate, heat reflecting fabric, and flannelette. Insulation bags were lightly tied closed over
the neck of the scrotum with a purse string and sutured to the skin adjacent to the top of the
scrotum. Semen samples and testicular ultrasonograms were obtained and analyzed once
just prior to insulation. Following removal of the insulation, semen collections were done
twice weekly, while testicular ultrasonograms were obtained once weekly for 8 weeks.
410 A.A. Arteaga et al. / Theriogenology 64 (2005) 408–415

Semen was collected by electroejaculation and evaluated according to the guidelines of

the Western Canadian Association of Bovine Practitioners [11]. The percentage of
progressively motile spermatozoa was determined on a wet-mount with phase contrast
microscopy under 400 magnification. Eosin–nigrosin stained semen smears were
prepared for the assessment of sperm morphology and examined with light microscopy
under 1000 magnification. A total of 200 spermatozoa were examined in each smear and
the percentage of morphologically normal spermatozoa was determined.
Ultrasonographic examinations of the testes were done with a B-mode ultrasound
scanner (Aloka SDD-900; Aloka Co., Tokyo, Japan) connected to a 7.5 MHz linear array
transducer. The ultrasound settings (focus, gains, brightness, and contrast) were
standardized and the same settings were used for all the examinations. Gel was used
as a coupling material between the transducer and the scrotum and minimum pressure was
applied to obtain the image. Each testis was examined by placing the transducer vertically
(parallel to the long axis of the testis) on the caudal aspect of the scrotum. Frozen images
included visualization of the mediastinum in order to have an image across the middle of
the testis. Ultrasonograms were analyzed with the ultrasound machine; the Aloka 900-SD
ultrasound has the capability of analyzing pixel–intensity in a selected area bounded by a
box cursor within the ultrasonogram. Ultrasonograms were analyzed using the spot
metering technique [9] in two 1 cm2 (1681 pixels) spots selected approximately 1 cm
above the mediastinum and approximately 2 cm from the edge of the image. The
parameters evaluated included the mean pixel–intensity level within the selected area (PI;
scale 0–63), and the number of pixels corresponding to the intensity level that occurs most
frequently in the area (NP). Means for PI and NP from the two spots from the two testes
were calculated to be used in the statistical analysis.
Mixed-models analysis with Tukey’s test was used to detect and locate group, time, and
group-by-time interaction effects on sperm motility and morphology, and on testicular PI
and NP using the Statistical Analysis System (SAS Institute; Cary, NC, USA); the covariate
structure that best fitted each end point was used [12]. Pearson’s correlation coefficients
and regression models were evaluated using Statistix (Statistix Analytical Software;
Tallahassee, FL, USA). Weekly sperm motility and morphology were averaged for
evaluation of correlations and regression models. Correlation coefficients were calculated
between sperm motility and morphology, and between PI and NP. If a significant
correlation was observed, linear regression models were evaluated using sperm motility
and morphology as dependent variables, and PI and NP as independent variables.
Correlation and regression models were evaluated considering semen quality at the same
week that ultrasound exams were performed and considering semen quality observed
during successive weeks after ultrasound examinations. Sensitivity, specificity, and
predictive values of PI and NP for predicting satisfactory semen quality (60% motile
spermatozoa and 70% morphologically normal spermatozoa) were also evaluated.

3. Results

There were group (P = 0.007) and time effects (P < 0.0001) on sperm motility. Overall,
group 1 had greater sperm motility than group 2 (66.4% versus 41.5%, respectively).
 A.A. Arteaga et al. / Theriogenology 64 (2005) 408–415 411

Fig. 1. Mean (S.E.M.) proportion of motile (A) and normal spermatozoa (B) after scrotal insulation for 4 days
(group 1; n = 8) or 8 days (group 2; n = 8) in bulls (G: group effect; T: time effect; G  T: group-by-time
interaction effect). (*) Values within group differ (P < 0.05) from the day before insulation (day 0). (a,b) Values
differ (P < 0.05) between groups within examination day.

Sperm motility decreased (P < 0.05) below pre-insulation levels in both groups 7 days
after insulation removal and returned to the pre-insulation levels after 38 days only in group
1 (Fig. 1A). There were group (P = 0.009), time (P < 0.0001) and group-by-time
interaction effects (P = 0.002) on the proportion of normal spermatozoa. The proportion of
normal spermatozoa decreased (P < 0.05) below pre-insulation levels between 10 and 31
days after insulation removal in group 1 and between 3 and 35 days in group 2. The
proportion of normal spermatozoa was lower (P < 0.05) in group 2 than in group 1
between 3 and 49 days after insulation removal. Despite the lack of significant difference
from pre-insulation, the proportion of normal spermatozoa in group 2 was only 59% at 56
days after insulation removal (Fig. 1B).
There were time effects (P < 0.001) on PI and NP. Testicular PI decreased (P < 0.001)
below pre-insulation levels at 14 and 28 days after insulation removal and tended
(P = 0.08) to be lower than pre-insulation on day 35 (Fig. 2A). Testicular NP decreased
below pre-insulation levels 21 days after insulation removal (P < 0.05 on Days 21, 49, and
56; Fig. 2B). There was no significant correlation between PI and NP, and sperm motility
412 A.A. Arteaga et al. / Theriogenology 64 (2005) 408–415

Fig. 2. Mean (S.E.M.) testicular ultrasonogram pixel–intensity level (PI; A) and number of pixels of the most
frequent intensity level of the area (NP; B) observed after scrotal insulation for 4 or 8 days in bulls (n = 16) (T: time
effect). (*) Values differ (P < 0.05) from the day before insulation (day 0).

and morphology at the same week that ultrasound exams were performed. However,
significant correlations between PI and NP, and the percentages of motile and
morphologically normal spermatozoa were observed when ejaculates collected at
successive weeks after ultrasound exams were considered. In general, correlations
increased for ejaculates obtained between 1 and 3 weeks after the ultrasound examination
(data not shown). Regression models indicated that PI and NP accounted for 13–25% of the
variation in sperm motility and morphology observed between 2 and 4 weeks after the
ultrasound examination (Table 1); only 6–7% of the variation was accounted for in the
ejaculate collected 1 week after ultrasound examination. There were moderate sensitivity
(64.5–74.2%) and negative predictive values (75.0–82.6%) when PI  26 and NP  150
were used as predictors of satisfactory semen quality for ejaculates collected 2–4 weeks
after ultrasound exam, but specificity (45.7–61.2%) and positive predictive values (33.3–
51.3%) were low (Table 2).

4. Discussion

The timing of decline in semen quality observed after scrotal insulation in the present
experiment was similar to previous studies [2–6]. Sperm motility and morphology reached
 A.A. Arteaga et al. / Theriogenology 64 (2005) 408–415 413

Table 1
Regression models for the percentages of motile and morphologically normal sperm observed weeks after
ultrasound exam using mean testicular ultrasonogram pixel–intensity level (PI) and number of pixels correspond-
ing to the intensity level that occurred most frequently (NP) as independent variables
Slope R2 Probability
Motile sperm
P
2 weeks ( R2 = 0.19; y-intercept = 259.3)
PI 3.02 0.1 0.0001
NP 0.83 0.09 0.0005
P
3 weeks ( R2 = 0.18; y-intercept = 242.7)
PI 3.14 0.12 0.0001
NP 0.68 0.06 0.005
P
4 weeks ( R2 = 0.16; y-intercept = 144.1)
PI 3.08 0.16 0.0001
Normal spermP
2 weeks ( R2 = 0.18; y-intercept = 254.6)
PI 2.0 0.03 0.008
NP P 1.04 0.15 < 0.0001
3 weeks ( R2 = 0.25; y-intercept = 284.4)
PI 3.11 0.1 0.0001
NP P 1.02 0.15 < 0.0001
4 weeks ( R2 = 0.13; y-intercept = 132.5)
PI 2.89 0.13 0.0005
P 2
R : total model R2.

the lowest levels 2–3 weeks after insulation removal in both groups and started to improve
after 4 weeks. Eight days of scrotal insulation appeared to have a more severe effect on
semen quality than insulation for 4 days, suggesting that scrotal insulation not only have a
detrimental effect on semen quality, but also that the severity of this detrimental effect is
related to the length of the period of insulation.
Visual assessment of testicular ultrasonograms in the absence of gross pathological
conditions has very limited diagnostic value, since there were no significant correlations
between visual analysis and semen quality in bulls [13]. Furthermore, one study found no

Table 2
Sensitivity, specificity, and predictive values (PV) of mean testicular ultrasonogram pixel–intensity level (PI) and
number of pixels corresponding to the intensity level that occurred most frequently (NP) to predict satisfactory
semen quality (60% motile spermatozoa and 70% morphologically normal spermatozoa) weeks after the
ultrasound examination
Sensitivity (%) Specificity (%) Positive PV (%) Negative PV (%)
PI  26
2 weeks 67.7 53.1 35.6 81.1
3 weeks 74.2 58.5 46.0 82.6
4 weeks 64.5 61.2 51.3 73.2
NP  150
2 weeks 71.0 45.7 33.3 80.4
3 weeks 74.2 50.8 41.8 80.5
4 weeks 71.0 55.1 50.0 75.0
414 A.A. Arteaga et al. / Theriogenology 64 (2005) 408–415

evident visible changes in testicular ultrasonograms of bulls with testicular degeneration

induced by scrotal insulation [14]. Similarly, the visual appearance of testicular
ultrasonograms remained similar throughout the present study, i.e. the parenchyma of the
testes appeared homogeneous with moderate echodensity and the mediastinum appeared as
a linear structure in the center of the testis with greater echodensity than the parenchyma.
The only exception was one bull in group 2 in which both the parenchyma and
mediastinum became more heterogeneous and less echodense.
Image analysis of ultrasonograms pixel–intensity can provide detailed information
about tissue changes in different organs including the testis [9,10]. In the present study,
both testicular PI and NP decreased in the first 2–3 weeks after scrotal insulation removal,
coincidently with the decrease in semen quality. In general, PI gradually returned to pre-
insulation levels accompanying improvements in semen quality (although exceptionally
high values were observed 21 days after insulation removal), while NP remained lower
than pre-insulation levels until the end of the experimental period. These observations
indicated that acute testicular degenerative changes produced by scrotal insulation were
reflected in altered testicular pixel–intensity, but there was no close association between
pixel–intensity and semen quality during the recovery period.
Others studies have failed to demonstrate correlations between testicular ultrasonogram
pixel–intensity and semen quality in bulls, either during breeding soundness evaluation or
after scrotal insulation [15,16]. Apart from the differences in ultrasound equipment and
image analysis techniques used in these studies when compared to the present study, the
most likely reason for the lack of correlations was the attempt to correlate testicular pixel–
intensity and semen quality observed at the same day. The health status of the seminiferous
epithelium at a given day is not reflected in the ejaculate until several weeks later,
depending on the sensitivity of specific germ cell types to insults (like scrotal insulation)
and the time required for the transport of spermatozoa released into the seminiferous
tubules throughout the epididymis. Testicular ultrasonogram analysis is more likely
associated with future than with present semen quality. This was obvious in the present
study as there were no correlations between testicular PI or NP with semen quality
observed at the same week of the ultrasound examination. However, the correlations of
testicular PI and NP with semen quality in ejaculates collected after the ultrasound
examination increased up to the 3rd week. Moreover, testicular PI and NP accounted for
13–25% of the variation in sperm motility and morphology in ejaculates collected 2–4
weeks after the ultrasound examination.
When the possibility of using testicular ultrasonogram analysis for predicting
satisfactory semen quality within 2–4 weeks after ultrasound examination was
investigated, it was evident that evaluation of PI and NP was more useful to predict
unsatisfactory semen quality based on the moderate sensitivity and negative predictive
values, but low specificity and positive predictive values. Cut-off values of 26 for PI and
150 for NP were selected in this study, based on the data distribution, but it is important to
point out that these values would be different when using different settings of the same
ultrasound scanner, when using a different scanner, or when using different image analysis
techniques. The advent of more sophisticated ultrasound equipment and techniques in
image analysis will likely improve the ability to predict semen quality by evaluation of
testicular ultrasonograms.
 A.A. Arteaga et al. / Theriogenology 64 (2005) 408–415 415

In conclusion, the deleterious effects of scrotal insulation on semen quality were

dependent on the duration of insulation and were associated with changes in testicular
ultrasonogram pixel–intensity. Testicular ultrasonogram pixel–intensity was associated
with semen quality observed 2–4 weeks after ultrasound examination, but not with semen
quality observed at the same time as the ultrasound examination.

Acknowledgement

Research supported by the Saskatchewan Agricultural Development Fund.

References

[1] Waites GMH. Temperature regulation and the testis. In: Johnson AD, Gomes WR, Vandermark NL, editors.
The testis. New York: Academic Press; 1970. p. 241–79.
[2] Kastelic JP, Cook RB, Coulter GH, Saacke RG. Insulating the scrotal neck affects semen quality and scrotal/
testicular temperatures in the bull. Theriogenology 1996;45:935–42.
[3] Barth AD, Bowman PA. The sequential appearance of sperm abnormalities after scrotal insulation or
dexamethasone treatment in bulls. Can Vet J 1994;34:93–102.
[4] Volger CJ, Saacke RG, Bame JH, DeJarnette JM, McGilliard ML. Effects of scrotal insulation on viability
characteristics of cryopreserved bovine spermatozoa. J Dairy Sci 1991;74:3827–35.
[5] Vogler CJ, Bame JH, DeJarnette JM, McGilliard ML, Saacke RG. Effects of elevated testicular temperature
on morphology characteristics of ejaculated spermatozoa in the bovine. Theriogenology 1993;40:1207–19.
[6] Nageswara-Rao VD, Ramamohana-Rao A. Influence of heat induced testicular degeneration on semen
characteristics and testicular histology in rams. Indian Vet J 1977;54:719–26.
[7] Setchell BP. Heat and the testis. J Reprod Fertil 1998;114:179–94.
[8] Powe TA, Cartee RE, Carson R, Wolfe D, Hudson R. B-mode ultrasonography of testicular pathology in the
bull. Agri-Practice 1988;9:43–5.
[9] Pierson RA, Adams GP. Computer-assisted imaging analysis, diagnostic ultrasonography and ovulation
induction: strange bedfellows. Theriogenology 1995;43:105–12.
[10] Chandolia RK, Bartlewski PM, Omeke BC, Beard AP, Rawlings NC, Pierson RA. Ultrasonography of the
developing reproductive tract in ram lambs: effects of a GnRH agonist. Theriogenology 1997;48:99–117.
[11] Barth AD. Bull breeding soundness evaluation. The Western Canadian Association of Bovine Practitioners;
2002.
[12] Littell RC, Henry PR, Ammerman CB. Statistical analysis of repeated measurements data using SAS
procedures. J Anim Sci 1998;76:1216–31.
[13] Eilts BE, Pechman RD. B-mode ultrasound observations of bull testes during breeding soundness
examinations. Theriogenology 1988;30:1169–76.
[14] Sidibe M, Franco LA, Fredriksson G, Madej A, Malmgren L. Effects of testosterone, LH and cortisol
concentrations, and on testicular ultrasonographic appearance of induced testicular degeneration in bulls.
Acta Vet Scand 1992;33:192–6.
[15] Kastelic JP, Cook RB, Pierson RA, Coulter GH. Relationship among scrotal and testicular characteristics,
sperm production, and seminal quality in 129 beef bulls. Can J Vet Res 2001;65:111–5.
[16] Brito LFC, Silva AEDF, Barbosa RT, Unanian MM, Kastelic JP. Effects of scrotal insulation on sperm
production, semen quality, and testicular echotexture in Bos indicus and Bos indicus  Bos taurus bulls.
Anim Reprod Sci 2003;79:1–15.