Comparison between ultrasonographic and caliper measurements of testicular volume in the dog

Animal Reproduction Science 108 (2008) 1–12

Comparison between ultrasonographic and calipermeasurements of testicular volume in the dog

Pagona G. Gouletsou a,∗, Apostolos D. Galatos b,Leonidas S. Leontides c

a Clinic of Obstetrics and Reproduction, Faculty of Veterinary Medicine,University of Thessaly, Trikalon 224, GR-43100 Karditsa, Greece

b Clinic of Surgery, Faculty of Veterinary Medicine, University of Thessaly, Karditsa, Greecec Laboratory of Epidemiology, Biostatistics and Animal Health Economics, Faculty of Veterinary Medicine,

University of Thessaly, Karditsa, Greece

Received 30 March 2007; received in revised form 12 June 2007; accepted 19 June 2007Available online 23 June 2007

Abstract

The aim of this study was to compare the accuracy of two methods used to clinically estimate testicularvolume in the dog.

Caliper and ultrasonographic measurements of testicular dimensions (length, width and height) ofboth testes were performed on 21 adult Beagles. Either measurement was taken in vivo over the scro-tal skin and again in vitro after castration, excluding or including, when possible, the epididymis. In asub-sample of 14 testes of 7 randomly selected dogs, the testicular volume was calculated by the for-mula of an ellipsoid: length (l) × width (w) × height (h) × 0.5236, and the empiric formula of Lambert:(l) × (w) × (h) × 0.71. The calculated volumes were then compared to the actual ones, which were estimatedby water displacement. For each measurement method, the formula that estimated more accurately the truetesticular volume was, subsequently, applied to the calculation of the testicular volume of the remainingtestes.

The formula of ellipsoid estimated testicular volume more accurately when in vivo and in vitro calipermeasurements and in vitro ultrasound measurements were performed. The formula of Lambert estimatedtesticular volume more accurately when in vivo ultrasound measurements were performed. The volumescalculated from the measurements of calipers over the scrotal skin overestimated true volumes, on aver-age, by 69 ± 27% (concordance correlation coefficient rccc = 0.49). By contrast, those calculated fromthe ultrasonography measurements were more accurate, overestimating the true volumes by 17 ± 24%(rccc = 0.81).

∗ Corresponding author.E-mail address: [email protected] (P.G. Gouletsou).

0378-4320/$ – see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.anireprosci.2007.06.020

mailto:[email protected]

dx.doi.org/10.1016/j.anireprosci.2007.06.020

2 P.G. Gouletsou et al. / Animal Reproduction Science 108 (2008) 1–12

Therefore, ultrasonography when the formula of Lambert is used is more accurate than caliper methodsfor the estimation of testicular volume in clinical practice and should be the method of choice.© 2007 Elsevier B.V. All rights reserved.

Keywords: Testis volume; Ultrasonography; Dog

1. Introduction

Accurate determination of testicular volume is of great benefit in the evaluation of patientswith a variety of disorders affecting, among other things, testicular growth and development(Paltiel et al., 2002). Furthermore, determination of testicular volume is important in assessingpubertal development and the effects of illness, trauma, orchitis, epididymitis, vasectomy andscrotal hernia, as far as their treatment on reproductive function (Chipkevitch et al., 1996; Paltielet al., 2002). Although, in some clinical settings, it is adequate to rank testicular size as eithersmall, medium or large, greater accuracy is required when one needs to closely follow changesin pubertal status, monitor testicular pathological processes, or evaluate the effect of therapeuticor hormonal agents on the testes. Testicular volume, largely, reflects spermatogenesis (Setchelland Brooks, 1988), because approximately 70–80% of testicular mass consists of seminiferoustubules, and is correlated with total sperm count, sperm motility, sperm morphology and dailysperm production (Olar et al., 1983; Takihara et al., 1983; Cartee et al., 1986, 1989; Behre et al.,1989; Arai et al., 1998).

For estimation of testicular dimensions or volume, a number of techniques have been used,including Prader (Prader, 1966) or Rochester orchidometers (Takihara et al., 1983), ruler orcalipers (Lambert, 1951; Hansen and With, 1952), and, relatively recently, ultrasonography(Rivkees et al., 1987; Fuse et al., 1990). The accuracy of these methods is still debated (Schiff etal., 2004). Experimental and clinical observations showed that in vivo measurements are proneto error because of observer and methodological biases (Behre et al., 1989). The major ones stemfrom the inclusion of part or all of the epididymis, the variability in the thickness and elasticity ofthe scrotal skin, the compression of the scrotal contents, and the observer variability in compar-ative palpation (Handelsman and Staraj, 1985). Ultrasonography is generally recognized as themost accurate means of in situ assessment of testicular volume (Rivkees et al., 1987; Behre et al.,1989; Fuse et al., 1990; Lenz et al., 1993; Diamond et al., 2000; Paltiel et al., 2002). However,some have reported that ultrasonography estimates of testicular volume are not the most accurate(Pugh et al., 1990; Ariturk and Ozates, 1993; Cayan et al., 2002), whilst others have found largevariability in the ultrasonographic estimates, based on the examiner and the formula used to cal-culate the testicular volume (Rivkees et al., 1987; Behre et al., 1989; Cartee et al., 1990; Carlsenet al., 2000; Tatsunami et al., 2006).

Various formulae for calculating testicular volume have been used; the most commonly appliedis the formula of an ellipsoid (volume = length × height × width × 0.5236), which is based on theassumption that the testis has the same shape (Fuse et al., 1990; England, 1991; Chipkevitch et al.,1996; Diamond et al., 2000; Gumbsch et al., 2002). Others used an empirical modification of theabove formula that was proposed by Lambert (1951), i.e. volume = length × height × width × 0.71(Ariturk and Ozates, 1993; Schiff et al., 2004; Shiraishi et al., 2005). To our knowledge, only oneattempt has been made to directly compare the accuracy of these formulae (Paltiel et al., 2002).

Therefore, the aims of our study were (a) to compare the accuracy of the measurements oflength, height and width of canine testes by using sliding calipers in vivo and in vitro, andultrasonography in vivo and in vitro, (b) to compute testicular volume from both calipers and

P.G. Gouletsou et al. / Animal Reproduction Science 108 (2008) 1–12 3

ultrasonography-derived measurements in vivo and in vitro by the formulae of ellipsoid and theempiric formula of Lambert, and to compare the accuracy of these formulae with respect to thetrue volume measured by water displacement, and (c) to determine the concordance betweeneither testicular volume measured with sliding calipers or ultrasonography in vivo and testicularvolume measured by these methods in vitro. Finally, we also investigated whether the inclusionor exclusion of the epididymis affected the accuracy of testicular measurements in vivo.

2. Materials and methods

2.1. Animals

The experiment was carried out under a license obtained from the Greek Ministry of Agricul-ture. Forty-two testes of 21 adult (age range 1–3 years) Beagle dogs, weighing between 11 and17 kg, were used.

2.2. Clinical examination

Before being included in the study each dog underwent a general clinical examination, fol-lowed by an evaluation and palpation of the scrotum, testes, epididymes and pampiniform plexusin order to ensure that no observable gross pathology was present on the external genitalia. Then,the dog was mildly sedated with a mixture of medetomidine (10 �g/kg) and morphine (1 mg/kg)given intramuscularly and restrained in the standing position and the scrotum was clipped. Mea-surements in millimeters of testicular length, width and height were obtained by using slidingcalipers, after stretching the scrotal skin tightly over the testes. No attempts were made to avoidthe epididymis during caliper in vivo measurements, as epididymis in the dog is strongly attachedto the testis, making impossible to measure the largest craniocaudal and dorsolateral diameter ofthe testis without including the epididymis.

2.3. Ultrasonographic examination

With the animal restrained in the standing position, acoustic gel was applied to the area of thescrotum. Scans were made with a real time scanner (AMI B7; Alliance Medical, Quebec, Canada)using 6.0 MHz sector transducer coupled directly to the skin. Images of the testes were obtainedin sagittal and transverse planes. Scanning was done by using light pressure to avoid a distortionof testicular shape. Testicular length, width and height were measured by using electronic cursorsintegrated into the ultrasound machine. Cursors were set at the borders of the tunica albuginea.At least three separate sagittal and transverse images of each testis were obtained and the largestmeasurement obtained for each testicular dimension was recorded and used for volume calculationand statistical analysis. An effort was made to avoid the inclusion of epididymis when performingtesticular measurements. The echogenicity, homogeneity and surface of the scrotal contents werealso assessed.

2.4. True testicular volume determination

Following ultrasonographic examination, anaesthesia was induced with propofol (3–4 mg/kg)and maintained with isoflurane 2% in oxygen and bilateral orchiectomies were performed. Sub-sequently, the testes were hanged up from the spermatic tone, in order to eliminate distortion of


testicular shape as a result of resting on a table, and their dimensions were obtained by calipers.Afterwards, they were immersed into normal saline and their dimensions were obtained by ultra-sound. The testicular length and height in vitro was calculated both including and excluding theepididymis. In total, six measurements of testicular dimensions were obtained for each of the 42testes: in vivo by sliding calipers (Cvv), in vivo by ultrasonography (Uvv), in vitro by slidingcalipers excluding epididymis (Cvt − e), in vitro by sliding calipers including epididymis (Cvt + e),in vitro by ultrasonography excluding epididymis (Uvt − e), and in vitro by ultrasonographyincluding epididymis (Uvt + e).

The epididymides and the vasculature were removed from 14 testes (of 7 randomly selecteddogs) and the testicular volume was calculated by water displacement, according to Archimedes’principle. The dimensions obtained by ultrasonography and calipers from these 14 testes were usedto calculate testicular volume by the formula of ellipsoid: volume = length (l) × width (w) × heath(h) × 0.5236, and by the empiric formula of Lambert: volume = (l) × (w) × (h) × 0.71. The for-mula that calculated testicular volume with the smallest mean difference from true mean volume,which was calculated by water displacement, was used to calculate testicular volume in theremaining 28 testes.

2.5. Statistical analysis

Paired t-tests were used to compare the mean values of dimensions and volume obtained byCvv, Cvt + e, Uvv, Uvt − e and Uvt + e to those obtained by Cvt − e and by water displacement,respectively. The level of agreement between Cvt − e obtained dimensions and volume and thoseobtained by Cvv, or Cvt + e, or Uvv, or Uvt − e, or Uvt + e, was evaluated by the concordancecorrelation coefficient (rccc). This statistic better reflects the level of agreement between thepaired results than the Pearson correlation coefficient (r) does, because it is computed from threeparameters, namely, a location-shift parameter which measures how far the data are above or belowthe equality line, a scale-shift parameter which measures the difference between the slope for thedata and the equality line (slope = 1), and the usual Pearson correlation coefficient which measureshow tightly clustered the data are around the line (Lin, 1989). The relative difference between rcccand r is shown by the ratio rccc/r. Furthermore, we also developed limits of agreement plots, whichwere used to investigate the differences between the pairs of dimension and volume results relativeto their mean value. Hence, we were able to determine whether the level of disagreement betweenthe two sets of paired results varied with the mean value of the measurement being measured. Allstatistical analyses were done in STATA (StataCorp., 2007. Stata Statistical Software: Release 9.5College Station, TX).

3. Results

3.1. Dimension measurements

Cvv derived length dimensions were larger than the Cvt − e ones (percentage difference27%), whilst Uvv derived length dimensions were slightly larger than Cvt − e ones (2%). Cvvderived height dimensions were larger than Cvt − e derived height dimensions (10.5%) whilstUvv derived height dimensions were smaller than Cvt − e derived height dimensions (25%). Cvvderived width dimensions were larger than Cvt − e derived width dimensions (19%), whilst Uvvderived width dimensions were larger than Cvt − e derived width dimensions (14%). There was ahigher degree of concordance between Cvt − e and Uvv measurements. Results are presented in

https://www.researchgate.net/publication/20436291_A_Concordance_Correlation-Coefficient_To_Evaluate_Reproducibility?el=1_x_8&enrichId=rgreq-0c4fbd54-2819-49dc-a70d-7823402da51f&enrichSource=Y292ZXJQYWdlOzYxMTc2MjA7QVM6MjEyMjYzOTg3MzU1NjUyQDE0Mjc2MTkwODY0MTA=


Table 1Comparison of testicular length estimated by sliding calipers (C) and ultrasonography (U) before (vv) and after (vt)castration, including (+e) and excluding (−e) epididymis with that estimated by calipers after castration (Cvt − e)

Measurementmethod

Length(mean ± S.D.)

Difference fromtrue length (mm)(mean ± S.D.)

Percentagedifference (%)(mean ± S.D.)

P-value Concordancecorrelationcoefficient (rccc)

95% CI ofrccc

Cvv 38.5 ± 5 8.1 ± 2.8 27 ± 9.7 <0.001 0.3 0.2–0.38Cvt − ea 30.4 ± 3.9 NA NA NA NA NACvt + e 38.2 ± 5 7.8 ± 3.1 26 ± 10 <0.001 0.3 0.2–0.4Uvv 30.9 ± 4 0.5 ± 2.9 2.1 ± 9.7 0.25 0.72 0.57–0.87Uvt − e 29 ± 3.8 −1.3 ± 2 −4 ± 6.6 0.0001 0.8 0.7–0.9Uvt + e 36.5 ± 4.2 6.1 ± 2.1 20.5 ± 7.5 <0.0001 0.4 0.3–0.5

NA: not applicable.a The comparison method.

Table 2Comparison of testicular height estimated by sliding calipers (C) and ultrasonography (U) before (vv) and after (vt)castration, including (+e) and excluding (−e) epididymis with that estimated by calipers after castration (Cvt − e)

Measurementmethod

Height(mean ± S.D.)




95% CI of rccc

Cvv 26.3 ± 4 2.5 ± 1.5 10.5 ± 6.3% <0.001 0.73 0.64–0.83Cvt − ea 23.8 ± 3.2 NA NA NA NA NACvt + e 27.2 ± 3.9 3.3 ± 1.8 26 ± 10 <0.001 0.6 0.5–0.72Uvv 17.6 ± 2.8 −6.2 ± 1.8 −25 ± 6.6 <0.001 0.26 0.18–0.33Uvt − e 22.4 ± 3.3 1.4 ± 1.3 −6 ± 5.5 0.0001 0.84 0.75–0.92


Tables 1–3.

3.2. Formulae for volume calculation

When compared to the volume measured by water displacement, the formula of ellipsoidestimated testicular volume more closely than that of Lambert, when dimension measurementswere obtained by Cvv, by Cvt − e, and by Uvt − e. However, the latter formula estimated more

Table 3Comparison of testicular of width estimated by sliding calipers (C) and ultrasonography (U) before (vv) and after (vt)castration, including (+e) and excluding (−e) epididymis with that estimated by calipers after castration (Cvt − e)

Measurementmethod

Width(mean ± S.D.)




95% CIof rccc

Cvv 24.6 ± 3.9 4 ± 2.2 19 ± 10.7 <0.001 0.48 0.35–0.61Cvt − ea 20.6 ± 3 NA NA NA NA NAUvv 23.4 ± 3.9 2.8 ± 2.9 14.1 ± 14.4 <0.001 0.5 0.32–0.67Uvt − e 23.3 ± 3.2 2.7 ± 2 13.5 ± 10 <0.0001 0.58 0.43–0.72



Table 4Differences and percentage differences (mean ± S.D.) from true volume, which was measured by water displacement, byeach formula

Measurementmethod

Difference from true volume (ml) Percentage difference (%)

Formula of ellipsoid Formula of Lambert Formula of ellipsoid Formula of Lambert

Cvv 5.4 ± 6.4 10.6 ± 4.9 57 ± 10 113 ± 14Cvt − e −0.6 ± 0.06 2.4 ± 1.7 −8 ± 5 24 ± 7Cvt + e 3 ± 2 7.4 ± 4 31 ± 7 77 ± 10Uvv −2.7 ± 1.2 −0.38 ± 1.8# −31 ± 11 −6 ± 16Uvt − e −0.95 ± 0.5 2 ± 1 −10 ± 5 21 ± 7Uvt + e 1.5 ± 1 5.4 ± 2.5 16 ± 8 58 ± 11

All values differ from true volume, measured by water displacement with the exception of that marked with #.

closely the testicular volume when dimensions were measured by Uvv. Data are presented inTable 4.

3.3. Volume calculation

We calculated testicular volumes of all studied dogs by Cvv, Cvt and Uvt with the formula ofellipsoid, and by Uvv with the formula of Lambert. The true testicular volume was calculated byCvt − e derived volume. Cvv derived volume overestimated true testicular volume by a mean of5.7 ml or 69%. Uvv derived volume overestimated true testicular volume by a mean of 1.3 ml or17%. However, Uvt − e derived volume overestimated true testicular volume by only a mean of0.13 ml or 2.3%. Details are presented in Table 5.

3.4. Concordance

The rccc values between all derived volumes are presented in Table 5. The limits of agreementplot for testicular volume measured either by Cvv or by Cvt − e showed that Cvv tended tooverestimate the true value and that, as the mean testicular volume increased, there was a tendencyfor higher level of disagreement (Fig. 1.).

Uvv derived volume had concordance correlation coefficient with Uvt + e derived volume ofrccc = 0.86. Uvv derived volume had concordance correlation coefficient with Uvt − e derivedvolume of rccc = 0.85. The limits of agreement plot for testicular volume measured either by Uvvor by Cvt − e showed that Uvv tended to overestimate true value and that, as the mean testicularvolume increased, there was a tendency for higher level of disagreement (Fig. 2.).

4. Discussion

Determination of testicular volume is important in assessing pubertal development and theeffects of illness and treatment on its function (Chipkevitch et al., 1996; Paltiel et al., 2002). In thepast, attempts were made to improve clinical assessment of testicular volume by a vast variety oforchidometers, but in most of them ultrasonographic measurements of testicular volume have beenused as the standard to which clinical measurements were compared (Behre et al., 1989; Costabileet al., 1992; Lenz et al., 1993; al Salim et al., 1995; Taskinen et al., 1996; Chipkevitch et al., 1996;Carlsen et al., 2000; Diamond et al., 2000; Cayan et al., 2002; Schiff et al., 2004). However,

P.G.G

ouletsouetal./A

nimalR

eproductionScience

108(2008)

1–127

Table 5Comparison of testicular volumes estimated by sliding calipers (C) and ultrasonography (U) before (vv) and after (vt) castration, including (+e) and excluding (−e) the epididymiswith those estimated by calipers after castration (Cvt − e)

Measurementmethod

Volume (ml)(mean ± S.D.)

Difference fromtrue volume (mm)(mean ± S.D.)

Percentage difference(%) (mean ± S.D.)

P-value Concordance correlationcoefficient (rccc)

95% CI of rccc Pearson’s correlationcoefficient (r)

rccc/r

Cvv 13.9 ± 6.3 5.7 ± 3.2 69 ± 27 <0.0001 0.49 0.38–0.59 0.94 0.521Cvt − ea 8.2 ± 3.5 NA NA NA NA NA NA NACvt + e 11.7 ± 5 3.5 ± 2.1 44 ± 18 <0.0001 0.66 0.55–0.76 0.94 0.701Uvv 9.5 ± 4.3 1.3 ± 2.1 17 ± 24 0.0002 0.81 0.7–0.91 0.87 0.924Uvt − e 8.3 ± 3.5 0.13 ± 1.3 2.3 ± 15 0.5 0.93 0.88–0.97 0.93 0.999Uvt + e 10.4 ± 4.2 2.2 ± 1.4 28 ± 17 <0.0001 0.79 0.71–0.88 0.95 0.840



Fig. 1. Ninety five percent limits of agreement plot for testicular volume measured either by sliding calipers in vivo orby sliding calipers in vitro, excluding the epididymis. As the mean testicular volume increased, there was a tendency forhigher level of disagreement.

only few of these reports addressed the issue of the accuracy and precision of ultrasonographicmeasurements of testicular volume (Behre et al., 1989; Costabile et al., 1992; Lenz et al., 1993).

In an attempt to evaluate the ultrasonographic volume estimation technique, Rivkees et al.(1987), estimated testicular volume of calves and dogs by Prader orchidometer and testiculardimensions by a straightedge or tape and by ultrasonography and reported a strong correlation(r = 0.98) between actual testicular volume and ultrasound findings. In contrast, testicular vol-ume calculated from length and width measurements exceeded actual testicular volume by nearly

Fig. 2. Limits of agreement plot for testicular volume measured either by ultrasonography in vivo or by sliding calipersin vitro, excluding the epididymis. As the mean testicular volume increased, there was a tendency for higher level ofdisagreement.


30%. Takihara et al. (1983) also observed a poor correlation between calipers measurement andtesticular size determined by water displacement. Behre et al. (1989) reported a high correlation(r = 0.99) between testicular volume measured by weighting and water displacement and volumedetermined by ultrasonography. In a similar study, Paltiel et al. (2002) compared the accuracyof Prader and Rochester orchidometers and ultrasonographic measurements of testicular volumewith the testicular volume calculated after weighted and found ultrasonography more accurate.Dornberger and Dornberger (1987) used ultrasonography for volume determination and intro-duced an empirical correction factor for the formula of ellipsoid in order to obtain comparableand highly correlated estimates. Ariturk and Ozates (1993), who evaluated testicular volumesby ultrasonography pre-operatively and by sliding calipers during surgery, by using the empiricformula of Lambert, found no difference between the two methods. Fuse et al. (1990) found thatthe correlation coefficient between testicular volume measured by sliding calipers on scrotal skinand the actual testicular volume measured by water displacement was rather low (r = 0.73), whilston the other hand, the volume measured by ultrasonography proved more accurate (r = 0.86). Thesame authors found that the actual testicular volume measured by water displacement and therespective volume measured in vitro by sliding calipers were almost the same (r = 0.97).

In dogs, Pugh et al. (1990) found no difference between the gross and ultrasonographic mea-surements. Finally, England (1991), who measured ultrasonographicaly dog’s testicular volume ina waterbath, using the formula of ellipsoid, found that there was a significant relationship betweencalculated testicular volume and actual testicular volume determined by water displacement(r = 0.96).

Sliding calipers were chosen in this study, so as all three dimensions could be comparedbetween clinical and ultrasonographic measurements. Also, the size of a three-dimensional objectas the testis would be more accurately estimated by using a combination of dimensions thatestimate the volume of an ellipsoid. Furthermore, by using sliding calipers, measurements invitro both including and excluding epididymis could be performed, in an attempt to determine theeffect of epididymis to the accuracy of testicular measurements. The Cvt − e derived volume wasdesignated as the standard because caliper is calibrated easier than ultrasound electronic calipers.Furthermore, it is more precise to touch the tunica albuginea of the testis with the caliper than toapply the electronic cursors on a frozen image of the ultrasound machine.

4.1. Dimension measurements

In our study, Cvv derived length and height dimensions were larger than the Cvt − e derivedones by 27% and 10.5%, respectively, when the epidydimis was excluded; when it was included,length dimensions were larger by 1% and height dimensions were smaller by 3%, respectively,showing that the inclusion of epididymis played an important role in testis length and heightdimension measurements in vivo. Cvv derived height measurements were slightly smaller thanCvt + e derived height measurements because during in vivo stabilization of the testis into thescrotum the epididymal body is compressed, whilst no such compression is apparent duringthe in vitro examination. Uvv derived length dimensions were larger than the Cvt − e derivedlength dimensions by only 2%, whilst height derived dimensions were smaller than Cvt − ederived height dimensions by 25%. The relatively large proportion of height underestimation(25%) and width overestimation (14%) by Uvv was probably the result of the pressure of theprobe to the testis. When no pressure was applied, i.e. the testes were suspended in waterbath,the percentage difference in height dimensions between Uvt − e and Cvt − e was much smaller(−6%). Uvv derived length dimensions were larger than Uvt − e derived length dimensions by


6.9%, whilst were smaller than Uvt + e derived length dimensions by 15%, indicating that inclusionof epididymis was unimportant in testis ultrasonography dimension measurements in vivo.

The data suggested that a higher degree of reliability was achieved when measurements oftesticular dimensions were obtained with ultrasonography, especially for length, even if a smalldistortion of testicular shape was obvious.

4.2. Formulae for volume calculation

From the two formulae that were used, the formula of ellipsoid had the smallest mean differencefrom true volume for testicular volume calculations by Cvv, Cvt and Uvt, whilst the formula ofLambert had the smallest mean difference from true volume for calculations of testicular volumesby Uvv. Our findings are in agreement with those of Paltiel et al. (2002). These findings are ofparticular importance, because most of the ultrasound machines use the formula of ellipsoid forautomatic calculation of volume (Paltiel et al., 2002) and as the formula of ellipsoid is the morecommonly referenced. Water displacement is a more direct method of volume measurement, but itis still inaccurate because of the difficulty to be performed when the volume of the testis is small,i.e. less than 6–7 ml, due to the capillary effect of the water inside small containers. Furthermore, itcannot be performed when the epididymis and vasculature should be left in place for histologicalsections.

4.3. Volume calculation

Uvv derived volumes were clearly superior to the Cvv derived ones, a fact reflected by over-estimation of 17% and 69% of true volume, respectively. The rccc between Cvv and Cvt − ederived volumes was 0.49, whilst it was 0.81 between Uvv and Cvt − e, confirming the superi-ority of ultrasound measurements. Testes are compressible, and during ultrasound examination,their dimensions may be distorted, as was shown by the high proportion of height underestimation(25%) and width overestimation (14%) by Uvv. However, the overall volume was unchanged, andwas accurately estimated by ultrasonography. Furthermore, we estimated a very high rccc = 0.93between Uvt − e and Cvt − e derived volumes; the former were larger than the latter by only2.3%. These statistics document that ultrasonography dimension measurements are precise andaccurate when performed under ideal conditions, i.e. with the testis suspending in a waterbath.Uvv derived volumes were larger than Uvt − e derived volumes by 14.6% (rccc = 0.85), whilstwere smaller than Uvt + e derived volumes by 10.9% (rccc = 0.86), indicating that the inclusion ofepididymis plays a small role in testis’ measurements in vivo by ultrasonography. On the contrary,Cvv derived volumes were larger than Cvt − e derived volumes by 69% (rccc = 0.49), whilst werelarger than Cvt + e derived volume by 26.8% (rccc = 0.86), indicating that inclusion of epididymisplays an important role in testis’ measurement in vivo by calipers.

We opted to calculate the rccc instead of r, that has been previously applied in similar studiescomparing the accuracy of methods to calculate testicular dimensions and volume ((Behre et al.,1989; Rivkees et al., 1987; Fuse et al., 1990; England, 1991). This was because r measures thedegree to which one set of test results varies linearly with a second set without directly comparingthe values obtained and ignoring the scales of the two sets of results. As we noted earlier, the rcccbetter reflects the level of agreement between two sets of results than r, because it is computed froma location-shift parameter, a scale-shift parameter and r. A rccc = 1 indicates perfect agreement(Lin, 1989). The results summarized in Table 5 clearly show that rccc estimates are more accuratelysummarizing the level of agreement between two sets of results than r does. The disagreement


between r and rccc is striking for Cvv volume measurements (r = 0.94, rccc = 0.49); the high valueof r indicated an accurate method although the overestimation of true volume by this method was69%.

5. Conclusions

The results of the present study indicate that the formula of ellipsoid provides superior estimateof testicular volume, and should be used in both clinical practice and research for testicular volumecalculations in vivo and in vitro by calipers and in vitro by ultrasound, whilst the formula of Lam-bert provides superior estimate for calculations of testicular volumes in vivo by ultrasound. Thevolume measured by sliding calipers on scrotal skin was found to be incorrect. Ultrasonographicmethod of testicular volume determination is more accurate and more precise than calipers, asreflected by measurements closer to true volume and higher values of rccc, and is considered tobe the best method. Ultrasonography provides an excellent tool for determining testicular volumewhen objective, accurate and reproducible measurements of testicular volume are required.

References

al Salim, A., Murchison, P.J., Rana, A., Elton, R.A., Hargreave, T.B., 1995. Evaluation of testicular volume by threeorchidometers compared with ultrasonographic measurements. Br. J. Urol. 76, 632–635.

Arai, T., Kitahara, S., Horiuchi, S., Sumi, S., Yoshida, K., 1998. Relationship of testicular volume to semen profiles andserum hormone concentrations in infertile Japanese males. Int. J. Fertil. Womens Med. 43, 40–47.

Ariturk, E., Ozates, M., 1993. Measurement of testicular volume by ultrasonography. Turk. J. Pediatr. 35, 177–180.Behre, H.M., Nashan, D., Nieschlag, E., 1989. Objective measurement of testicular volume by ultrasonography: evaluation

of the technique and comparison with orchidometer estimates. Int. J. Androl. 12, 395–403.Carlsen, E., Andersen, A.G., Buchreitz, L., Jorgensen, N., Magnus, O., Matulevicuus, V., Nermoen, I., Petersen, J.H.,

Punab, M., Suominen, J., Zilaitiene, B., Giwercman, A., 2000. Inter-observer variation in the results of the clinicalandrological examination including estimation of testicular size. Int. J. Androl. 23, 248–253.

Cartee, R.E., Gray, B.W., Powe, T.A., Hudson, R.S., Whitesides, J., 1989. Preliminary implications of B-mode ultrasonog-raphy of the testicles of beef bulls with normal breeding soundness examinations. Theriogenology 31, 1149–1157.

Cartee, R.E., Powe, T.A., Gray, B.W., Hudson, R.S., Kuhlers, D.L., 1986. Ultrasonographic evaluation of normal boartesticles. Am. J. Vet. Res. 47, 2543–2548.

Cartee, R.E., Rumph, P.F., Kenter, D.C., Cooney, J.C., Frank, D., Haught, J., Leong, P., Humphries, M., Amaratunga,P., Zampaglioni, N., 1990. Evaluation of drug-induced prostatic involution in dogs by transabdominal B-modeultrasonography. Am. J. Vet. Res. 51, 1773–1778.

Cayan, S., Akbay, E., Bozlu, M., Doruk, E., Yildiz, A., Acar, D., Kanik, E.A., Ulusoy, E., 2002. Diagnosis of pediatricvaricoceles by physical examination and ultrasonography and measurement of the testicular volume: using the praderorchidometer versus ultrasonography. Urol. Int. 69, 293–296.

Chipkevitch, E., Nishimura, R.T., Tu, D.G., Galea-Rojas, M., 1996. Clinical measurement of testicular volume in adoles-cents: comparison of the reliability of 5 methods. J. Urol. 156, 2050–2053.

Costabile, R.A., Skoog, S., Radowich, M., 1992. Testicular volume assessment in the adolescent with a varicocele. J.Urol. 147, 1348–1350.

Diamond, D.A., Paltiel, H.J., DiCanzio, J., Zurakowski, D., Bauer, S.B., Atala, A., Ephraim, P.L., Grant, R., Retik,A.B., 2000. Comparative assessment of pediatric testicular volume: orchidometer versus ultrasound. J. Urol. 164,1111–1114.

Dornberger, V., Dornberger, G., 1987. Comparative volumetry of human testes using special types of testicular sonography,Prader’s orchidometer, Schirren’s circle and sliding caliber. Andrologia 19, 487–496.

England, G.C.W., 1991. Relationship between ultrasonographic appearance, testicular size, spermatozoal output andtesticular lesions in the dog. J. Small Anim. Pract. 32, 306–311.

Fuse, H., Takahara, M., Ishii, H., Sumiya, H., Shimazaki, J., 1990. Measurement of testicular volume by ultrasonography.Int. J. Androl. 13, 267–272.


Gumbsch, P., Gabler, C., Holzmann, A., 2002. Colour-coded duplex sonography of the testes of dogs. Vet. Rec. 151,140–144.

Handelsman, D.J., Staraj, S., 1985. Testicular size: the effects of aging, malnutrition, and illness. J. Androl. 6, 144–151.Hansen, P.F., With, T.K., 1952. Clinical measurement of the testes in boys and men. Acta Med. Scand., 457–465.Lambert, B., 1951. The frequency of mumps and of mumps orchitis and the consequences for sexuality and fertility. Acta

Genet. Stat. Med. 2, 1–166.Lenz, S., Giwercman, A., Elsborg, A., Cohr, K.H., Jelnes, J.E., Carlsen, E., Skakkebaek, N.E., 1993. Ultrasonic testicular

texture and size in 444 men from the general population: correlation to semen quality. Eur. Urol. 24, 231–238.Lin, L.I., 1989. A concordance correlation coefficient to evaluate reproducibility. Biometrics 45, 255–268.Olar, T.T., Amann, R.P., Pickett, B.W., 1983. Relationships among testicular size, daily production and output of sperma-

tozoa, and extragonadal spermatozoa reserves of the dog. Biol. Reprod. 29, 1114–1120.Paltiel, H.J., Diamond, D.A., Di Canzio, J., Zurakowski, D., Borer, J.G., Atala, A., 2002. Testicular volume: comparison

of orchidometer and US measurements in dogs. Radiology 222, 114–119.Prader, A., 1966. Testicular size: assessment and clinical importance. Triangle 7, 240–243.Pugh, C.R., Konde, L.G., Park, R.D., 1990. Testicular ultrasound in the normal dog. Vet. Radiol. 31, 195–199.Rivkees, S.A., Hall, D.A., Boepple, P.A., Crawford, J.D., 1987. Accuracy and reproducibility of clinical measures of

testicular volume. J. Pediatr. 110, 914–917.Schiff, J.D., Li, P.S., Goldstein, M., 2004. Correlation of ultrasonographic and orchidometer measurements of testis

volume in adults. BJU Int. 93, 1015–1017.Setchell, B.P., Brooks, D.E., 1988. Anatomy, vasculature, innervation and fluids of the male reproductive tract. In: Knobil,

E., Neill, J.D. (Eds.), The Physiology of Reproduction. Raven, New York, NY, pp. 753–836.Shiraishi, K., Takihara, H., Kamiryo, Y., Naito, K., 2005. Usefulness and limitation of punched-out orchidometer in

testicular volume measurement. Asian J. Androl. 7, 77–80.Takihara, H., Sakatoku, J., Fujii, M., Nasu, T., Cosentino, M., Cockett, A.T., 1983. Significance of testicular size

measurement in andrology. I. A new orchidometer and its clinical application. Fertil. Steril. 39, 836–840.Taskinen, S., Taavitsainen, M., Wikstrom, S., 1996. Measurement of testicular volume: comparison of 3 different methods.

J. Urol. 155, 930–933.Tatsunami, S., Matsumiya, K., Tsujimura, A., Itoh, N., Sasao, T., Koh, E., Maeda, Y., Eguchi, J., Takehara, K., Nishida,

T., Miyano, S., Tabata, C., Iwamoto, T., 2006. Inter/intra investigator variation in orchidometric measurements oftesticular volume by ten investigators from five institutions. Asian J. Androl. 8, 373–378.

Documents

Comparison between ultrasonographic and caliper measurements of testicular volume in the dog