Key words
elastography - ultrasonography - testicular microlithiasis - scrotum - testis cancer
Introduction
Ultrasonography (US) is the preferred and widely accepted first choice imaging modality
for the evaluation of testicular abnormalities [1]. B-mode US has a very high sensitivity when investigating the scrotum. Furthermore,
new additions to B-mode US such as color Doppler, contrast-enhanced ultrasound, and
recently elastography have been developed over time [2]. Elastography is a noninvasive method to evaluate tissue stiffness [3], which can be assessed by either strain or shear wave. Strain elastography uses
external transducer compression and continuously compresses the tissue of interest
[4], whereas shear wave is a quantitative method [5]
[6]
[7]. Since lesions in testicles are superficial, both elastography methods could be
used. However, elastography should be recommended as part of a multiparametric ultrasound
examination approach, also suggested by Sidhu [8].
Elastography is used in various organs [7]
[9]
[10]
[11] and also in the testicles for investigating testicular torsion [12]
[13], testicular cancer [14]
[15] infarction [16]
[17], epidermoid cysts [18], undescended testes [19], infertile men [20], hemodialysis patients [21], testicular microlithiasis, and other lesions [22]
[23]
[24]. It can be challenging to distinguish hard tumors in soft tissue using only B-mode
ultrasound. Therefore, elastography is instrumental in differentiating normal from
pathological tissue [5]. In an elasticity phantom comparing strain and shear wave elastography, Carlsen
et al. [25] showed that the two methods were equally applicable in intermediate levels of elasticity,
but strain was better in hard and soft targets.
The aim of this study was to compare shear wave elastography (SWE) in patients with
normal testicular tissue, TML, and testicular cancer.
Materials and methods
Study population
A total of 248 consecutive patients from the Department of Radiology were included
during the period between September 2013 and March 2016. All included patients were
outpatients referred by their general practitioner for a clinical US scrotal examination
due to symptoms such as pain, discomfort, lump or testicular mass. The exclusion criteria
were torsion, inflammation, previously diagnosed testicular cancer, an age of less
than 18 years or any linguistic difficulties regarding understanding of the informed
consent. The patients were divided into three groups: normal testicular tissue, TML,
and testicular cancer.
Elastography image technique
All patients underwent B-mode US investigation followed by SWE. In order to avoid
signal artifacts, care was taken not to compress the testicle during the elastography
investigation. SRR, MRP, or one of four certified radiologists (all had more than
five years of ultrasound experience and between six months to 5 years of elastography
experience) performed the SWE measurements. First a B-mode investigation was performed
followed by SWE. All examinations were performed separately for the right and left
testicle and on the same ultrasound machine. The three SWE measurements were performed
in the center, upper and lower pole. In patients with testicular cancer, the three
SWE measurements were performed in the center, upper and lower pole of the lesion.
If an elastography measurement error occurred, a new measurement was performed within
seconds. We used a 9 MHz linear 9L4 probe for all examinations (Siemens S3000 ultrasound
machine (Acuson Corporation, Siemens, Mountain View, CA, USA) with Virtual TouchTM Tissue Quantification software). The frame rate per second (frs) was 14; the thermal
index for soft tissue (TIS) was 0.8; the thermal index for bone (TIB) was 1.2; and
the mechanical index was 1.6. A detection pulse measured the speed of the shear waves,
and the shear wave velocity was quantified in meters per second from the region of
interest (ROI). A B-mode image was used to place the ROI in order to perform elastography.
The ROI size was a 10×10 mm box, which could not be altered. Three shear wave velocity
measurements were performed in each testicle with the patient in the supine position.
The SWE results are presented in meters per second (m/s). All images and data were
recorded in the Picture Archive Communication System (PACS, Easyviz Impax Workstation,
Medical Insight, Valby, Denmark). Elastography is easily performed without discomfort
or pain to the patient.
Statistical analysis
We estimated the mean testicular tissue stiffness in the three groups of men using
a linear mixed effects model to account for the cluster structure and estimated random
variation. Data constituted a hierarchy of clusters of correlated observations: multiple
measurements performed on the same testicle have a higher similarity than in comparison
with those performed on the contralateral testicles. Likewise, the measurements in
the same man are expected to be more similar than if compared with those performed
in other men. Using the linear mixed effects model, it is possible to take random
variation into account that can be attributed to heterogeneity between men, and also
between testicles in the estimated elastography values.
Age and TML were adjusted for in the analysis, but neither was significant in the
linear mixed effects model and hence were removed from the model. The linear mixed
effects model was tested for normality by qnorm-plots of the residuals, and showed
no cause for concern. The proportions of random variation attributed to inhomogeneity
between men and between testicles (within each man), and measurement error stemming
from triple measurements on each testicle was estimated.
A p-value ≤0.05 was considered statistically significant. STATA statistical software
(version 14.1, STATA Corporation, College Station, TX, USA) was used for the analysis.
Ethics and approvals
The Danish Data Protection Agency and The Regional Scientific Ethical Committees for
Southern Denmark (ID: S-20120144) approved the study. All patients signed an informed
consent after receiving both oral and written information.
Results
We included 248 consecutive patients with a total of 492 testicles (246 right testicles
(146 (59.3%) normal tissue, 90 (36.6%) TML, 10 (4.1%) cancers) and 246 left testicles
(159 (64.6%) normal tissue, 78 (31.7%) TML, 9 (3.7%) cancers). One of the 19 patients
with testicular cancer had bilateral cancer.
The mean age in men with normal testicular tissue was 47.6 (range: 20–79 years), and
in men with TML and testicular cancer it was 47.3 (19–94 years) and 39.7 (25–79 years),
respectively.
We found a statistically significantly higher estimated mean velocity in the testicular
cancer group compared to men with normal testicular tissue and TML (p<0.001). Also,
men with TML had a higher estimated mean velocity compared to men with normal tissue:
0.79 m/s vs. 0.76 m/s (p=0.027) ([Table 1]).
Fig. 1 a, b Shear wave elastography of the left and right testicle of a 27-year-old man with
no TML referred to an ultrasound of the scrotum due to pain.
Fig. 2 a, b Shear wave elastography of the left and right testicle in a 30-year-old man with
multiple TML (grade 3) in both testicles. TML was diagnosed about a year ago. Referred
to our Department of Radiology due to suspicion of a lump in the right testicle.
Fig. 3 a, b Shear wave elastography of both testicles in a 31-year-old man with a tumour in the
right testicle. The tumour was a seminoma with no TML. The patient was referred due
to tumour suspicion.
Fig. 4 a, b Shear wave elastography of both testicles in a 29-year-old man with bilateral TML
and a tumour in the left testicle. Referred to the Department of Radiology due to
a lump in the left testicle. The tumour was a seminoma.
Table 1 Estimated shear wave elastography stiffness in testicles with normal tissue, testicular
microlithiasis, and cancer.
|
Estimated values*
|
Mean (m/s)
|
S.E.
|
95% CI
|
P-value
|
|
Normal testicular tissue (n=130)
|
0.76
|
0.006
|
0.75–0.78
|
0.027
|
|
Testicular microlithiasis (n=99)
|
0.79
|
0.008
|
0.77–0.81
|
0.02
|
|
Testicular cancer (n=19)
|
1.92
|
0.059
|
1.80–2.03
|
<0.001
|
Abbreviation: m/s=meters per second; S.E.=standard error; CI=confidence interval;
TML=testicular microlithiasis
* The mean testicular stiffness values predicted by the random effects models when
adjusted for random effects among patients, between testicles, and measurement errors
In the group of patients with testicular cancer, the unaffected testicle had an observed
mean value of 0.77 m/s. In men with unilateral TML, the observed mean of the unaffected
testicle (without TML) was 0.81 m/s. The estimated elastography findings in the three
groups of men are summarized in [Table 1]. [Table 2] shows the estimated means from the three groups using the linear mixed effects model.
Table 2 Random variations between the patients and the testicles from the empirical elastography
measurements.
|
Source of random variation
|
Variance
|
Proportion variance (%)
|
|
Between patients
|
0.0026242
|
2.6
|
|
Between testicles
|
0.0129973
|
62.1
|
|
Residual/measurement error
|
0.0098929
|
35.3
|
|
Total
|
0.0255144
|
100
|
Nine testicles with cancer were diagnosed with seminoma and 10 testicles with non-seminoma.
TML was present in one patient with non-seminoma and in four with seminoma. The tumors
had a mean diameter of 3.6 cm (range: 0.6–8.2 cm)
[Fig. 1]
[]
[]
[4].
Discussion
Main findings
We found a higher estimated statistically significant stiffness in patients with testicular
cancer compared to men with normal tissue. Additionally, the stiffness in men with
TML was higher than in men with normal tissue. This difference, however, is considered
too small to have any clinical relevance.
Comparison with other studies
A study of thyroid nodules showed that the presence of calcifications significantly
increased the stiffness [26]. We also found men with TML to have increased stiffness compared to men with normal
tissue (p=0.027). In our study heterogeneity between testicles belonging to the same
man was responsible for nearly two-thirds of random variation (62%), but measurement
error (35%) and heterogeneity between testicles (3%) also mattered. Although the difference
in stiffness between testicles with normal tissue and TML was statistically significant,
it was very modest and unlikely to be of clinical importance. An explanation may be
the deposition of microcalcifications throughout the testicle, as also in the thyroid
gland [26]. This consideration is also shared in the recently published guidelines on TML management,
recommending that a limited group of men with risk factors be offered US follow-up,
since epidemiological studies and recent reports do not consider TML to be premalignant
[27].
Overall, the observed mean velocity in men with testicular cancer was 2.10 m/s, but
we found a calculated predicted mean of 1.92 m/s. Trottmann et al. investigated seven
patients with germ cell tumors and five with seminoma using virtual touch imaging
quantification (VTIQ), and the results were a mean shear wave velocity of 1.94 m/s
in germ cell tumors and 2.42 m/s in seminoma [22]. Macron et al. investigated the testes of 20 healthy patients using SWE and found
a mean SWE of 0.81 m/s and a mean VTIQ of 1.07 m/s [28]. VTIQ seems to measure the stiffness higher than SWE, which could be caused by the
very small ROI of the VTIQ. Both studies had small sample sizes and did not take random
effects into account. Variation could also be related to the ultrasound machine, transducer,
and software being used. Trottman et al. also showed a statistically significant difference
between mean shear wave velocity in normal testicular tissue (1.17 m/s), seminomas
(2.42 m/s) and germ cell tumors (1.94 m/s) (p=0.002) [22]. An explanation for the observed velocity of 1.17 m/s in normal testicular tissue
compared to our estimated mean of 0.76 m/s could be different elastography ultrasound
machines. Another study [29] investigated 60 healthy males and found velocities at the center of the testicle
to be significantly lower (0.90 m/s) than the mean velocities of 1.15 m/s in the inferior
and superior parts of the testicles, respectively (p<0.001). Whether this applies
to testicular tumor tissue remains unknown.
Shin et al. compared shear wave velocities between different ultrasound machines and
transducers in a phantom and found the difference in mean shear wave velocities between
three machines to be statistically significant (p≤0.002) [30]. This implies that caution is needed when using absolute cut-off values in the comparison
of measurements from different ultrasound machines or transducers.
The main strengths of this study are the large number of patients and the inclusion
of random variation in the data analysis. However, using random variation requires
caution in the comparison of estimated results with observed results. We consider
the use of SWE to be an asset, since it does not use manual transducer compression
and therefore the testicular tissue is not displaced during measurement. Hence, SWE
is operator-independent. The main limitation of this single center study is that it
was only possible to obtain three velocity measurements in each testicle due to the
daily workflow in our radiology department. More measurements would have yielded more
precise estimates. This was also confirmed by our linear mixed effects model, which
showed that increasing the number of measurements rather than the number of patients
would increase the precision of the mean velocity from the three groups. Another limitation
may be the relatively small number of patients with testicular malignancy (n=19) included
in the study. However, the mean diameter of the tumors was large in our study population,
and the ROI could easily fit most lesions. Further studies including a higher number
of patients with testicular lesions are warranted.
In conclusion, our study shows that the presence of TML increases stiffness slightly,
but the range is still within that of normal testicles. Increased velocity may indicate
testicular malignancy in testicular lesions, and the use of elastography seems a promising
method for the evaluation of testicular lesions.
