- Email: [email protected]
Qualitative and quantitative contrast-enhanced ultrasonography for the characterisation of non-palpable testicular tumours
Clinical Radiology 73 (2018) 322.e1e322.e9
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Qualitative and quantitative contrast-enhanced ultrasonography for the characterisation of non-palpable testicular tumours A. Luzurier a, *, F. Maxwell a, J.M. Correas b, c, d, G. Benoit e, V. Izard f, S. Ferlicot g, J.P. Teglas h, M.F. Bellin a, e, L. Rocher a, d, e Department of Adult Diagnosis and Interventional Radiology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France b Department of Adult Radiology, Necker University Hospital, 149 rue de S evres, 75015 Paris, France c Paris Descartes University, 15 Rue de l’Ecole de M edecine, 75006 Paris, France d Institut Langevin, ESPCI Paris, PSL Research University CNRS UMR 7587, INSERM U979, 17 rue Moreau, 75012 Paris, France e Paris South Medical University, 63 rue Gabriel P eri, 94270 Le Kremlin Bicetre, France f Department of Urology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France g Department of Pathology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France h Department of Public Health and Epidemiology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France a
ˇ
ˇ
ˇ
ˇ
ˇ
article in formation Article history: Received 27 February 2017 Received in revised form 15 August 2017 Accepted 9 October 2017
AIM: To assess the diagnostic performance of conventional ultrasound (US) and contrastenhanced ultrasonography (CEUS) in the differential diagnosis of non-palpable intratesticular tumours. MATERIALS AND METHODS: The local ethics review board approved the protocol, and all of the patients provided written informed consent. Between December 2011 and February 2014, men with non-palpable testicular tumours and normal tumour markers who were referred for surgery were included. The tumours were analysed by conventional US, including B-mode and colour Doppler US (CDUS) as well as by CEUS. Morphological aspects and qualitative and quantitative CEUS criteria, based on visual enhancement and timeeintensity curves, were assessed for each lesion. RESULTS: Forty patients were ultimately included. Based on histopathological results, the tumours were classified into three groups: benign tumours (n¼16), malignant tumours (n¼15), and burned-out tumours (n¼9). In B-mode, the morphological aspects were significantly different between benign and malignant tumours (p-values from 0.0002 to 0.008). Qualitative and quantitative analyses of the CEUS images revealed that burned-out tumours exhibited significantly less enhancement than malignant and benign tumours: in burned-out tumours, timeeintensity curves were flat, whereas in both benign and malignant tumours the curves had a bell-shaped pattern. All intensity parameters were lower for burned-out tumours compared to benign and malignant tumours (p-value from 0.0001 to 0.026). Both benign and
ne ral * Guarantor and correspondent: A. Luzurier, Department of Adult Diagnostic and Inteventional Radiology, Bicetre University Hospital, 78 avenue du Ge Leclerc, 94275 Le Kremlin Bicetre, France. Tel.: þ33145213402; fax: þ33154213209. E-mail address: [email protected] (A. Luzurier). https://doi.org/10.1016/j.crad.2017.10.007 0009-9260/Ó 2017 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
322.e2
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
malignant tumours enhanced strongly, however, and no significant difference between the two was noted (p-value from 0.0721 to 0.0953). CONCLUSION: Unlike conventional US, which enable benign lesions to be differentiated from malignant or burned-out tumours, CEUS failed to enabled differentiation between benign lesions and malignant vascularised testicular tumours. CEUS appears to have the potential, however, to differentiate burned-out tumours from vascularised testicular tumours. Ó 2017 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction The testes are the most frequent site of malignant tumours in males who are in their third decade of life. Histologically, 90e95% of these lesions are germ cell tumours, such as seminomas (30e40%), and non-seminomatous tumours (60e70%).1 The remaining 5e10% of testicular lesions are stromal tumours (mostly Leydig and Sertoli cell tumours), which are benign in 90% of cases.2 In rare cases, advanced germ cell tumours can undergo fibrous involution. These lesions are called burned-out tumours (BOT).3e5 The discovery of non-palpable testicular lesions is becoming more common, as testicular ultrasonography (US) indications have become broader; becoming mandatory, for instance, in the male infertility workup. Historically, testicular focal lesions have been treated by radical orchiectomy, as testicular biopsies are not recommended.6 In this population of young men, preservation of fertility is an important consideration. Furthermore, whereas the most common type of palpable testicular lesions in young men are malignant carcinomas, >80% of non-palpable testicular lesions discovered at US may in fact be benign.7 Technical advances in US have allowed for better characterisation of testicular lesions, and management of nonpalpable testicular lesions has hence changed drastically. Nowadays, testis-sparing surgery with frozen section analysis is deemed to be appropriate, provided that the lesion measures <25 mm across and that it is not in contact with the rete testis, in order to preserve testicular vascularisation.8,9 Moreover, the European Society of Urogenital Radiology (ESUR) scrotal imaging subcommittee recommends active monitoring of small testicular lesions (i.e., <5 mm across) in young/infertile patients, provided they have benign US patterns and negative tumour markers.7 Unenhanced B-mode US is the reference standard for the characterisation of testicular lesions, particularly to describe their morphological features; while colour Doppler is the preferred mode for visualising their macrovascularisation.10 Scrotal MRI and shear-wave elastography are promising imaging techniques, although they are still undergoing evaluation.7,11e14 Over the last decade, contrastenhanced US (CEUS) has proven to be suitable to analyse microvascularisation in various organs. Indeed, CEUS provides qualitative results as well as quantitative data with timeeintensity curves extracted from the image using specific software.15,16 There have only been a few studies, however, demonstrating the suitability of CEUS for the
characterisation of focal testicular tumours. Thus the aim of the present study was hence to assess the diagnostic performance of conventional US and CEUS to investigate nonpalpable testicular tumours, using qualitative evaluation and quantitative parameters.
Materials and methods Population and study design de The local ethics committee approved the study (Comite protection des personnes, Ile de France VII, CPP 13 051), and all of the patients provided written informed consent. This monocentric study was carried out between December 2011 and February 2014. Of the 3,256 inpatients and outpatients referred for scrotal US during this period, 49 patients who were diagnosed with a non-palpable testicular focal tumour discovered during US and referred for surgery, were enrolled consecutively and prospectively underwent CEUS. Patients with non-tumorous lesions, such as haematomas, infarctions, adrenal rests, and focal orchitis, were not part of this study: those diagnoses were performed based on clinical and biological data and follow-up. After the CEUS, nine patients were secondarily excluded in light of movement artefacts (due either to the probe or the patient) that compromised the data (n¼2), the small size of the lesion precluding being able to define an appropriate region of interest (ROI; n¼5), or when the lesion was too close to the hilum due to overlap with the vessels’ convergence (n¼2). Thus, ultimately a total of 40 patients were included (Fig 1). The final diagnosis was based on histological criteria (radical orchiectomy or tumorectomy). Three groups were identified: benign tumours, malignant tumours, and burned-out tumours (BOT). Five patients with BOT had been included in a case series of 10 patients with BOT that was revealed during investigations for infertility.5 Qualitative and quantitative CEUS evaluations were not reported in this published study.
Conventional US and CEUS The testicular US was performed by two radiologists with expertise in urogenital US and qualified in the use of CEUS. The radiologists reviewed the standardised reports, images, and videos in the picture archiving and communication system (PACS). An Aplio 500 US device (Toshiba Medical Systems Co., Ltd, Nasu, Japan), with two superficial probes
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
322.e3
Figure 1 Flow chart. MT, malignant tumours; BT, benign tumours; BOT, burned-out tumours.
(a 5-12 MHz linear array and an 8-18 MHZ linear array) was used. Computer-assisted calculation of the testicular volume was based on use of the following formula: volume ¼ length width depth 0.523. Using B-mode, the testicular volume; the tumour characteristics including size, echogenicity (e.g., hypoechoic or not), and margins (e.g., irregular or not); the features of the adjacent parenchyma (e.g., homogeneous or heterogeneous); and the presence of microlithiasis. Vascularisation was studied using colour Doppler US. To adequately evaluate blood flow, the Doppler parameters were adjusted to the most sensitive settings that did not introduce artefacts using a 3.5 cm/s scale. The pattern of vascularisation was categorised as peripheral, central, or mixed (peripheral and central). CEUS was performed by a bolus injection of 4.8 ml of Sonovue (Bracco Imaging, France) into a superficial vein of the forearm. The intravenous tube was immediately flushed with isotonic saline solution. Image acquisition lasted at least 3 minutes or until the uptake disappeared completely, and it was performed at a low mechanical index (<0.5). Video clips were sent for perfusion software processing (VueBox 4.3, Bracco Suisse SA, Geneva, Switzerland). The first step was qualitative analysis: for each lesion, the uptake of contrast was assessed and noted as being faster, similar, or delayed compared to the adjacent parenchyma. Video clips were then analysed using the perfusion software to determine quantitative perfusion patterns (VueBox 4.3, Bracco Suisse SA, Geneva, Switzerland). Using a ROIs placed (in consensus) in the lesion and in the adjacent parenchyma, the software extracted timeeintensity curves and 11 quantitative perfusion parameters15: (1) peak enhancement (a.u.); (2) mean time transit: full-width at
half-maximum of the curve (s); (3) area under the curve during wash-in (a.u.); (4) rise time: the time it took for the contrast signal to increase from the baseline level to peak enhancement (s); (5) time to peak: the time needed to reach peak enhancement (s); (6) wash-in rate ¼ maximal slope, calculated as the peak enhancement / rise time (a.u.); (7) perfusion index, calculated as the area under the curve during wash-in divided by the rise time (a.u); (8) fallout time: the time needed for the descending slope to reach a contrast signal intensity of zero (s); (9) wash-out rate ¼ minimal slope (a.u.); (10) area under the curve during wash-out (a.u.); (11) area under the curve during wash-in and wash-out (a.u)
Statistical analysis The three groups, i.e., malignant tumours, benign tumours, and BOT, were compared using Fisher’s exact test for non-continuous variables (unenhanced US criteria and qualitative CEUS), and with the KruskaleWallis test for continuous variables (testicular volume, lesion size, and CEUS criteria).
Results Clinical data The clinical data are presented in Table 1. Forty patients who were 20e58 years of age (median age of 35.5) were included. Twenty-three patients (57.5%) were referred for hypofertility, seven (17.5%) due to a history of contralateral scrotal lesion, and two (5%) had a history of retroperitoneal germ cell tumours. Five patients (12.5%) consulted for scrotal pain, one (2.5%) for haemospermia, one (2.5%) for gynaecomastia, and one (2.5%) prior to gender
322.e4
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
Table 1 Clinical characteristics.
No. of patients Age (years) Size (mm) Patient with metastatic statusa Radical orchiectomya
Malignant tumour
Seminomas
Malignant Sertoli cell tumours
Benign tumours
Benign Leydig cell tumours
Benign Sertoli cell tumours
BOT
15 39 (29; 53) 11 (6; 29) 0 (0%) 14 (93%)
14 37 (29; 53) 10.5 (6; 29) 0 (0%) 13 (93%)
1 42 20 0 (0%) 1 (100%)
16 32.5 (20; 47) 5.8(2; 13) 0 (0%) 2 (12.5%)
15 32.5 (20; 47) 6(2; 13) 0 (0%) 1 (6.7%)
1 29 7.5 0 (0%) 1 (100%)
9 40 (20; 58) 16(10; 22) 3 (33%) 9 (100%)
Unless indicated otherwise, data are medians with ranges in parentheses. a The values represent the number of patients with the percentage in parentheses.
reassignment. Five patients (12.5%) had a single testis. Sixteen patients (40%) had a benign tumour: 15 Leydig cell tumours and one Sertoli cell tumour. Fifteen patients (37.5%) had a malignant tumour: 14 seminomas and one patient had a malignant Sertoli cell tumour. None of these patients had metastatic retroperitoneal lymph nodes. Nine patients (22.5%) had a BOT. Three of these had metastatic retroperitoneal lymph nodes (two with metastatic seminomatous tumours, and one patient had a metastatic nonseminomatous tumour (a teratoma). Based on clinical, biological, US, and magnetic resonance imaging (MRI) findings, patients with a lesion suspected of being malignant or BOT underwent radical orchiectomy (14/ 15 malignant tumours and 9/9 BOT), except for one patient with a single testis who was managed by testis-sparing surgery. Lesions suspected of being benign were managed by tumorectomy (n¼14/15).
Unenhanced US findings The US features are presented in Table 2. The morphological features of malignant tumours and BOT differed significantly from those of benign tumours. The benign tumours (median size: 5.8 mm; min: 2 mm; max: 13 mm) were significantly smaller than both the BOT (median size: 16 mm; min: 10 mm; max: 22 mm; p¼0.0001) and the malignant tumours (median size: 11 mm; min: 6 mm; max: 29 mm; p¼0.0002). No significant difference was noted in terms of the testicular volume. All of the lesions were
hypoechoic. Irregular margins were significantly more frequent with both malignant tumours (n¼11, p<0.0001) and BOT (n¼9, p<0.0001) than with benign tumours (n¼1). The adjacent parenchyma were more often heterogeneous with both malignant tumours (n¼11, p<0.0001) and BOT (n¼9, p<0.0001) than with benign tumours (n¼0). Microlithiasis was significantly more frequent with malignant tumours (n¼13) than with benign tumours (n¼2; p<0.001). Based on colour Doppler imaging, peripheral vascularisation was more frequent with benign tumours (n¼15) than with both malignant tumours (n¼3, p<0.0001) and burnedout tumours (n¼0, p<0.0001). Based on colour Doppler imaging, none of the BOT were vascularised. Examples of this are shown in Fig 2.
CEUS findings The qualitative findings are presented in Table 3. No significant differences were observed between benign and malignant tumours. Faster wash-in was observed in 11 benign and 11 malignant tumours (p¼1.000). Five benign tumours and two malignant tumours had a similar level of enhancement (p¼0.394). Delayed enhancement was observed in two malignant tumours (p¼0.226), but not in benign tumours. Conversely, delayed enhancement was significantly more frequent in BOT (n¼6) than in either malignant tumours (n¼2; p¼0.021) or benign tumours (n¼0; p<0.0001). The quantitative findings are presented in Table 4. Both benign and malignant tumours enhanced
Table 2 Characteristics of lesions (in B-mode US and colour Doppler). Groups
p-Value p-Value
Benign (n¼16) Malignant (n¼15) Burned-out (n¼9) Lesion size (mm) Testicular volume (ml) Hypoechogenicityc Irregular marginsc (%) Mixed vascularizationc Peripheral vascularizationc Heterogeneity of the adjacent parenchymac Microlithiasis
Benign vs. malignant Malignant vs. BOT Benign vs. BOT
5.8(2; 13) 10.4 (6;22.6) 16(100%) 1(6%) 0 15(94%) 0
11 (6;29) 10.8 (5.6;18.2) 15(100%) 11(73%) 12(80%) 3(20%) 11(73%)
16(10; 22) 9 (6.1;11) 9(100%) 9(100%) 0 0 9(100%)
0.0001a 0.3877a 0.000b 0.000b 0.000b 0.000b
0.0002a 0.9056a 0.008b 0.001b 0.008b 0.001b
0.0001a 0.2949a 0.259b 0.000b 0.266b 0.253b
0.2831a 0.1524a 0.000b 0.000b 0.000b
2(12%)
13(87%)
5(56%
0.000b
0.002b
0.150b
0.058b
Unless indicated otherwise, data are medians with ranges in parentheses. a p-Value after comparison of the three groups with the non-parametric KruskaleWallis test. b p-Value after comparison of the three groups with Fisher’s exact test. c The values represent the number of patients, percentages in parentheses.
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
322.e5
Figure 2 Unenhanced contrast US: B-mode and Doppler. (a) US imaging of a benign testicular tumour: B-mode reveals a homogeneous hypoechoic lesion, with regular margins, and homogeneous adjacent parenchyma without microlithiasis. Vascularization is mostly peripheral on colour-coded Doppler. (b) Malignant testicular tumour: B-mode reveals a heterogeneous hypoechoic lesion with irregular margins and microlithiasis in the adjacent parenchyma. Colour-Doppler shows that vascularization is mixed, both central and peripheral. (c) Burned-out tumour: B-mode reveals a heterogeneous hypoechoic lesion with irregular margins and heterogeneous adjacent parenchyma. No vascularization is seen using power Doppler.
strongly, with a pronounced wash-in immediately followed by a wash-out, as shown by the timeeintensity curves (Figs 3 and 4); however, the quantitative parameters were not significantly different between benign and malignant tumours. BOT exhibited a weaker level of enhancement
than both benign and malignant lesions, with an almost flat timeeintensity curve (Fig 5). Nine parameters were significantly different between BOT, benign, and malignant tumours (Table 4). The peak enhancement, wash-in and wash-out rates, perfusion index,
Table 3 Qualitative analysis. Groups
Rapid wash-in Similar wash-in Delayed wash-in
Benign (n¼16)
Malignant (n¼15)
Burned-out (n¼9)
11(69%) 5(31%) 0
11(73%) 2(13%) 2(13%)
0 3(33.3%) 6(67%)
Data represent number of lesions with percentage in parentheses. a p-Value after comparison with Fisher’s exact test.
p-Valuea
p-Valuea Benign vs. Malignant
Malignant vs. BOT
Benign vs. BOT
0.001 0.565 0.000
1.000 0.394 0.226
0.001 0.615 0.021
0.001 1 0.000
322.e6
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
Table 4 Quantitative parameters. Benign (n¼16)
Peak enhancement (a.u.) Wash-in rate (a.u.) Area under the curve during wash-in Rise time (s) Time to peak (s) Mean time transit (s) Perfusion index (a.u.) Area under the curve (a.u.) Fall time (s) Wash-out rate (a.u.) Area under the curve during wash-out (a.u.)
Malignant (n¼15)
Burned-out (n¼9)
p-Valuea
p-Valuea Benign vs. malignant
Benign vs. BOT
Malignant vs. BOT
2422 (412;12,452) 324 (5;,3673) 16191 (2,549; 68,597)
2284(81,26,018) 252 (10;3,881) 12591 (514; 16,0803)
166 (2;554) 18 (0;120) 2580 (19; 2459)
0.0004 0.0006 0.0021
0.707 0.7369 0.953
0.0001 0.0001 0.0007
0.026 0.0026 0.0038
10.2 (4.7; 19.0) 15.5 (7.7;37.6) 60.1 (22.8;110.5) 1544 (273; 7,858) 48746 (7,476;18,5521) 16.8 (9.0; 34.8) 143 (16;1,547) 31380 (4,927; 11,6924)
11.2 (5.6; 22.8) 18.9 (8.3;32.1) 49.3 (26.6;419.8) 1419 (50; 16,597) 41850 (1,398;47,7912) 26.3 (10.2; 62.5) 118 (5;1,444) 29529 (884; 317,109)
14.5 (8.1; 55.1) 26.1 (9.;63.2) 106.3 (36.4;523.4) 116 (1; 378) 1070 (0; 11,926) 25.3 (0; 34.8) 5 (0;22) 671 (0; 8,842)
0.0795 0.3022 0.0373 0.0006 0.0003 0.1927 0.0011 0.0003
0.2280 0.6781 0.6212 0.7369 0.8588 0.0721 0.5401 0.7669
0.0315 0.1570 0.0174 0.0001 0.0002 0.5006 0.0001 0.0003
0.1998 0.1998 0.0295 0.0026 0.0004 0.4014 0.0067 0.0004
Data are medians with ranges in parentheses. a p-Value after comparison with the non-parametric KruskaleWallis test.
Figure 3 Peak enhancement in a Leydig cell tumour. Note the difference between the yellow curve (the lesion), and the pink curve (adjacent parenchyma).
and areas under the curve were significantly lower in BOT than in either benign tumours or malignant tumours (the pvalues ranged from 0.0001 to 0.0067). The mean time transit was significantly higher in BOT than in either benign tumours (p¼0.0174) or malignant tumours (p¼0.0295). The rise time was significantly higher in BOT than in benign tumours (p¼0.0315), while there was no difference between BOT or malignant tumours (p¼0.1998).
Discussion The findings of the present study show that unenhanced US can be used to identify features that enable benign lesions to be differentiated from malignant and BOT. Thus, irregular margins and heterogeneity of the adjacent
parenchyma were significantly more frequent in malignant and BOT than in benign tumours, whereas benign lesions were well-circumscribed with homogeneous adjacent pulp. With colour Doppler imaging, peripheral vascularisation was significantly associated with benign tumours, whereas mixed vascularisation was significantly more frequent in malignant lesions. By contrast, colour Doppler US did not reveal vascularisation of BOT. CEUS is an imaging technique that involves microbubble contrast agents that remain strictly intravascular; therefore, it is a good technique for studying microvascularisation. The present study has revealed significant differences in the enhancement of BOT compared to benign and malignant testicular tumours. Both malignant and benign testicular tumours exhibited strong enhancement, characterised by substantial wash-in and wash-out on quantitative and
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
322.e7
Figure 4 Peak enhancement in a seminoma. Note the strong enhancement of the seminoma (yellow curve) compared to the adjacent parenchyma (pink curve).
Figure 5 CEUS in a burned-out tumour: The lesion (yellow curve) does not enhance compared to the adjacent parenchyma (pink curve).
qualitative CEUS. Conversely, enhancement of BOTs was poor, as reflected by flat timeeintensity curves and low perfusion parameters. BOT are primary germ cell tumours of the testes that are replaced by fibrosis after having undergone partial or total regression. Most of them are diagnosed belatedly, being detected by the discovery of metastases such as retroperitoneal lymphadenopathies.17e19 The regression of malignant
cells and their replacement by fibrosis results in poorly vascularised areas that are reflected on CEUS as an absence of contrast medium uptake after injection of microbubbles, a flat timeeintensity curve, and low perfusion parameters. Non-tumorous lesions, such as haematomas or infarctions, are also known to exhibit a lack of perfusion on CEUS.20,21 They are, however, associated with sudden pain and a change in the US pattern during follow-up.
322.e8
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
BOT are usually diagnosed at B-mode US when a large testicular lesion with irregular margins and heterogeneous adjacent parenchyma with microlithiasis is seen.3,17e19,22 Although the differential diagnosis of malignant tumours can be difficult with B-mode US, colour Doppler US is generally sufficient to achieve this, as it can reveal the hypovascularisation in BOT that does not occur in most malignant tumours. This study showed that in some difficult cases CEUS provided imaging data that enabled BOT to be differentiated from active benign and malignant tumours. Unlike BOT, both benign and malignant tumours tended to enhance strongly, with a pronounced wash-in followed by a wash-out, which is in accordance with the findings of Lock et al. and Huang and Sidhu.23,24 Contrary to previous results, however, in the present study none of the qualitative and quantitative parameters were statistically different between benign and malignant tumours. Isidori et al. found that malignant tumours exhibited a rapid wash-out, whereas delayed wash-out was observed in benign tumours.20 Rapid wash-out in malignant tumours could be explained by a more anarchic vascularisation and the presence of many shunts. In the study by Drudi et al., the wash-in of Leydig cell tumours was higher and more rapid than seminomas, as demonstrated by a shorter time to peak, higher peak enhancement, and wash-in rate in Leydig cell tumours. This difference is thought to be due to a high blood flow velocity and an increased microvessel density compared to seminomas, despite the overlapping values.25 In the present study, neither the wash-in nor the wash-out were different between benign (mostly Leydig cell tumours) and malignant tumours (mostly seminomas). These contradictory results reveal that microvascularisation of testicular tumours is difficult to predict by CEUS, and this is presumably due to the histological heterogeneity of malignant lesions. Benign tumours, which primarily comprise Leydig cell tumours, are generally homogeneous lesions with no necrotic, haemorrhagic, or cystic component.10 These tumours present a rich vascularisation which explains the substantial enhancement seen with microbubbles. A pronounced uptake of contrast with a quick wash-out was observed in all of such lesions in our study. On the other hand, the range of malignant tumours is more variable. Non-seminomatous tumours are known to be heterogeneous, with necrosis, cystic areas, and calcifications; whereas seminomas are generally more homogeneous.26,27 These different histological features are reflected as different enhancement behaviours with CEUS. In the present study, 97.7% of the malignant tumours were seminomas (n¼15/16), which enhanced strongly. The results may well have been different with a higher proportion of non-seminomatous tumours. The age of the present patients (median of 39 years of age), and the inclusion of non-palpable lesions only can explain this low proportion of non-seminomatous tumours. Regardless of the histological type, some germ cell tumours can partially regress and undergo fibrous involution before turning into BOT. Germ cell tumours with no fibrous involution should enhance strongly, while those that are
partially fibrous should exhibit less enhancement. These reasons explain why malignant testicular tumours do not have a stereotyped CEUS enhancement. Unlike Vandaele et al., the homogeneity or heterogeneity of wash-in and wash-out was not analysed in the lesions in light of their small size.28 Indeed, this evaluation could have been biased, as malignant tumours are generally larger than benign tumours, which are sometimes only a few millimetres in size. The main clinical challenge in this context is the distinction between lesions that will be treated by total orchiectomy, i.e., malignant tumours and BOT, from those that do not require a radical surgical treatment, i.e., benign tumours and non-tumorous lesions. As reported in the literature, non-tumorous lesions, such as haematomas or infarctions, are known to be non-vascularised by CEUS, and are thus readily recognisable if a degree of doubt persists after unenhanced US.21,23,24 According to the present study and the published literature, unenhanced US provides information that allows non-palpable benign tumours to be distinguished from malignant tumours and BOT.10 For small lesions (<5 mm) with benign US patterns and negative tumour markers in sub-fertile/young patients, the ESUR hence recommends an active monitoring instead of a surgery.7 Moreover, testissparing surgery may be appropriate for lesions measuring <25 mm in sub-fertile men.8,9 On standard US, distinguishing between malignant and BOT can be difficult, but their management is the same. CEUS adds little value in that context. CEUS could be used to distinguish benign tumours and BOT if there is a degree of doubt after unenhanced US, as their management is different, although this is rarely the case in common practice. The clinical benefit of CEUS in the investigation of nonpalpable testicular tumours is therefore poor. The relatively small number of patients in the present study limits interpretation of the present findings, as some results could have been statistically significant had more patients been included. Many lesions had to be excluded. Small lesions (i.e. with a mean size of 3 mm) were not included because they could not be accurately discerned using the CEUS mode. The placement of the ROI is difficult in small lesions, as also mentioned in the study by Isidori et al.20 Moreover, any movement of the probe or the patient makes the quantitative analysis uninterpretable. These limitations are more related to the technique than to the study itself, yet they are significant as many incidentally diagnosed testicular tumours are very small in size. Further, some lesions had to be excluded due to their localisation. Indeed, CEUS of lesions close to the hilum was affected by artefacts caused by the surrounding converging vessels. In the present study, the prevalence of benign stromal tumours (mostly Leydig cell tumours) and BOT over germ cell tumours was much higher than in the general population (germ cell tumours account for about 90e95% of all testicular tumours, whereas stromal tumours account for only about 5e10%). This difference could induce a selection bias. This difference may be due to the fact that the authors’ institution is a referral centre for testicular imaging.
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
In conclusion, this study assessed the potential of CEUS for distinguishing between non-palpable BOT and active tumours. Colour Doppler US may, however, be an adequate way to differentiate between these two kinds of tumours. Unlike unenhanced US, CEUS did not provide any imaging data that allows non-palpable benign and malignant vascularised tumours to be differentiated. Lastly, the use of CEUS on small tumours is subject to technical limitations, which could restrict its widespread use.
References 1. Manecksha RP, Fitzpatrick JM. Epidemiology of testicular cancer. BJU Int 2009;104:1329e33. 2. Grem JL, Robins HI, Wilson KS, et al. Metastatic Leydig cell tumour of the testis. Report of three cases and review of the literature. Cancer 1986;58:2116e9. 3. Fabre E, Jira H, Izard V, et al. “Burned-out” primary testicular cancer. BJU Int 2004;94:74e8. 4. Casella R, Rochlitz C, Sauter G, et al. “Burned out” testicular tumour: a rare form of germ cell neoplasias. Schweiz Med Wochenschr 1999;129:235e40. 5. Rocher L, Glas L, Bellin MF, et al. Burned-out testis tumours in asymptomatic infertile men: multiparametric sonography and MRI findings. J Ultrasound Med 2016, https://doi.org/10.7863/ultra.15.08037. s C, et al. CCAFU recommendations 2013: 6. Durand X, Rigaud J, Avance testicular germ cell cancer. Prog Urol 2013;23:145e60. 7. Rocher L, Ramchandani P, Belfield J, et al. Incidentally detected nonpalpable testicular tumours in adults at scrotal ultrasound: impact of radiological findings on management radiologic review and recommendations of the ESUR scrotal imaging subcommittee. Eur Radiol 2016;26:2268e78. 8. Palermo G, Antonucci M, Recupero SM, et al. Focal surgery in testis cancer: current state of the art. Urologia 2013;80:290e6. 9. Brunocilla E, Gentile G, Schiavina R, et al. Testis-sparing surgery for the conservative management of small testicular masses: an update. Anticancer Res 2013 Nov;33:5205e10. 10. Maxwell F, Izard V, Ferlicot S, et al. Colour Doppler and ultrasound characteristics of testicular Leydig cell tumours. Br J Radiol 2016;89:20160089. 11. Sanharawi IE, Correas J-M, Glas L, et al. Non-palpable incidentally found testicular tumours: differentiation between benign, malignant, and burned-out tumours using dynamic contrast-enhanced MRI. Eur J Radiol 2016;85:2072e82. 12. Goddi A, Sacchi A, Magistretti G, et al. Real-time tissue elastography for testicular lesion assessment. Eur Radiol 2012;22:721e30. 13. Aigner F, De Zordo T, Pallwein-Prettner L, et al. Real-time sonoelastography for the evaluation of testicular lesions. Radiology 2012;263:584e9.
322.e9
14. Zeng B, Chen F, Qiu S, et al. Application of quasistatic ultrasound elastography for examination of scrotal lesions. J Ultrasound Med 2016;35:253e61. 15. Dietrich CF, Averkiou MA, Correas J-M, et al. An EFSUMB introduction into dynamic contrast-enhanced ultrasound (DCE-US) for quantification of tumour perfusion. Ultraschall Med 2012;33:344e51. 16. Piscaglia F, Nolsøe C, Dietrich CF, et al. The EFSUMB guidelines and recommendations on the clinical practice of contrast enhanced ultrasound (CEUS): update 2011 on non-hepatic applications. Ultraschall Med 2012;33:33e59. 17. Tasu J-P, Faye N, Eschwege P, et al. Imaging of burned-out testis tumour: five new cases and review of the literature. J Ultrasound Med 2003;22:515e21. 18. Gurioli A, Oderda M, Vigna D, et al. Two cases of retroperitoneal metastasis from a completely regressed burned-out testicular cancer. Urologia 2013;80:74e9. 19. Qureshi JM, Feldman M, Wood H. Metastatic “burned-out” germ cell tumour of the testis. J Urol 2014;192:936e7. 20. Isidori AM, Pozza C, Gianfrilli D, et al. Differential diagnosis of nonpalpable testicular lesions: qualitative and quantitative contrastenhanced US of benign and malignant testicular tumours. Radiology 2014;273:606e18. €der C, Lock G, Schmidt C, et al. Real-time elastography and 21. Schro contrast-enhanced ultrasonography in the evaluation of testicular masses: a comparative prospective study. Ultrasound Med Biol 2016;42:1807e15. 22. Balalaa N, Selman M, Hassen W. Burned-out testicular tumour: a case report. Case Rep Oncol 2011;4:12e5. €der C, Schmidt C, et al. Contrast-enhanced ultrasound and 23. Lock G, Schro real-time elastography for the diagnosis of benign Leydig cell tumours of the testis d a single center report on 13 cases. Ultraschall Med 2014;35:534e9. 24. Huang DY, Sidhu PS. Focal testicular lesions: colour Doppler ultrasound, contrast-enhanced ultrasound and tissue elastography as adjuvants to the diagnosis. Br J Radiol 2012;85:41e53. 25. Drudi FM, Valentino M, Bertolotto M, et al. CEUS timeeintensity curves in the differentiation between Leydig cell carcinoma and seminoma: a multicenter study. Ultraschall Med 2015, https://doi.org/10.1055/s-00341398841. 26. Woodward PJ, Sohaey R, O’Donoghue MJ, et al. From the archives of the AFIP: tumours and tumour like lesions of the testis: radiologicepathologic correlation. RadioGraphics 2002;22:189e216. 27. Coursey Moreno C, Small WC, Camacho JC, et al. Testicular tumours: what radiologists need to knowddifferential diagnosis, staging, and management. RadioGraphics 2015;35:400e15. 28. Vandaele P, Marcelli F, Ouzzane A, et al. Testicular nodules of infertile men and contrast enhanced ultrasonography: preliminary study. Prog Urol 2015;25:274e81.
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Qualitative and quantitative contrast-enhanced ultrasonography for the characterisation of non-palpable testicular tumours A. Luzurier a, *, F. Maxwell a, J.M. Correas b, c, d, G. Benoit e, V. Izard f, S. Ferlicot g, J.P. Teglas h, M.F. Bellin a, e, L. Rocher a, d, e Department of Adult Diagnosis and Interventional Radiology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France b Department of Adult Radiology, Necker University Hospital, 149 rue de S evres, 75015 Paris, France c Paris Descartes University, 15 Rue de l’Ecole de M edecine, 75006 Paris, France d Institut Langevin, ESPCI Paris, PSL Research University CNRS UMR 7587, INSERM U979, 17 rue Moreau, 75012 Paris, France e Paris South Medical University, 63 rue Gabriel P eri, 94270 Le Kremlin Bicetre, France f Department of Urology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France g Department of Pathology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France h Department of Public Health and Epidemiology, Bicetre University Hospital, 78 avenue du G en eral Leclerc, 94270 Le Kremlin Bicetre, France a
ˇ
ˇ
ˇ
ˇ
ˇ
article in formation Article history: Received 27 February 2017 Received in revised form 15 August 2017 Accepted 9 October 2017
AIM: To assess the diagnostic performance of conventional ultrasound (US) and contrastenhanced ultrasonography (CEUS) in the differential diagnosis of non-palpable intratesticular tumours. MATERIALS AND METHODS: The local ethics review board approved the protocol, and all of the patients provided written informed consent. Between December 2011 and February 2014, men with non-palpable testicular tumours and normal tumour markers who were referred for surgery were included. The tumours were analysed by conventional US, including B-mode and colour Doppler US (CDUS) as well as by CEUS. Morphological aspects and qualitative and quantitative CEUS criteria, based on visual enhancement and timeeintensity curves, were assessed for each lesion. RESULTS: Forty patients were ultimately included. Based on histopathological results, the tumours were classified into three groups: benign tumours (n¼16), malignant tumours (n¼15), and burned-out tumours (n¼9). In B-mode, the morphological aspects were significantly different between benign and malignant tumours (p-values from 0.0002 to 0.008). Qualitative and quantitative analyses of the CEUS images revealed that burned-out tumours exhibited significantly less enhancement than malignant and benign tumours: in burned-out tumours, timeeintensity curves were flat, whereas in both benign and malignant tumours the curves had a bell-shaped pattern. All intensity parameters were lower for burned-out tumours compared to benign and malignant tumours (p-value from 0.0001 to 0.026). Both benign and
ne ral * Guarantor and correspondent: A. Luzurier, Department of Adult Diagnostic and Inteventional Radiology, Bicetre University Hospital, 78 avenue du Ge Leclerc, 94275 Le Kremlin Bicetre, France. Tel.: þ33145213402; fax: þ33154213209. E-mail address: [email protected] (A. Luzurier). https://doi.org/10.1016/j.crad.2017.10.007 0009-9260/Ó 2017 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
322.e2
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
malignant tumours enhanced strongly, however, and no significant difference between the two was noted (p-value from 0.0721 to 0.0953). CONCLUSION: Unlike conventional US, which enable benign lesions to be differentiated from malignant or burned-out tumours, CEUS failed to enabled differentiation between benign lesions and malignant vascularised testicular tumours. CEUS appears to have the potential, however, to differentiate burned-out tumours from vascularised testicular tumours. Ó 2017 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction The testes are the most frequent site of malignant tumours in males who are in their third decade of life. Histologically, 90e95% of these lesions are germ cell tumours, such as seminomas (30e40%), and non-seminomatous tumours (60e70%).1 The remaining 5e10% of testicular lesions are stromal tumours (mostly Leydig and Sertoli cell tumours), which are benign in 90% of cases.2 In rare cases, advanced germ cell tumours can undergo fibrous involution. These lesions are called burned-out tumours (BOT).3e5 The discovery of non-palpable testicular lesions is becoming more common, as testicular ultrasonography (US) indications have become broader; becoming mandatory, for instance, in the male infertility workup. Historically, testicular focal lesions have been treated by radical orchiectomy, as testicular biopsies are not recommended.6 In this population of young men, preservation of fertility is an important consideration. Furthermore, whereas the most common type of palpable testicular lesions in young men are malignant carcinomas, >80% of non-palpable testicular lesions discovered at US may in fact be benign.7 Technical advances in US have allowed for better characterisation of testicular lesions, and management of nonpalpable testicular lesions has hence changed drastically. Nowadays, testis-sparing surgery with frozen section analysis is deemed to be appropriate, provided that the lesion measures <25 mm across and that it is not in contact with the rete testis, in order to preserve testicular vascularisation.8,9 Moreover, the European Society of Urogenital Radiology (ESUR) scrotal imaging subcommittee recommends active monitoring of small testicular lesions (i.e., <5 mm across) in young/infertile patients, provided they have benign US patterns and negative tumour markers.7 Unenhanced B-mode US is the reference standard for the characterisation of testicular lesions, particularly to describe their morphological features; while colour Doppler is the preferred mode for visualising their macrovascularisation.10 Scrotal MRI and shear-wave elastography are promising imaging techniques, although they are still undergoing evaluation.7,11e14 Over the last decade, contrastenhanced US (CEUS) has proven to be suitable to analyse microvascularisation in various organs. Indeed, CEUS provides qualitative results as well as quantitative data with timeeintensity curves extracted from the image using specific software.15,16 There have only been a few studies, however, demonstrating the suitability of CEUS for the
characterisation of focal testicular tumours. Thus the aim of the present study was hence to assess the diagnostic performance of conventional US and CEUS to investigate nonpalpable testicular tumours, using qualitative evaluation and quantitative parameters.
Materials and methods Population and study design de The local ethics committee approved the study (Comite protection des personnes, Ile de France VII, CPP 13 051), and all of the patients provided written informed consent. This monocentric study was carried out between December 2011 and February 2014. Of the 3,256 inpatients and outpatients referred for scrotal US during this period, 49 patients who were diagnosed with a non-palpable testicular focal tumour discovered during US and referred for surgery, were enrolled consecutively and prospectively underwent CEUS. Patients with non-tumorous lesions, such as haematomas, infarctions, adrenal rests, and focal orchitis, were not part of this study: those diagnoses were performed based on clinical and biological data and follow-up. After the CEUS, nine patients were secondarily excluded in light of movement artefacts (due either to the probe or the patient) that compromised the data (n¼2), the small size of the lesion precluding being able to define an appropriate region of interest (ROI; n¼5), or when the lesion was too close to the hilum due to overlap with the vessels’ convergence (n¼2). Thus, ultimately a total of 40 patients were included (Fig 1). The final diagnosis was based on histological criteria (radical orchiectomy or tumorectomy). Three groups were identified: benign tumours, malignant tumours, and burned-out tumours (BOT). Five patients with BOT had been included in a case series of 10 patients with BOT that was revealed during investigations for infertility.5 Qualitative and quantitative CEUS evaluations were not reported in this published study.
Conventional US and CEUS The testicular US was performed by two radiologists with expertise in urogenital US and qualified in the use of CEUS. The radiologists reviewed the standardised reports, images, and videos in the picture archiving and communication system (PACS). An Aplio 500 US device (Toshiba Medical Systems Co., Ltd, Nasu, Japan), with two superficial probes
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
322.e3
Figure 1 Flow chart. MT, malignant tumours; BT, benign tumours; BOT, burned-out tumours.
(a 5-12 MHz linear array and an 8-18 MHZ linear array) was used. Computer-assisted calculation of the testicular volume was based on use of the following formula: volume ¼ length width depth 0.523. Using B-mode, the testicular volume; the tumour characteristics including size, echogenicity (e.g., hypoechoic or not), and margins (e.g., irregular or not); the features of the adjacent parenchyma (e.g., homogeneous or heterogeneous); and the presence of microlithiasis. Vascularisation was studied using colour Doppler US. To adequately evaluate blood flow, the Doppler parameters were adjusted to the most sensitive settings that did not introduce artefacts using a 3.5 cm/s scale. The pattern of vascularisation was categorised as peripheral, central, or mixed (peripheral and central). CEUS was performed by a bolus injection of 4.8 ml of Sonovue (Bracco Imaging, France) into a superficial vein of the forearm. The intravenous tube was immediately flushed with isotonic saline solution. Image acquisition lasted at least 3 minutes or until the uptake disappeared completely, and it was performed at a low mechanical index (<0.5). Video clips were sent for perfusion software processing (VueBox 4.3, Bracco Suisse SA, Geneva, Switzerland). The first step was qualitative analysis: for each lesion, the uptake of contrast was assessed and noted as being faster, similar, or delayed compared to the adjacent parenchyma. Video clips were then analysed using the perfusion software to determine quantitative perfusion patterns (VueBox 4.3, Bracco Suisse SA, Geneva, Switzerland). Using a ROIs placed (in consensus) in the lesion and in the adjacent parenchyma, the software extracted timeeintensity curves and 11 quantitative perfusion parameters15: (1) peak enhancement (a.u.); (2) mean time transit: full-width at
half-maximum of the curve (s); (3) area under the curve during wash-in (a.u.); (4) rise time: the time it took for the contrast signal to increase from the baseline level to peak enhancement (s); (5) time to peak: the time needed to reach peak enhancement (s); (6) wash-in rate ¼ maximal slope, calculated as the peak enhancement / rise time (a.u.); (7) perfusion index, calculated as the area under the curve during wash-in divided by the rise time (a.u); (8) fallout time: the time needed for the descending slope to reach a contrast signal intensity of zero (s); (9) wash-out rate ¼ minimal slope (a.u.); (10) area under the curve during wash-out (a.u.); (11) area under the curve during wash-in and wash-out (a.u)
Statistical analysis The three groups, i.e., malignant tumours, benign tumours, and BOT, were compared using Fisher’s exact test for non-continuous variables (unenhanced US criteria and qualitative CEUS), and with the KruskaleWallis test for continuous variables (testicular volume, lesion size, and CEUS criteria).
Results Clinical data The clinical data are presented in Table 1. Forty patients who were 20e58 years of age (median age of 35.5) were included. Twenty-three patients (57.5%) were referred for hypofertility, seven (17.5%) due to a history of contralateral scrotal lesion, and two (5%) had a history of retroperitoneal germ cell tumours. Five patients (12.5%) consulted for scrotal pain, one (2.5%) for haemospermia, one (2.5%) for gynaecomastia, and one (2.5%) prior to gender
322.e4
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
Table 1 Clinical characteristics.
No. of patients Age (years) Size (mm) Patient with metastatic statusa Radical orchiectomya
Malignant tumour
Seminomas
Malignant Sertoli cell tumours
Benign tumours
Benign Leydig cell tumours
Benign Sertoli cell tumours
BOT
15 39 (29; 53) 11 (6; 29) 0 (0%) 14 (93%)
14 37 (29; 53) 10.5 (6; 29) 0 (0%) 13 (93%)
1 42 20 0 (0%) 1 (100%)
16 32.5 (20; 47) 5.8(2; 13) 0 (0%) 2 (12.5%)
15 32.5 (20; 47) 6(2; 13) 0 (0%) 1 (6.7%)
1 29 7.5 0 (0%) 1 (100%)
9 40 (20; 58) 16(10; 22) 3 (33%) 9 (100%)
Unless indicated otherwise, data are medians with ranges in parentheses. a The values represent the number of patients with the percentage in parentheses.
reassignment. Five patients (12.5%) had a single testis. Sixteen patients (40%) had a benign tumour: 15 Leydig cell tumours and one Sertoli cell tumour. Fifteen patients (37.5%) had a malignant tumour: 14 seminomas and one patient had a malignant Sertoli cell tumour. None of these patients had metastatic retroperitoneal lymph nodes. Nine patients (22.5%) had a BOT. Three of these had metastatic retroperitoneal lymph nodes (two with metastatic seminomatous tumours, and one patient had a metastatic nonseminomatous tumour (a teratoma). Based on clinical, biological, US, and magnetic resonance imaging (MRI) findings, patients with a lesion suspected of being malignant or BOT underwent radical orchiectomy (14/ 15 malignant tumours and 9/9 BOT), except for one patient with a single testis who was managed by testis-sparing surgery. Lesions suspected of being benign were managed by tumorectomy (n¼14/15).
Unenhanced US findings The US features are presented in Table 2. The morphological features of malignant tumours and BOT differed significantly from those of benign tumours. The benign tumours (median size: 5.8 mm; min: 2 mm; max: 13 mm) were significantly smaller than both the BOT (median size: 16 mm; min: 10 mm; max: 22 mm; p¼0.0001) and the malignant tumours (median size: 11 mm; min: 6 mm; max: 29 mm; p¼0.0002). No significant difference was noted in terms of the testicular volume. All of the lesions were
hypoechoic. Irregular margins were significantly more frequent with both malignant tumours (n¼11, p<0.0001) and BOT (n¼9, p<0.0001) than with benign tumours (n¼1). The adjacent parenchyma were more often heterogeneous with both malignant tumours (n¼11, p<0.0001) and BOT (n¼9, p<0.0001) than with benign tumours (n¼0). Microlithiasis was significantly more frequent with malignant tumours (n¼13) than with benign tumours (n¼2; p<0.001). Based on colour Doppler imaging, peripheral vascularisation was more frequent with benign tumours (n¼15) than with both malignant tumours (n¼3, p<0.0001) and burnedout tumours (n¼0, p<0.0001). Based on colour Doppler imaging, none of the BOT were vascularised. Examples of this are shown in Fig 2.
CEUS findings The qualitative findings are presented in Table 3. No significant differences were observed between benign and malignant tumours. Faster wash-in was observed in 11 benign and 11 malignant tumours (p¼1.000). Five benign tumours and two malignant tumours had a similar level of enhancement (p¼0.394). Delayed enhancement was observed in two malignant tumours (p¼0.226), but not in benign tumours. Conversely, delayed enhancement was significantly more frequent in BOT (n¼6) than in either malignant tumours (n¼2; p¼0.021) or benign tumours (n¼0; p<0.0001). The quantitative findings are presented in Table 4. Both benign and malignant tumours enhanced
Table 2 Characteristics of lesions (in B-mode US and colour Doppler). Groups
p-Value p-Value
Benign (n¼16) Malignant (n¼15) Burned-out (n¼9) Lesion size (mm) Testicular volume (ml) Hypoechogenicityc Irregular marginsc (%) Mixed vascularizationc Peripheral vascularizationc Heterogeneity of the adjacent parenchymac Microlithiasis
Benign vs. malignant Malignant vs. BOT Benign vs. BOT
5.8(2; 13) 10.4 (6;22.6) 16(100%) 1(6%) 0 15(94%) 0
11 (6;29) 10.8 (5.6;18.2) 15(100%) 11(73%) 12(80%) 3(20%) 11(73%)
16(10; 22) 9 (6.1;11) 9(100%) 9(100%) 0 0 9(100%)
0.0001a 0.3877a 0.000b 0.000b 0.000b 0.000b
0.0002a 0.9056a 0.008b 0.001b 0.008b 0.001b
0.0001a 0.2949a 0.259b 0.000b 0.266b 0.253b
0.2831a 0.1524a 0.000b 0.000b 0.000b
2(12%)
13(87%)
5(56%
0.000b
0.002b
0.150b
0.058b
Unless indicated otherwise, data are medians with ranges in parentheses. a p-Value after comparison of the three groups with the non-parametric KruskaleWallis test. b p-Value after comparison of the three groups with Fisher’s exact test. c The values represent the number of patients, percentages in parentheses.
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
322.e5
Figure 2 Unenhanced contrast US: B-mode and Doppler. (a) US imaging of a benign testicular tumour: B-mode reveals a homogeneous hypoechoic lesion, with regular margins, and homogeneous adjacent parenchyma without microlithiasis. Vascularization is mostly peripheral on colour-coded Doppler. (b) Malignant testicular tumour: B-mode reveals a heterogeneous hypoechoic lesion with irregular margins and microlithiasis in the adjacent parenchyma. Colour-Doppler shows that vascularization is mixed, both central and peripheral. (c) Burned-out tumour: B-mode reveals a heterogeneous hypoechoic lesion with irregular margins and heterogeneous adjacent parenchyma. No vascularization is seen using power Doppler.
strongly, with a pronounced wash-in immediately followed by a wash-out, as shown by the timeeintensity curves (Figs 3 and 4); however, the quantitative parameters were not significantly different between benign and malignant tumours. BOT exhibited a weaker level of enhancement
than both benign and malignant lesions, with an almost flat timeeintensity curve (Fig 5). Nine parameters were significantly different between BOT, benign, and malignant tumours (Table 4). The peak enhancement, wash-in and wash-out rates, perfusion index,
Table 3 Qualitative analysis. Groups
Rapid wash-in Similar wash-in Delayed wash-in
Benign (n¼16)
Malignant (n¼15)
Burned-out (n¼9)
11(69%) 5(31%) 0
11(73%) 2(13%) 2(13%)
0 3(33.3%) 6(67%)
Data represent number of lesions with percentage in parentheses. a p-Value after comparison with Fisher’s exact test.
p-Valuea
p-Valuea Benign vs. Malignant
Malignant vs. BOT
Benign vs. BOT
0.001 0.565 0.000
1.000 0.394 0.226
0.001 0.615 0.021
0.001 1 0.000
322.e6
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
Table 4 Quantitative parameters. Benign (n¼16)
Peak enhancement (a.u.) Wash-in rate (a.u.) Area under the curve during wash-in Rise time (s) Time to peak (s) Mean time transit (s) Perfusion index (a.u.) Area under the curve (a.u.) Fall time (s) Wash-out rate (a.u.) Area under the curve during wash-out (a.u.)
Malignant (n¼15)
Burned-out (n¼9)
p-Valuea
p-Valuea Benign vs. malignant
Benign vs. BOT
Malignant vs. BOT
2422 (412;12,452) 324 (5;,3673) 16191 (2,549; 68,597)
2284(81,26,018) 252 (10;3,881) 12591 (514; 16,0803)
166 (2;554) 18 (0;120) 2580 (19; 2459)
0.0004 0.0006 0.0021
0.707 0.7369 0.953
0.0001 0.0001 0.0007
0.026 0.0026 0.0038
10.2 (4.7; 19.0) 15.5 (7.7;37.6) 60.1 (22.8;110.5) 1544 (273; 7,858) 48746 (7,476;18,5521) 16.8 (9.0; 34.8) 143 (16;1,547) 31380 (4,927; 11,6924)
11.2 (5.6; 22.8) 18.9 (8.3;32.1) 49.3 (26.6;419.8) 1419 (50; 16,597) 41850 (1,398;47,7912) 26.3 (10.2; 62.5) 118 (5;1,444) 29529 (884; 317,109)
14.5 (8.1; 55.1) 26.1 (9.;63.2) 106.3 (36.4;523.4) 116 (1; 378) 1070 (0; 11,926) 25.3 (0; 34.8) 5 (0;22) 671 (0; 8,842)
0.0795 0.3022 0.0373 0.0006 0.0003 0.1927 0.0011 0.0003
0.2280 0.6781 0.6212 0.7369 0.8588 0.0721 0.5401 0.7669
0.0315 0.1570 0.0174 0.0001 0.0002 0.5006 0.0001 0.0003
0.1998 0.1998 0.0295 0.0026 0.0004 0.4014 0.0067 0.0004
Data are medians with ranges in parentheses. a p-Value after comparison with the non-parametric KruskaleWallis test.
Figure 3 Peak enhancement in a Leydig cell tumour. Note the difference between the yellow curve (the lesion), and the pink curve (adjacent parenchyma).
and areas under the curve were significantly lower in BOT than in either benign tumours or malignant tumours (the pvalues ranged from 0.0001 to 0.0067). The mean time transit was significantly higher in BOT than in either benign tumours (p¼0.0174) or malignant tumours (p¼0.0295). The rise time was significantly higher in BOT than in benign tumours (p¼0.0315), while there was no difference between BOT or malignant tumours (p¼0.1998).
Discussion The findings of the present study show that unenhanced US can be used to identify features that enable benign lesions to be differentiated from malignant and BOT. Thus, irregular margins and heterogeneity of the adjacent
parenchyma were significantly more frequent in malignant and BOT than in benign tumours, whereas benign lesions were well-circumscribed with homogeneous adjacent pulp. With colour Doppler imaging, peripheral vascularisation was significantly associated with benign tumours, whereas mixed vascularisation was significantly more frequent in malignant lesions. By contrast, colour Doppler US did not reveal vascularisation of BOT. CEUS is an imaging technique that involves microbubble contrast agents that remain strictly intravascular; therefore, it is a good technique for studying microvascularisation. The present study has revealed significant differences in the enhancement of BOT compared to benign and malignant testicular tumours. Both malignant and benign testicular tumours exhibited strong enhancement, characterised by substantial wash-in and wash-out on quantitative and
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
322.e7
Figure 4 Peak enhancement in a seminoma. Note the strong enhancement of the seminoma (yellow curve) compared to the adjacent parenchyma (pink curve).
Figure 5 CEUS in a burned-out tumour: The lesion (yellow curve) does not enhance compared to the adjacent parenchyma (pink curve).
qualitative CEUS. Conversely, enhancement of BOTs was poor, as reflected by flat timeeintensity curves and low perfusion parameters. BOT are primary germ cell tumours of the testes that are replaced by fibrosis after having undergone partial or total regression. Most of them are diagnosed belatedly, being detected by the discovery of metastases such as retroperitoneal lymphadenopathies.17e19 The regression of malignant
cells and their replacement by fibrosis results in poorly vascularised areas that are reflected on CEUS as an absence of contrast medium uptake after injection of microbubbles, a flat timeeintensity curve, and low perfusion parameters. Non-tumorous lesions, such as haematomas or infarctions, are also known to exhibit a lack of perfusion on CEUS.20,21 They are, however, associated with sudden pain and a change in the US pattern during follow-up.
322.e8
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
BOT are usually diagnosed at B-mode US when a large testicular lesion with irregular margins and heterogeneous adjacent parenchyma with microlithiasis is seen.3,17e19,22 Although the differential diagnosis of malignant tumours can be difficult with B-mode US, colour Doppler US is generally sufficient to achieve this, as it can reveal the hypovascularisation in BOT that does not occur in most malignant tumours. This study showed that in some difficult cases CEUS provided imaging data that enabled BOT to be differentiated from active benign and malignant tumours. Unlike BOT, both benign and malignant tumours tended to enhance strongly, with a pronounced wash-in followed by a wash-out, which is in accordance with the findings of Lock et al. and Huang and Sidhu.23,24 Contrary to previous results, however, in the present study none of the qualitative and quantitative parameters were statistically different between benign and malignant tumours. Isidori et al. found that malignant tumours exhibited a rapid wash-out, whereas delayed wash-out was observed in benign tumours.20 Rapid wash-out in malignant tumours could be explained by a more anarchic vascularisation and the presence of many shunts. In the study by Drudi et al., the wash-in of Leydig cell tumours was higher and more rapid than seminomas, as demonstrated by a shorter time to peak, higher peak enhancement, and wash-in rate in Leydig cell tumours. This difference is thought to be due to a high blood flow velocity and an increased microvessel density compared to seminomas, despite the overlapping values.25 In the present study, neither the wash-in nor the wash-out were different between benign (mostly Leydig cell tumours) and malignant tumours (mostly seminomas). These contradictory results reveal that microvascularisation of testicular tumours is difficult to predict by CEUS, and this is presumably due to the histological heterogeneity of malignant lesions. Benign tumours, which primarily comprise Leydig cell tumours, are generally homogeneous lesions with no necrotic, haemorrhagic, or cystic component.10 These tumours present a rich vascularisation which explains the substantial enhancement seen with microbubbles. A pronounced uptake of contrast with a quick wash-out was observed in all of such lesions in our study. On the other hand, the range of malignant tumours is more variable. Non-seminomatous tumours are known to be heterogeneous, with necrosis, cystic areas, and calcifications; whereas seminomas are generally more homogeneous.26,27 These different histological features are reflected as different enhancement behaviours with CEUS. In the present study, 97.7% of the malignant tumours were seminomas (n¼15/16), which enhanced strongly. The results may well have been different with a higher proportion of non-seminomatous tumours. The age of the present patients (median of 39 years of age), and the inclusion of non-palpable lesions only can explain this low proportion of non-seminomatous tumours. Regardless of the histological type, some germ cell tumours can partially regress and undergo fibrous involution before turning into BOT. Germ cell tumours with no fibrous involution should enhance strongly, while those that are
partially fibrous should exhibit less enhancement. These reasons explain why malignant testicular tumours do not have a stereotyped CEUS enhancement. Unlike Vandaele et al., the homogeneity or heterogeneity of wash-in and wash-out was not analysed in the lesions in light of their small size.28 Indeed, this evaluation could have been biased, as malignant tumours are generally larger than benign tumours, which are sometimes only a few millimetres in size. The main clinical challenge in this context is the distinction between lesions that will be treated by total orchiectomy, i.e., malignant tumours and BOT, from those that do not require a radical surgical treatment, i.e., benign tumours and non-tumorous lesions. As reported in the literature, non-tumorous lesions, such as haematomas or infarctions, are known to be non-vascularised by CEUS, and are thus readily recognisable if a degree of doubt persists after unenhanced US.21,23,24 According to the present study and the published literature, unenhanced US provides information that allows non-palpable benign tumours to be distinguished from malignant tumours and BOT.10 For small lesions (<5 mm) with benign US patterns and negative tumour markers in sub-fertile/young patients, the ESUR hence recommends an active monitoring instead of a surgery.7 Moreover, testissparing surgery may be appropriate for lesions measuring <25 mm in sub-fertile men.8,9 On standard US, distinguishing between malignant and BOT can be difficult, but their management is the same. CEUS adds little value in that context. CEUS could be used to distinguish benign tumours and BOT if there is a degree of doubt after unenhanced US, as their management is different, although this is rarely the case in common practice. The clinical benefit of CEUS in the investigation of nonpalpable testicular tumours is therefore poor. The relatively small number of patients in the present study limits interpretation of the present findings, as some results could have been statistically significant had more patients been included. Many lesions had to be excluded. Small lesions (i.e. with a mean size of 3 mm) were not included because they could not be accurately discerned using the CEUS mode. The placement of the ROI is difficult in small lesions, as also mentioned in the study by Isidori et al.20 Moreover, any movement of the probe or the patient makes the quantitative analysis uninterpretable. These limitations are more related to the technique than to the study itself, yet they are significant as many incidentally diagnosed testicular tumours are very small in size. Further, some lesions had to be excluded due to their localisation. Indeed, CEUS of lesions close to the hilum was affected by artefacts caused by the surrounding converging vessels. In the present study, the prevalence of benign stromal tumours (mostly Leydig cell tumours) and BOT over germ cell tumours was much higher than in the general population (germ cell tumours account for about 90e95% of all testicular tumours, whereas stromal tumours account for only about 5e10%). This difference could induce a selection bias. This difference may be due to the fact that the authors’ institution is a referral centre for testicular imaging.
A. Luzurier et al. / Clinical Radiology 73 (2018) 322.e1e322.e9
In conclusion, this study assessed the potential of CEUS for distinguishing between non-palpable BOT and active tumours. Colour Doppler US may, however, be an adequate way to differentiate between these two kinds of tumours. Unlike unenhanced US, CEUS did not provide any imaging data that allows non-palpable benign and malignant vascularised tumours to be differentiated. Lastly, the use of CEUS on small tumours is subject to technical limitations, which could restrict its widespread use.
References 1. Manecksha RP, Fitzpatrick JM. Epidemiology of testicular cancer. BJU Int 2009;104:1329e33. 2. Grem JL, Robins HI, Wilson KS, et al. Metastatic Leydig cell tumour of the testis. Report of three cases and review of the literature. Cancer 1986;58:2116e9. 3. Fabre E, Jira H, Izard V, et al. “Burned-out” primary testicular cancer. BJU Int 2004;94:74e8. 4. Casella R, Rochlitz C, Sauter G, et al. “Burned out” testicular tumour: a rare form of germ cell neoplasias. Schweiz Med Wochenschr 1999;129:235e40. 5. Rocher L, Glas L, Bellin MF, et al. Burned-out testis tumours in asymptomatic infertile men: multiparametric sonography and MRI findings. J Ultrasound Med 2016, https://doi.org/10.7863/ultra.15.08037. s C, et al. CCAFU recommendations 2013: 6. Durand X, Rigaud J, Avance testicular germ cell cancer. Prog Urol 2013;23:145e60. 7. Rocher L, Ramchandani P, Belfield J, et al. Incidentally detected nonpalpable testicular tumours in adults at scrotal ultrasound: impact of radiological findings on management radiologic review and recommendations of the ESUR scrotal imaging subcommittee. Eur Radiol 2016;26:2268e78. 8. Palermo G, Antonucci M, Recupero SM, et al. Focal surgery in testis cancer: current state of the art. Urologia 2013;80:290e6. 9. Brunocilla E, Gentile G, Schiavina R, et al. Testis-sparing surgery for the conservative management of small testicular masses: an update. Anticancer Res 2013 Nov;33:5205e10. 10. Maxwell F, Izard V, Ferlicot S, et al. Colour Doppler and ultrasound characteristics of testicular Leydig cell tumours. Br J Radiol 2016;89:20160089. 11. Sanharawi IE, Correas J-M, Glas L, et al. Non-palpable incidentally found testicular tumours: differentiation between benign, malignant, and burned-out tumours using dynamic contrast-enhanced MRI. Eur J Radiol 2016;85:2072e82. 12. Goddi A, Sacchi A, Magistretti G, et al. Real-time tissue elastography for testicular lesion assessment. Eur Radiol 2012;22:721e30. 13. Aigner F, De Zordo T, Pallwein-Prettner L, et al. Real-time sonoelastography for the evaluation of testicular lesions. Radiology 2012;263:584e9.
322.e9
14. Zeng B, Chen F, Qiu S, et al. Application of quasistatic ultrasound elastography for examination of scrotal lesions. J Ultrasound Med 2016;35:253e61. 15. Dietrich CF, Averkiou MA, Correas J-M, et al. An EFSUMB introduction into dynamic contrast-enhanced ultrasound (DCE-US) for quantification of tumour perfusion. Ultraschall Med 2012;33:344e51. 16. Piscaglia F, Nolsøe C, Dietrich CF, et al. The EFSUMB guidelines and recommendations on the clinical practice of contrast enhanced ultrasound (CEUS): update 2011 on non-hepatic applications. Ultraschall Med 2012;33:33e59. 17. Tasu J-P, Faye N, Eschwege P, et al. Imaging of burned-out testis tumour: five new cases and review of the literature. J Ultrasound Med 2003;22:515e21. 18. Gurioli A, Oderda M, Vigna D, et al. Two cases of retroperitoneal metastasis from a completely regressed burned-out testicular cancer. Urologia 2013;80:74e9. 19. Qureshi JM, Feldman M, Wood H. Metastatic “burned-out” germ cell tumour of the testis. J Urol 2014;192:936e7. 20. Isidori AM, Pozza C, Gianfrilli D, et al. Differential diagnosis of nonpalpable testicular lesions: qualitative and quantitative contrastenhanced US of benign and malignant testicular tumours. Radiology 2014;273:606e18. €der C, Lock G, Schmidt C, et al. Real-time elastography and 21. Schro contrast-enhanced ultrasonography in the evaluation of testicular masses: a comparative prospective study. Ultrasound Med Biol 2016;42:1807e15. 22. Balalaa N, Selman M, Hassen W. Burned-out testicular tumour: a case report. Case Rep Oncol 2011;4:12e5. €der C, Schmidt C, et al. Contrast-enhanced ultrasound and 23. Lock G, Schro real-time elastography for the diagnosis of benign Leydig cell tumours of the testis d a single center report on 13 cases. Ultraschall Med 2014;35:534e9. 24. Huang DY, Sidhu PS. Focal testicular lesions: colour Doppler ultrasound, contrast-enhanced ultrasound and tissue elastography as adjuvants to the diagnosis. Br J Radiol 2012;85:41e53. 25. Drudi FM, Valentino M, Bertolotto M, et al. CEUS timeeintensity curves in the differentiation between Leydig cell carcinoma and seminoma: a multicenter study. Ultraschall Med 2015, https://doi.org/10.1055/s-00341398841. 26. Woodward PJ, Sohaey R, O’Donoghue MJ, et al. From the archives of the AFIP: tumours and tumour like lesions of the testis: radiologicepathologic correlation. RadioGraphics 2002;22:189e216. 27. Coursey Moreno C, Small WC, Camacho JC, et al. Testicular tumours: what radiologists need to knowddifferential diagnosis, staging, and management. RadioGraphics 2015;35:400e15. 28. Vandaele P, Marcelli F, Ouzzane A, et al. Testicular nodules of infertile men and contrast enhanced ultrasonography: preliminary study. Prog Urol 2015;25:274e81.