MRI in the Characterization and Local Staging of Testicular Neoplasms

Genitourinary Imaging Original Research Tsili et al. MRI of Testicular Neoplasms Genitourinary Imaging Original Research thina. Tsili 1 Maria I. rgyropoulou 1 Dimitrios Giannakis 2 Nikolaos Sofikitis 2 Konstantine Tsampoulas 1 Tsili, rgyropoulou MI, Giannakis D, Sofikitis N, Tsampoulas K Keywords: magnetic resonance, MR imaging, staging, testicular neoplasms, testis DOI:10.2214/JR.09.3256 Received June 20, 2009; accepted after revision ugust 25, 2009. Presented at the 2005 annual meeting of the merican Roentgen Ray Society, New Orleans, L. 1 Department of linical Radiology, University Hospital of Ioannina, Leoforos S. Niarchou, 45500, Platia Pargis, 2, 45332, Ioannina, Greece. ddress correspondence to.. Tsili (a_tsili@yahoo.gr). 2 Department of Urology, University Hospital of Ioannina, Ioannina, Greece. ME This article is available for ME credit. See www.arrs.org for more information. JR 2010; 194:682 689 0361 803X/10/1943 682 merican Roentgen Ray Society MRI in the haracterization and Local Staging of Testicular Neoplasms OJETIVE. The purpose of this study was to assess the role of MRI in the preoperative characterization and local staging of testicular neoplasms. SUJETS ND METHODS. MRI was performed on 33 patients referred because a testicular mass had been detected clinically and sonographically. oth T1- and T2-weighted sequences were performed with a 1.5-T MRI unit. Gadolinium chelate was administered IV in all cases. We recorded the presence of a lesion and whether the histologic diagnosis of testicular malignancy could have been predicted on the basis of MRI features. For testicular neoplasms, local extension of disease was studied. The MRI findings were correlated with the surgical and histopathologic results. RESULTS. Histologic examination revealed 36 intratesticular lesions, 28 (78%) of which were malignant and eight benign. Thirteen malignant testicular tumors (46%) were confined within the testis, 12 (43%) had invaded the testicular tunicae or epididymis, and three (11%) had invaded the spermatic cord. The sensitivity and specificity of MRI in differentiating benign from malignant intratesticular lesions were 100% (95% I, 87.9 100%) and 87.5% (95% I, 52.9 97.7%). The rate of correspondence between MRI and histologic diagnosis in the local staging of testicular tumors was 92.8% (26/28). ONLUSION. MRI is a good diagnostic tool for the evaluation of testicular disease. It is highly accurate in the preoperative characterization and local staging of testicular neoplasms. I maging has an important role in the investigation of testicular masses. Sonography, although the primary imaging technique for the evaluation of scrotal contents, does not always allow confident characterization of the nature of a testicular mass [1 3]. MRI of the scrotum has been used as an alternative technique that performs well in the morphologic evaluation and tissue characterization of scrotal disease [4 16]. The advantages of MRI are simultaneous imaging of both testicles and both sides of the inguinal region, acquisition of adequate anatomic information, and satisfactory tissue contrast [4 16]. MRI may be a valuable problem-solving tool in the assessment of testicular disease when sonographic findings are equivocal or inconclusive [10, 11]. The technique also may be accurate for differentiation of intratesticular from extratesticular mass lesions and may yield additional anatomic and morphologic information about intratesticular masses [4 16]. The primary goal in the evaluation of a palpable scrotal mass is determining its lo- cation. Most solid intratesticular masses should be considered malignant, and radical orchiectomy is the treatment of choice. To avoid unnecessary orchiectomy, however, it is extremely important to recognize the imaging features of various benign intratesticular mass lesions, including orchitis, hemorrhage, ischemia and infarction, fibrosis, and dilatation of the rete testis [4 19]. lthough organ-sparing surgery is not generally indicated in the care of patients with testicular masses, it is suggested in the following circumstances: suspected intratesticular benign lesion, synchronous bilateral testicular neoplasms, metachronous contralateral testicular tumors with normal preoperative testosterone levels, and tumor in one testis in association with normal preoperative testosterone levels [20 22]. Preoperative imaging evaluation of the local stage of disease is mandatory in the care of patients for whom organ-sparing surgery is planned [20 22]. The purpose of our study was to assess the role of MRI in the preoperative characterization and local staging of testicular neoplasms. 682 JR:194, March 2010

MRI of Testicular Neoplasms Subjects and Methods Forty patients were prospectively enrolled in this study, which was approved by the institutional review board of our hospital. ll patients were referred to the urology department because of clinical and sonographic detection of a testicular mass. Written informed consent was obtained before the examinations. Five patients were excluded from the study because of normal MRI findings in four cases and the MRI finding of a varicocele in one case. One patient with the MRI finding of bilateral intratesticular lesions was excluded from the data analysis because he was lost to follow-up without histologic confirmation. nother patient with a testicular mass was excluded because he declined orchiectomy. In this case, the histologic diagnosis was obtained from the report of a biopsy of a left supraclavicular lymph node that showed an undifferentiated metastatic neoplasm probably of germ cell origin. Thirty-three patients (mean age, 35 years; range, 17 70 years) with surgical pathologic confirmation (radical orchiectomy, 29 patients; testicular biopsy, three patients; radical orchiectomy and biopsy of the contralateral testis, one patient) were the cohort in this study. MRI Technique ll the MRI examinations were performed with a 1.5-T system (Gyroscan and Intera, Philips Healthcare) with a 17-cm circular surface coil (27 patients) or a pelvic phased-array coil (six patients). Patients were examined in the supine position. towel was placed under the scrotum to maintain a comparable distance of both testicles from the coil. The penis was kept on the anterior abdominal wall in most cases. xial, coronal, and sagittal images were obtained with unenhanced and contrast-enhanced T1-weighted spinecho sequences (TR/TE, 500 650/13 15) and a T2-weighted fast spin-echo sequence (4,000/100 120) with 3- to 4-mm slice thickness and a 0.5- mm gap. The image matrix was 180 256 mm, and the field of view was 240 270 mm. xial fatsuppressed T1-weighted sequences were repeated if a lesion with high T1 signal intensity was found. In all cases, 0.2 mmol/kg gadodiamide (Omniscan, GE Healthcare) was administered IV. MR Image Interpretation Image analysis included prospective reading of mass lesion characteristics. MR images were studied by two radiologists who did not know the surgical and histopathologic results, and discrepancies were resolved by consensus. We recorded the presence or absence of an intratesticular lesion and the possible diagnosis based on the MRI features with the surgical and pathologic results as the standard of reference. Tumor size, margins, signal homogeneity or heterogeneity, and patterns of contrast enhancement were evaluated. Lesion signal intensity was characterized as isointense, hyperintense, or hypointense compared with the normal contralateral testis. The MRI criteria used to characterize malignant testicular tumors were based on previously published reports [12, 13]. Specifically, the presence of a multinodular intratesticular lesion of mainly low signal intensity on T2-weighted images (Fig. 1) or an inhomogeneous mass with variable signal intensity on T2-weighted images (Fig. 2) was considered suggestive of malignancy [12, 13]. ontrast enhancement, especially if heterogeneous, also was considered indicative of malignancy. The coexistence of areas of hemorrhage, detected with high signal intensity on both T1-weighted and fatsuppressed T1-weighted images was considered suggestive of a malignant diagnosis. The presence of necrosis, detected as hyperintense areas on T2- weighted images not becoming enhanced after contrast administration, and extension of the neoplasm to the testicular tunicae, epididymis (Fig. 2), or spermatic cord confirmed the diagnosis of malignancy. The sensitivity and specificity with exact 95% I for each and the overall accuracy of MRI in differentiating benign from malignant intratesticular lesions were calculated. In patients with malignant testicular tumors, the accuracy of MRI in the assessment of local stage of disease was evaluated. The local extent of testicular neoplasms was classified according to the recommendations of the National omprehensive ancer Network and the European Germ ell ancer onsensus Group [3, 9, 23, 24] (Table 1). n intratesticular tumor surrounded by a rim of normal testicular parenchyma or intact testicular tunicae detected as a continuous hypointense halo in the periphery of the testicle (Fig. 1) was classified category T1 disease. In cases in which the neoplasm interrupted the testicular tunicae with or without a mass in the paratesticular space (Figs. 2 and 2), the local category of disease was categorized T2. Enlargement and contrast enhancement of the spermatic cord due to neoplastic involvement were categorized T3 (Fig. 3). Tumoral invasion of the scrotal wall was considered category T4. Results t surgery and histopathologic examination, 36 intratesticular lesions were found in 33 patients. Twenty-eight (78%) of the lesions were malignant, and eight (22%) were benign. Twenty-seven patients had intrates- Fig. 1 27-year-old man with seminoma of right testicle. Pathologic finding was category pt1., Transverse T1-weighted MR image shows multinodular right testicular lesion is isointense (arrow) compared with normal contralateral testis (star)., Sagittal T2-weighted MR image shows hypointense bands (small arrows) within mass that have greater enhancement than tumoral tissue and correspond to pathologic finding of fibrous septa. Part of tumor is close to testicular tunica, which was found intact (large arrow) at histopathologic examination., Sagittal contrast-enhanced T1-weighted MR image shows areas that did not become enhanced (black and white arrows) and correspond to areas of necrosis found at histologic examination. JR:194, March 2010 683

Tsili et al. ticular malignant tumors, three patients had benign testicular lesions, two patients had benign intratesticular masses in both testicles, and one patient had both a malignant and a benign testicular lesion. The malignant testicular tumors were 13 seminomas, 13 nonseminomatous germ cell tumors (three cases of embryonal carcinoma and teratoma; one case of embryonal carcinoma and seminoma; two cases of embryonal carcinoma; two cases of mature cystic teratoma; one case of embryonal carcinoma, teratoma, and yolk sac tumor; one case of yolk sac tumor alone; two cases of embryonal carcinoma, teratoma, choriocarcinoma, and yolk sac tumor; and one case of embryonal carcinoma, teratoma, yolk sac tumor, and seminoma), and two primary testicular diffuse large -cell lymphomas. The mean maximal diameter of the testicular neoplasms was 5.1 cm (range, 1.2 12 cm). Regarding local extension of testicular tumors, the pathologic reports showed 13 neoplasms (46%) were confined within the testis (category T1), including eight nonseminomatous germ cell neoplasms, four seminomas, and Fig. 2 45-year-old man with nonseminomatous tumor (teratocarcinoma, yolk sac tumor, and seminoma) of right testicle. Pathologic category was pt2. and, Transverse () and sagittal () T2-weighted images show large extremely heterogeneous testicular mass replacing right testicle. Lesion has peripheral dark halo (arrow, ) that correlated with histologic finding of fibrous capsule. Tumor interrupts testicular tunicae and invades epididymis (arrow, ), as confirmed at pathologic examination., Sagittal contrast-enhanced T1-weighted image shows inhomogeneous enhancement of mass. reas (arrow) that did not become enhanced after gadolinium administration correlated with histologic finding of areas of necrosis. TLE 1: lassification of Testicular Neoplasms Stage of Primary Tumor pt0 pt1 pt2 pt3 pt4 No evidence of primary tumor Extent of Primary Tumor ssessed fter Radical Orchiectomy Tumor confined to the testis, no vascular or lymphatic invasion, possible invasion of the tunica albuginea but not the tunica vaginalis Tumor confined to the testis with vascular or lymphatic invasion or extending through the tunica albuginea with involvement of the tunica vaginalis Tumor invading the spermatic cord with or without vascular or lymphatic invasion Tumor invading the scrotal wall with or without vascular or lymphatic invasion one lymphoma. mong those malignant tumors, histologic examination showed five were close to the testicular tunicae without signs of neoplastic infiltration. In 12 cases (43%), invasion of the testicular tunicae or the epididymis (three cases) by the neoplasm was proved histologically (category T2). These 12 tumors included five nonseminomatous neoplasms, six seminomatous tumors, and one lymphoma. Finally, pathologic examination showed that three seminomatous testicular tumors (11%) had invaded the spermatic cord (category T3). The pathologic diagnoses of benign tumor in two patients were bilateral tubular ectasia of the rete testis coexisting with a large spermatocele. Histologic examination revealed two diagnoses of hemorrhagic necrosis, involving both the testis and the epididymis in two patients. The other diagnoses were one case of fibrosis in a patient with a seminoma in the contralateral testis and one case of granulomatous orchitis. MRI findings led to correct preoperative characterization of all intratesticular malignant tumors and seven of eight cases of benign testicular lesions in this study. The sensitivity and specificity of the technique in differentiating benign from malignant intratesticular lesions were 100% (95% I, 87.9 100%) and 87.5% (95% I, 52.9 97.7%). The diagnostic performance of MRI in the preoperative characterization of malignant testicular tumors is summarized in Table 2. The mean diameter of testicular neoplasms detected at MRI was 5.2 cm (range, 1.1 12), which was in accordance with the histopathologic results. On T1-weighted images, all malignant testicular tumors detected were mainly isointense compared with the normal contralateral testis (Figs. 1 and 3). In two patients with mature cystic testicular teratoma, fatty components within the tumors were seen with high and low signal intensity on T1-weighted and fat-suppressed T1-weighted images, respectively. reas of hemorrhage, detected as hyperintense foci on both T1-weighted and fat-suppressed T1- weighted images (Fig. 3) were found histologically in 10 intratesticular tumors, including four seminomas, five nonseminomatous tumors, and one testicular lymphoma. The signal intensity of malignant testicular tumors was variable on T2-weighted images. Thirteen tumors were relatively homogeneous, having low signal intensity on T2-weighted images. These tumors included 11 seminomas (Figs. 1, 3, and 3), one embryonal carcinoma, and one testicular lymphoma (Fig. 4). In two patients, seminomatous tumors were inhomogeneous on T2-weighted images, although mainly hypointense. In all 13 cases of seminoma, hypointense bands were found 684 JR:194, March 2010

MRI of Testicular Neoplasms Fig. 3 40-year-old man with right testicular seminoma invading spermatic cord (category pt3)., Transverse T1-weighted image shows large right testicular tumor. rea of high signal intensity (arrow) within lesion corresponds to histopathologic finding of hemorrhage. and, oronal T2-weighted images depict tumor of mainly low signal intensity invading distal part of ipsilateral spermatic cord (arrow, ). TLE 2: Diagnostic Performance of MRI in Diagnosis of Malignant Testicular Tumors haracteristic Value (%) Sensitivity 100 Specificity 87.5 Positive predictive value 96.5 Negative predictive value 100 ccuracy 96.4 Fig. 4 52-year-old man with primary diffuse large -cell lymphoma of left testicle. Part of tumor was in proximity to testicular tunicae, which proved intact at pathologic examination (category pt1). and, oronal T2-weighted () and contrast-enhanced T1-weighted () images show relatively homogeneous left intratesticular tumor of low signal intensity on T2-weighted images. Testicular tunicae (long arrow, ) appear intact as hypointense halo in periphery of testicular parenchyma. Mass is strongly and inhomogeneously enhanced. Moderate left hydrocele (short arrow) also is evident. within the tumors on T2-weighted images, corresponding histologically with fibrous septa (Fig. 1). Thirteen testicular neoplasms were heterogeneous on T2-weighted images, detected as areas of low and areas of high signal intensity. The histologic diagnosis included 12 nonseminomatous tumors (Figs. 2 and 2) and one testicular lymphoma. hypointense halo in the periphery of malignant testicular tumors (Fig. 2) was visualized on T2- weighted images in nine cases in this study, all nonseminomatous germ cell tumors. t pathologic examination this halo was found to correspond to a fibrous capsule. fter gadolinium administration, all testicular neoplasms exhibited heterogeneous enhancement (Figs. 1, 2, and 4), which was more conspicuous in nonseminomatous neoplasms (Fig. 2). In all cases of seminoma, fibrous septa showed greater enhancement than the rest of the tumoral tissue (Fig. 1). reas within tumors of high signal intensity on T2-weighted images that did not become enhanced after contrast administration (Figs. 1 and 2) were found to be areas of necrosis at histologic examination. These findings were made in 12 neoplasms, including seven seminomas, four nonseminomatous tumors, and one lymphoma. One false-positive finding was made in this study. The patient had granulomatous orchitis diagnosed after testicular biopsy. The MRI examination showed a 0.6-cm hypointense intratesticular lesion that became enhanced after gadolinium administration and was falsely interpreted as malignant. With MRI, 87.5% (7/8) of benign intratesticular mass lesions in this study were characterized correctly. Two of the patients had bilateral dilatation of the rete testis detected as multicystic masses with signal intensity identical to that of water, that is, low on T1-weighted images and high on T2-weighted images. These lesions were located in the region of the mediastinum testis and did not become enhanced after contrast administration. The mean diameter was 2 cm. In both cases, a large unilocular cystic mass was found to coexist in the para testicular space in continuity with the larger intratesticular lesion. The histologic finding was spermatocele. The MRI findings were strongly suggestive of a benign lesion in a patient with testicular fibrosis. These findings included a sharply demarcated intratesticular mass (mean diameter, 1.9 cm) with low signal intensity on T1-weighted images and very low signal intensity on T2- weighted images that did not become enhanced after gadolinium administration (Fig. 5). In the same patient, a large testicular seminoma JR:194, March 2010 685

Tsili et al. invading the epididymis and the spermatic cord was seen in the contralateral hemiscrotum. On T1-weighted images, a hyperintense testis exhibiting lack of enhancement was detected in both patients with hemorrhagic testicular necrosis (Fig. 6). The testis was of very low signal intensity on T2-weighted images of one patient (Fig. 6) and heterogeneous with a spotty pattern of very low signal intensity on T2-weighted images of the other patient. In both patients, the epididymis was hypointense on T2-weighted images (Fig. 6). The MRI criteria used to characterize malignant versus benign intratesticular lesions in this study are summarized in Table 3. With MRI, 12 of 13 category T1 testicular tumors (92%) were correctly staged. In 11 cases, neoplasms detected in proximity to Fig. 5 47-year-old man with left testicular seminoma invading ipsilateral spermatic cord (not shown) and right testicular fibrosis proven at histologic examination after biopsy of right testis. and, Transverse () and coronal () T2-weighted MR images show right upper pole intratesticular lesion (arrow) with very low signal intensity, suggesting presence of fibrous tissue., Transverse contrast-enhanced T1-weighted image shows no lesion enhancement (arrow), confirming benign diagnosis. the testicular tunicae were assessed as intact at imaging (Figs. 1 and 4), and this finding was proved at pathologic examination. In one case, the tumor was falsely characterized as T2 owing to inability at MRI to recognize intact testicular tunicae in the vicinity of the neoplasm. The histologic diagnosis was mixed germ cell tumor in this patient. MRI was accurate in the assessment of the local extent of 11 of 12 category T2 malignant testicular tumors (92%). In these cases, localized interruption of the low-signal-intensity testicular tunicae (Fig. 7), accompanied by a contrast-enhancing mass in the para testicular space (Fig. 2) in three cases, was the MRI finding suggesting the local stage of disease. One case of category T2 embryonal carcinoma was incorrectly categorized at MRI. In this case, the hypointense testicular tunicae were seen as intact, a small rim of normal-appearing testicular parenchyma was detected in the periphery of the neoplasm, and the tumor was incorrectly categorized T1. Finally, all T3 neoplasms were correctly categorized at MRI examination (Fig. 3), the presence of a contrastenhancing component in the spermatic cord suggesting neoplastic infiltration. Overall, the accuracy of MRI examination in the assessment of the local extent of malignant testicular tumors in our study was 92.8% (26/28). Discussion The clinical manifestations of scrotal masses are variable. Sonography is the first option and often the only imaging method for evaluating acute and nonacute scrotal disease [1, Fig. 6 56-year-old man with hemorrhagic necrosis of right testicle., Transverse T1-weighted image shows hyperintense right testicle (arrow)., oronal T2-weighted MR image shows very low signal intensity of right testicle (short arrow) and low signal intensity of right epididymis (long arrow). Star indicates small left hydrocele., Transverse contrast-enhanced T1-weighted image shows high signal intensity of right testicle (arrow) and lack of testicular enhancement after gadolinium administration. 686 JR:194, March 2010

MRI of Testicular Neoplasms TLE 3: MRI Findings in Diagnosis of Malignant Versus enign Testicular Masses: Signal Intensity of Lesion ompared With That of Normal Testicular Parenchyma Lesion signal intensity Malignant Diagnosis Isointense on T1-weighted image Hypointense on T2-weighted image Heterogeneous signal on T2-weighted image reas of hemorrhage reas of necrosis ontrast enhancement Enhancement of tumoral septa Heterogeneous enhancement Lesion extension Extension or invasion of the testicular tunicae Extension or invasion of the epididymis Extension or invasion of the spermatic cord 2]. MRI of the scrotum has been satisfactory in the evaluation of diverse scrotal diseases [4 16, 25 33]. The advantages of the technique include acquisition of precise anatomic information, satisfactory tissue contrast, and imaging in various planes [4 16, 25 33]. MRI has proved accurate in the differentiation of extratesticular from intratesticular disease, being superior to sonography in cases in which the scrotum is markedly enlarged [7, 25 28]. MRI findings with respect to tumor location, morphologic features, and tissue characterization can aid in narrowing the differential diagnosis in cases of extratesticular masses [25 28]. lthough most solid intratesticular masses should be considered malignant, it is important to recognize benign testicular entities for which radical orchiectomy is unnecessary. MRI of the testicles may yield satisfactory results in preoperative characterization of the histologic nature of various intratesticular mass lesions in terms of morphologic information and showing the presence of fat, fibrous tissue, fluid, and solid tissue within the masses. possible diagnosis of benign lesion based on MRI features may improve patient care and decrease the number of unnecessary radical surgical procedures [7 11, 16, 18, 19, 30, 31]. In cases of testicular malignancy, preoperative characterization of the histologic type of testicular tumors as enign Diagnosis Tubular ectasia of the rete testis Hypointense on T1-weighted image Hyperintense on T2-weighted image Fibrosis Hypointense on T1-weighted image Extremely hypointense on T2-weighted image Hemorrhagic necrosis Hyperintense on T1-weighted image Extremely hypointense on T2-weighted image No enhancement No lesion extension seminomatous and nonseminomatous also is possible [12, 13]. MRI, including dynamic contrast-enhanced imaging, also may yield additional information about testicular enhancement, thereby improving the diagnostic performance of the technique in lesion characterization [29, 32, 33]. Several groups of investigators [9, 20 22] have proposed organ-sparing surgery in cases of benign intratesticular lesions in patients with bilateral testicular neoplasms or a tumor in a solitary testis. Reliable preoperative evaluation of the local stage of disease is necessary in the care of patients referred for testis-sparing surgery. In our series, we evaluated the diagnostic performance of MRI in preoperative characterization of testicular neoplasms and, in cases of malignancy, prediction of the local extent of the disease. Our findings showed MRI accurate in the preoperative characterization of testicular neoplasms with an overall accuracy of 96.4%, a value in accordance with other published reports [8, 10, 11]. ll cases of testicular neoplasms (n = 28), and all but one case of benign testicular lesions (n = 7) were correctly characterized on the basis of the MRI findings in this study. The first report, to our knowledge, on the utility of MRI in the differentiation of malignant and benign testicular masses was by Thurnher et al. [7]. This group of authors used a low-field-strength system to examine 20 patients with intratesticular lesions, including 14 testicular neoplasms. Only one case of multiple testicular cysts was correctly assessed; the other benign intratesticular lesions, including four cases of inflammation and one case of testicular infarction, were falsely misdiagnosed as malignant. Subsequent studies [8, 10, 11], conducted with a 1.5-T MRI unit, had more satisfactory results in the characterization of intratesticular lesions with MRI. Schultz-Lampel et al. [8], in a study involving 88 patients with clinically suspected testicular cancer, found that MRI findings enabled correct characterization of all cases of both malignant and benign scrotal disease, although the benign diagnoses included mainly paratesticular lesions. In a study of 29 testicu- Fig. 7 41-year-old man with embryonal carcinoma of left testicle. Pathologic finding was tumor invading tunica vaginalis (category pt2). and, Transverse () and sagittal () T2-weighted MR images show inhomogeneous left testicular neoplasm. Localized interruption of testicular tunicae by tumor (arrow) and moderate left hydrocele (star, ) are evident. JR:194, March 2010 687

Tsili et al. lar masses, Serra et al. [11] found that MRI had an overall accuracy of 91% in differentiating benign from malignant testicular lesions after inconclusive clinical and sonographic evaluations. Muglia et al. [10] reported reliable characterization of 15 of 17 cases of testicular tumors and pseudotumors on the basis of MRI findings in patients referred because of nondiagnostic sonographic findings. In that study, two of the cases falsely diagnosed as malignant at MRI examination were proved at pathologic examination to correspond to chronic orchitis. This situation was similar to the false-positive case of granulomatous orchitis in our study. On images, scrotal inflammatory disease can mimic neoplasia in all phases of evolution, and clinical correlation is necessary for correct diagnosis. Granulomatous orchitis, however, is considered a common cause of false-positive results of scrotal imaging [7, 9, 10]. lthough the disease primarily involves the epididymis, in rare instances an isolated intratesticular mass detected at imaging is easily misdiagnosed as malignant [7, 9, 10]. In one of our patients, a small intratesticular lesion was depicted as mainly hypointense on T2-weighted images and with enhancement after gadolinium administration; malignancy therefore could not be excluded. The MRI criteria we used to characterize malignant testicular tumors included the presence of a mainly hypointense mass (15 of 28 cases) or a heterogeneous lesion (13 of 28 cases) on T2-weighted images that became inhomogeneously enhanced after gadolinium administration. The signal intensity of all testicular neoplasms was similar to that of the contralateral testis on T1-weighted images, being accompanied by areas of hemorrhage in 10 cases (36%). The presence of fat within two mature cystic testicular teratomas was visualized as high signal intensity on T1-weighted images. dramatic decrease in signal intensity was observed on fat-suppressed images. Preoperative differentiation of these tumors from testicular epidermoid cyst is extremely important [30, 31]. Epidermoid cysts of the testis are rare benign intratesticular lesions, often containing fatlike components [30, 31]. The absence of contrast enhancement in a fat-containing testicular mass is necessary to suggest the benign diagnosis in these patients [30, 31]. In our study, both lesions exhibited contrast enhancement that strongly suggested malignancy. reas of necrosis, visualized as areas of high signal intensity on T2-weighted images that did not become enhanced after contrast administration, were seen in 11 of 28 tumors (39%) in our study. ands of low signal intensity on T2-weighted images with more pronounced enhancement than the rest of the tumor were seen in 13 neoplasms in this study, all seminomatous. In a previous report [13], this feature was reported as indicative of the diagnosis of seminoma. hypointense halo was found in the periphery of nine of 28 neoplasms (32%) in our series. This finding was proved at histologic examination to correspond to a fibrous capsule. The presence of a pseudocapsule in cases of testicular malignancy is reported to facilitate tumor enucleation when testis-sparing surgery is planned [20]. Our study showed MRI examination accurate in the detection of this tumor capsule. In this study, the overall accuracy of MRI in estimating the local extent of malignant testicular tumors was 93%. lower accuracy (63%) was reported by Thurnher et al. [7] in using MRI to estimate the local stage of 14 primary malignant testicular neoplasms, probably because of the lower field strength (0.35 T) of the MRI unit used in that study. Twelve of 13 category T1 (92%) and 11 of 12 category T2 (92%) tumors were correctly assessed on the basis of the MRI findings in our series. The presence of intact testicular tunicae, detected as a hypointense halo on T2- weighted images in the periphery of the testis, was the main imaging feature used to discriminate category T1 and T2 malignant testicular tumors. More specifically, 11 of 12 category T1 neoplasms were visualized close to the testicular tunicae on MR images. However, the testicular tunicae were found intact on images in these cases, and this finding was proved at histopathologic examination. In one case, the tumor was surrounded by a rim of normal testicular tissue, confirming that it was confined within the testis, as proved pathologically. The localized interruption of the hypointense testicular tunicae found at MRI was detected in 11 of 12 category T2 tumors, suggesting the local stage of the disease in these cases. In two patients, MRI proved inaccurate in evaluation of the local extent of the neoplasms. T1 tumor was overstaged because of failure to recognize as intact the testicular tunicae surrounding the mass. category T2 neoplasm was incorrectly classified as T1 at imaging. In this patient, the testicular tunicae were assessed as intact, and a normal-appearing testicular rim was found to surround the neoplasm. The histologic result, however, was invasion of the tunica vaginalis. Invasion of the epididymis (category T2, n = 3) and the spermatic cord (category T3, n = 3) was correctly assessed in all cases in our study. The presence of an inhomogeneously enhancing mass involving the area of the epididymis or the spermatic cord was the main imaging finding suggesting the local stage of disease in these patients. Our study showed that MRI yielded accurate diagnostic information about the benign nature of intratesticular mass lesions, leading to a specific histologic diagnosis in most cases, namely, in patients with tubular ectasia of the rete testis, testicular fibrosis, and testicular hemorrhagic necrosis. The overall accuracy of MRI in the characterization of benign testicular lesions was 87.5% (seven of eight lesions), and the negative predictive value was 100%. The absence of contrast enhancement was the most sensitive sign for predicting the benign nature of intratesticular masses in this study. Our results are in accordance with those in other reports [10, 11]. Tubular ectasia of the rete testis is a relatively common benign intratesticular entity that can mimic neoplasia at clinical and sonographic evaluation [9, 17 19]. The MRI diagnosis, however, usually is straightforward in these patients, as it was in our study, and surgery can be avoided [9, 17 19]. Tubular ectasia is typically located in the mediastinum testis, is often bilateral, as it was in both of our patients, and frequently is associated with a spermatocele, as in one of these patients. The MRI findings are typical, including multicystic masses of variable size with signal intensity similar to that of water (low signal intensity on T1-weighted images, high signal intensity on T2-weighted images) [9, 17 19]. fter gadolinium administration, no lesion enhancement is seen [9, 17 19]. These features were found in both patients in this study who had bilateral tubular ectasia of the rete testis. Testicular fibrosis can be a worrisome finding at ultrasound examination. The MRI criteria used to characterize this benign entity are typical, including the presence of lesions of low and very low signal intensity on T1- and T2-weighted images, respectively, that do not become enhanced after gadolinium administration [9, 10] (Fig. 5). These findings were made in our study. The preoperative diagnosis of testicular hemorrhagic necrosis based on MRI findings also is possible [29]. The presence of a hyperintense testis on T1-weighted images with very low signal intensity or with a spotty or streaked pattern of very low signal intensity on T2-weighted images without enhancement after contrast administration is considered strongly suggestive of the diagno- 688 JR:194, March 2010

MRI of Testicular Neoplasms sis of hemorrhagic necrosis of the testis, as found in two cases in our study [29]. Our study showed MRI accurate in the preoperative characterization of testicular malignancy and assessment of the local extent of disease. We found both high sensitivity (100%; I, 87.9 100%) and high specificity (87.5%; I, 52.9 97.7%) in differentiating benign from malignant intratesticular lesions. The rate of correspondence between MRI and histopathologic diagnosis in regard to local extension of testicular neoplasms also was satisfactory (92.8%). Our results suggest that although it cannot be considered the first imaging technique in the investigation of testicular masses, MRI of the scrotum may be an efficient technique for testicular imaging. suggested presurgical diagnosis of benign testicular lesion at MRI evaluation may obviate unnecessary radical orchiectomy, reducing the incidence of diagnostic surgical exploration for testicular disease and improving patient care. Furthermore, preoperative MRI evaluation of the local extent of malignant testicular tumors may be recommended in the care of patients for whom testis-sparing surgery is planned, such as those with bilateral germ cell tumors or tumor in a solitary testis. There were limitations to our study. First, our sample included only a small number of benign testicular lesions. It is known, however, that malignant intratesticular lesions greatly outnumber benign ones. Furthermore, only a single reading of the MRI findings by two radiologists in consensus was performed, so interobserver reliability was not assessed. major limitation of this study was that although all patients underwent both sonographic and MRI evaluation of the scrotum, no direct evaluation of the results with these techniques was performed. comparison of the diagnostic performances of sonography and MRI might justify the role of MRI compared with ultrasound in the diagnosis and characterization of testicular disease. We conclude that MRI is an efficient diagnostic tool for the evaluation of testicular masses. It is accurate in the preoperative differentiation of benign and malignant intratesticular masses, facilitating accurate estimation of the local extent of disease in patients with malignant tumors. References 1. Dogra VS, Gottlieb RH, Oka M, Rubens DJ. Sonography of the scrotum. Radiology 2003; 227: 18 36 2. enson, Doubilet PM, Richie JP. Sonography of the male genital tract. JR 1989; 153:705 713 3. Sohaib S, Koh DW, Husband JE. The role of imaging in the diagnosis, staging and management of testicular cancer. JR 2008; 191:387 395 4. Schnall M. Magnetic resonance imaging of the scrotum. Semin Roentgenol 1993; 28:19 30 5. aker LL, Hajek P, urkhard TK, et al. Magnetic resonance imaging of the scrotum: normal anatomy. Radiology 1987; 163:89 92 6. aker LL, Hajek P, urkhard TK, et al. Magnetic resonance imaging of the scrotum: pathologic conditions. Radiology 1987; 163:93 98 7. Thurnher S, Hricak H, aroll PR, Pobiel RS, Filly R. Imaging the testis: comparison between MR imaging and US. Radiology 1988; 167:631 636 8. Schultz-Lampel D, ogaert G, Thuroff JW, Schlegel E, ramer. MRI for evaluation of scrotal pathology. Urol Res 1991; 19:289 292 9. Woodward PJ, Sohaey R, O Donoghue MJ, Green DE. Tumors and tumorlike lesions of the testis: radiologic pathologic correlation. RadioGraphics 2002; 22:189 216 10. Muglia V, Tucci S, Elias J, Trad S, ilbey J, ooperberg PL. Magnetic resonance imaging of scrotal diseases: when it makes the difference. Urology 2002; 59:419 423 11. Serra D, Hricak H, oakley FV, et al. Inconclusive clinical and ultrasound evaluation of the scrotum: impact of magnetic resonance imaging on patient management and cost. Urology 1998; 51:1018 1021 12. Johnson JO, Mattrey RF, Phillipson J. Differentiation of seminomatous from nonseminomatous testicular tumors with MR imaging. JR 1990; 154:539 543 13. Tsili, Tsampoulas, Giannakopoulos X, et al. MRI in the histologic characterization of testicular neoplasms. JR 2007; 189:1473; [web]w331 W337 14. Fritzche PJ, Hricak H, Kogan, Winkler ML, Tanagho E. Undescended testis: value of MR imaging. JR 1987; 164:169 173 15. aker LL, Hajek P, urkhard TK, Mattrey RF. Polyorchidism: evaluation by MR. JR 1987; 148:305 306 16. Fernandez-Perez G, Tardaguila FM, Velasco M. Radiologic findings of segmental testicular infarction. JR 2005; 184:1587 1593 17. rown DL, enson, Doherty FJ, et al. ystic testicular mass caused by dilated rete testis: sonographic findings in 31 cases. JR 1992; 158:1257 1259 18. Tartar VM, Trambert M, alsara ZN, Mattrey RF. Tubular ectasia of the testicle: sonographic and MR imaging appearance. JR 1993; 160:539 542 19. Rouvière O, ouvier R, Pangaud, Jeune, Dawahra M, Lyonnet D. Tubular ectasia of the rete testis: a potential pitfall in scrotal imaging. Eur Radiol 1999; 9:1862 1868 20. Yossepowitch O, aniel J. Role of organ-sparing surgery in germ cell tumors of the testis. Urology 2004; 63:421 427 21. lbers P, lbrecht W, lgaba F, et al. Guidelines on testicular cancer. Eur Urol 2005; 48:885 894 22. Heidenreich, Weissbach L, Holtl W, et al. Organ sparing surgery for malignant germ cell tumor of the testis. J Urol 2001; 166:2161 2165 23. Krege S, eyer J, Souchon R, et al. European consensus on diagnosis and treatment of germ cell cancer: a report of the second meeting of the European Germ ell ancer onsensus group (EGG). Part 1. Eur Urol 2008; 53:473 496 24. [No authors listed]. International germ cell consensus classification: a prognostic factor-based staging system for metastatic germ cell cancers. International Germ ell ancer ollaborative Group. J lin Oncol 1997; 15:594 603 25. kbar S, Sayyed T, Jafri SZ, Hasteh F, Neil JS. Multimodality imaging of paratesticular neoplasms and their rare mimics. RadioGraphics 2003; 23:1461 1476 26. Woodward PJ, Schwab M, Sesterhenn I. Extratesticular scrotal masses: radiologic pathologic correlation. RadioGraphics 2003; 23:215 240 27. Tsili, Tsampoulas, Giannakopoulos X, et al. Solitary fibrous tumor of the epididymis: MRI features. r J Radiol 2005; 78:565 568 28. Patel MD, Silva. MRI of an adenomatoid tumor of the tunica albuginea. JR 2004; 182:415 417 29. Watanabe Y, Nagayama M, Okumura, et al. MR imaging of testicular torsion: features of testicular hemorrhagic necrosis and clinical outcomes. J Magn Reson Imaging 2007; 26:100 108 30. ho JH, hang J, Park H, Lee JG, Son H. Sonographic and MR imaging findings of testicular epidermoid cysts. JR 2002; 178:743 748 31. Langer JE, Ramchandani P, Siegelman ES, anner MP. Epidermoid cysts of the testicle: sonographic and MR imaging features. JR 1999; 173:1295 1299 32. Watanabe Y, Dohke M, Ohkubo K, et al. Scrotal disorders: evaluation of testicular enhancement patterns at dynamic contrast-enhanced subtraction MR imaging. Radiology 2000; 217:219 227 33. hoyke PL. Dynamic contrast-enhanced MR imaging of the scrotum: reality check. Radiology 2000; 217:14 15 FOR YOUR INFORMTION This article is available for ME credit. See www.arrs.org for more information. JR:194, March 2010 689