MRI of Patients With Suspected Scrotal or Testicular Lesions: Diagnostic Value in Daily Practice

Transcription

1 Genitourinary Imaging Original Research Mohrs et al. Scrotal and Testicular MRI Genitourinary Imaging Original Research Oliver K. Mohrs 1,2 Henrik Thoms 1 Tobias Egner 3 Anne Brunier 1 Michael Eiers 2 Hans-Ulrich Kauczor 2 Peter Hallscheidt 2 Mohrs OK, Thoms H, Egner T, et al. Keywords: inflammation, mass, MRI, scrotum, testicles DOI: /AJR Received June 9, 2011; accepted after revision January 27, Department of Cardiovascular Imaging at Alice Hospital, Darmstadt Radiology, Dieburger Strasse 29-31, D Darmstadt, Germany. Address correspondence to O. K. Mohrs (mohrs@radiologie-darmstadt.de). 2 Department of Diagnostic and Interventional Radiology, University of Heidelberg, Heidelberg, Germany. 3 Department of Urology, Missio Hospital Würzburg, Würzburg, Germany. AJR 2012; 199: X/12/ American Roentgen Ray Society MRI of Patients With Suspected Scrotal or Testicular Lesions: Diagnostic Value in Daily Practice OBJECTIVE. The purpose of this study was to evaluate retrospectively the diagnostic value of MRI in the care of patients with suspected scrotal disorders. MATERIALS AND METHODS. Testicle-containing scrotal compartments were evaluated by examiners blinded to all clinical information in an interdisciplinary consensus reading after MRI was performed with standardized T1-, T2-, and contrast-enhanced T1-weighted sequences. Image quality was evaluated on an increasing 4-point scale. Nineteen binary diagnostic parameters were allocated 1 point each if the diagnosis was correct and were summarized into categories in terms of localization, pattern, and characterization of the lesions. The results were compared with histopathologic results in 22 cases, with the surgical report in one case, and with a combined clinical reference comprising medical reports and scrotal ultrasound, follow-up, or laboratory results in 61 cases. RESULTS. A total of 165 scrotal compartments containing testicles were analyzed in 84 patients; three patients underwent single-sided orchiectomy before MRI. The mean image quality score was 3.2 ± 0.7. Percentage of maximum diagnostic points, sensitivity, specificity, and positive and negative predictive values were 100% for localization of scrotal lesions and analysis of the pattern of testicular lesions. The values were 95%, 87%, 97%, 88%, and 97% for analysis of the pattern of nontesticular lesions and 95%, 92%, 97%, 91%, and 97% for classification of scrotal lesions. CONCLUSION. Scrotal lesions can be reliably detected, precisely localized, and characterized with MRI in everyday practice. The differential diagnosis was narrowed by use of MRI. This finding is of great clinical importance because more precise treatment strategies can be determined and surgical exploration or orchiectomy avoided by more patients. T he incidence of testicular cancer has been rising and has almost doubled over the past 40 years [1]. Intratesticular masses are usually malignant. However, the findings are sometimes contradictory or inconclusive. Correct diagnosis and characterization of scrotal and testicular masses are important for optimal treatment, including resection planning to avoid orchiectomy for some subtypes of benign tumors for which enucleation is an alternative [2, 3]. Moreover, differentiation of malignant from benign masses and from tumorlike lesions such as congenital, traumatic, and inflammatory disorders is clinically relevant. The routine urologic pathway involves manual palpation and ultrasound of suspected scrotal and testicular lesions. Because of the limited diagnostic capability of sonography, neoplasms may be diagnosed in inconclusive cases, and orchiectomy performed. The testi- cles of some of these patients probably could be saved if a better noninvasive method were available for characterizing these lesions. The literature on testicular and scrotal MRI is sparse and only presents reviews [4 8], case reports [9, 10], and a small number of original studies that address isolated diagnostic aspects such as congenital disorders [11], scrotal ischemia [12, 13], and specific neoplasms [14, 15]. Most of these studies were performed with a small number of patients and did not always involve state-of-the-art MRI. It remains unclear whether MRI might play an additional role in the diagnostic pathway for characterizing scrotal disorders in everyday practice. The purpose of this study was to evaluate, in a retrospective interdisciplinary consensus reading in which the observers were blinded, the diagnostic value of MRI in the routine care of patients with suspected scrotal or testicular lesions. AJR:199, September

2 Mohrs et al. Materials and Methods Study Sample A total of 100 men and boys with suspected scrotal or testicular lesions were referred for MRI for clinical reasons between 2005 and 2010 and were consecutively enrolled in this retrospective analysis. Clinical MRI was a noninvasive method of enhancing accuracy in confirming and classifying suspected lesions. Sixteen patients were excluded from the data analysis because they refused surgical treatment or were lost to follow-up. The mean age of the other 84 patients was 40 ± 18 (SD) years (range, 9 75 years). The combined information from medical (n = 84), surgical (n = 23), and histopathologic reports (n = 22), scrotal gray-scale and color-encoded ultrasound reports (n = 84), and laboratory results (n = 84) served as the clinical reference for identifying the culprit lesion. The results were compared with histopathologic results in 22 cases, with the surgical report in one case, and with a combined clinical reference comprising medical reports and scrotal ultrasound, follow-up, or laboratory results in 61 cases. MRI Sixty-three patients underwent MRI with a 1.5-T system (Magnetom Avanto SQ Class, Siemens Healthcare), and in 21 patients with a 1-T system (Genesis Signa, GE Healthcare). All patients were supine and headfirst for the examination. The scrotum was supported by a towel placed between the thighs to avoid motion artifacts and to maintain the same distance from the coil. For signal detection, a high-resolution phased-array coil was used in a centered position toward the scrotum. The scrotal imaging protocol comprised the following sequences with typical parameters: axial fat-suppressed T1-weighted turbo spin echo (TSE) (TR/TE, 703/11; flip angle, 150 ; matrix size, ; voxel size, mm; slice thickness, 3 mm; gap, 0.6 mm; 24 slices), coronal T1-weighted spin echo (TR/TE, 596/16; flip angle, 90 ; matrix size, ; voxel size, mm; slice thickness, 3 mm; gap, 0.6 mm; 24 slices), axial fat-suppressed T2-weighted TSE (TR/TE, 5140/95; flip angle, 150 ; matrix size, ; voxel size, mm; slice thickness, 3 mm; gap, 0.6 mm; 24 slices), and coronary fat-suppressed T2-weighted TSE (TR/ TE, 5010/95; flip angle, 150 ; matrix size, ; voxel size, mm; slice thickness, 3 mm; gap, 0.6 mm; 24 slices). Approximately 1 minute after manual administration of 0.1 mmol/kg body weight gadopentetate dimeglumine (Magnevist, Bayer Schering Pharma) into an antecubital vein, axial and coronary fat-suppressed T1-weighted TSE images (TR/TE, 703/11; flip angle, 150 ; matrix size, ; voxel size, mm; slice thickness, 3 mm; gap, 0.6 mm; 24 slices) were obtained. Image Analysis To combine interdisciplinary knowledge, the MRI datasets were evaluated by consensus of three observers (two radiologists experienced in either uroradiology or MRI and one urologist). The examiners were blinded to all clinical data. The MR image quality was graded on a 4-point scale: excellent (4 points), good (3 points), moderate (2 points), or poor (1 point). If multiple lesions were present in a scrotal compartment, a clinically identified culprit lesion (e.g., neoplasm or orchitis) was defined and analyzed, and the nonculprit lesions (e.g., common incidental findings) were classified as additional lesions and did not count for analysis. A total of 19 single parameters were summarized in four diagnostic categories and analyzed in a binary manner (yes-no) for every culprit lesion for categories A D. If assessable, each parameter was allocated 1 point if the diagnosis was correct compared with the available references. All lesions were compared with clinical findings, laboratory results, and scrotal ultrasound findings. For patients who underwent surgery, the MRI findings were also compared with the surgical and histopathologic results. Category A represented identification and localization of the clinically identified culprit lesions as follows: 1, identified; 2, main part intratesticular; 3, main part extratesticular. Category B described the individual pattern of the testicular lesions as follows: 4, cystic; 5, solid; 6, mixed pattern; 7, edema; 8, hemorrhage; and 9, tunica albuginea involvement. Category C described the individual pattern of the nontesticular lesions as follows: 10, cystic; 11, solid; 12, mixed; 13, vascular or tubular; 14, edema or liquid; and 15, hemorrhage. In categories B and C, cystic lesions were diagnosed if there was homogeneously high signal intensity on T2-weighted images and low signal intensity on T1-weighted images without enhancement. Solid lesions were diagnosed with low T2 signal intensity and low T1 signal intensity and enhancement. A mixed pattern was a combination of the two. Edema was expected in solid tissue with diffuse hyperintensity on T2-weighted fat-suppressed images and hypointensity on T1- weighted images. Hemorrhage was confirmed in lesions with high T1 signal intensity and varying T2 signal intensity depending on the age of the hematoma without loss of signal intensity on fatsuppressed images. The parenchymal pattern was verified by comparison with the references and was based on knowledge of vascularization, tissue characterization, and enhancement pattern of the reference diagnosis in the literature [4 18]. The lesion could have multiple aspects in this category. Category D comprised the classification of the lesions as follows: 16, malignant; 17, benign tumorlike lesion; 18, benign inflammatory lesion; and 19, benign traumatic lesion. Typical anatomic criteria and morphologic patterns described in the literature were used to classify the lesions [4 16]. In category D, seminoma was diagnosed in testicular lesions with relatively homogeneously low T2 signal intensity, intermediate signal intensity on T1-weighted images, and possible enhancing fibrovascular septa with low signal intensity on unenhanced T1- and T2-weighted images [5, 6, 8, 14, 15, 17, 18]. A nonseminomatous lesion was expected in testicular masses with heterogeneous signal intensity with isointensity to hyperintensity on T1-weighted images and areas of high and low signal intensity on T2-weighted images with possible hemorrhage and heterogeneous enhancement [5, 6, 8, 14, 15, 17, 18]. Teratoma was diagnosed in a lesion in which fat was present with hyperintensity on T1- and T2-weighted images, loss of signal intensity on fat-suppressed images, and enhancement [15]. Aspects of malignancy were areas of hemorrhage, necrotic parts, invasive growth of the tumor with destruction of testicular septation, infiltration of the extratesticular space, and destruction of anatomic borders [14, 17 20]. Epididymitis was diagnosed in cases of diffuse enlargement of the epididymis with a possible decrease of T2 signal intensity and hypervascularity with marked enhancement [8, 17]. Orchitis was diagnosed in cases of heterogeneous areas of low signal intensity on T2-weighted images, hypointense bands in the testis with a possible decrease in enhancement, and more enhancing regions. Epididymoorchitis was a combination of epididymitis and orchitis. In liquid lesions with a regular enhancing wall, an abscess was diagnosed [5, 8, 17]. With regard to cystic lesions, topographic localization led to a diagnosis of testicular cyst, testicular appendix, epididymal appendix, or spermatocele [7, 17]. Dilated rete testis was diagnosed in a lesion with an enlarged mediastinum testis with low signal intensity on T1-weighted images and high signal intensity on T2-weighted images without enhancement [6, 8]. Varicocele was diagnosed in cases of spermatic cord mass with a serpiginous appearance with signal intensity varying from high on T2-weighted images to flow void, depending on blood flow velocity and enhancement [7, 17]. Hematoma was suggested in cases of nonenhancing lesions with high T1 signal intensity without loss of signal intensity on fat-suppressed images and varying T2 signal intensity depending on the age of the hematoma [8, 17, 18]. 610 AJR:199, September 2012

3 Scrotal and Testicular MRI Statistical Analysis Image quality was assessed as mean ± SD. For the four diagnostic categories and for each of the 19 single parameters, the total number of points and the sensitivity, specificity, and positive (PPV) and negative (NPV) predictive values were calculated. Results Image Quality The mean overall image quality was 3.2 ± 0.7 points on the 4-point scale. The datasets were given 4 points (excellent) in 30 cases, 3 points (good) in 42 cases, 2 points (moderate) in 12 cases, and 1 point (poor) in no case. A total of 165 scrotal compartments containing testicles were analyzed in 84 patients; three patients underwent single-sided orchiectomy before MRI. On the basis of the reference diagnosis, 70 patients had a clinically identified culprit lesion in a single compartment and five patients in both compartments, resulting in a total of 80 scrotal lesions containing neoplasms in eight cases (seminoma in six cases; nonseminoma in two cases, one differentiated teratoma and one Leydig cell tumor), traumatic lesions in six cases (one case of hemorrhage and five of scrotal hematoma or hematocele), inflammatory lesions in 36 cases (epididymitis, 22 cases; epididymoorchitis and orchitis, six cases; aftermath of inflammation, two cases; and abscess, six cases), and tumorlike lesions in 25 cases (testicular hydatid cyst, two cases; testicular cyst, two cases; appendix epididymis, two cases; spermatocele, 11 cases; varicocele, six cases; atheromatous cyst of deferens duct, one case; and hydrocele without other pathologic changes, one case). The other 85 scrotal compartments did not contain a culprit lesion. Figures 1 5 show examples of scrotal lesions. TABLE 1: Diagnostic Accuracy of Major Categories and Single Parameters in a Binary Approach Category A: Identification and Localization of the Clinically Identified Culprit Lesions With MRI, all clinically identified culprit lesions (n = 80) in 165 scrotal compartments were identified and localized correctly either as testicular (28 lesions) or extratesticular (52 lesions) compared with the reference diagnosis, resulting in 100% yielded points, sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV). Category B: Individual Pattern of Testicular Lesions The individual pattern of the 28 testicular lesions was correctly analyzed as cystic (three lesions), solid (12 lesions), mixed (five lesions), edema (11 lesions), hemorrhagic (two lesions), or tunical involvement (four lesions) in all cases. The yielded points, sensitivity, specificity, NPV, and PPV were 100%. Category C: Individual Pattern of the Extratesticular Lesions In all 52 extratesticular culprit lesions, the solid pattern (parameter 11) was described correctly compared with the reference diagnosis. Most of the other single parameters in category C reached yielded points, sensitivity, specificity, NPV, and PPV of less than 100% (Table 1). A total of 12 lesions with a n Yielded Points (%) Sensitivity (%) Specificity (%) Positive Predictive Value (%) Negative Predictive Value (%) Category A: identification and localization (495/495) , Identification of culprit lesion 80/ (165/165) , Localized as testicular 28/ (165/165) , Localized as extratesticular 52/ (165/165) Category B: pattern of testicular lesions (168/168) , Cystic 3/ (28/28) , Solid 12/ (28/28) , Mixed 5/ (28/28) , Edema 11/ (28/28) , Hemorrhage 2/ (28/28) , Tunica albuginea involvement 4/ (28/28) Category C: pattern of nontesticular lesions (297/312) , Cystic 12/52 96 (50/52) , Solid 0/ (52/52) , Mixed 5/52 94 (49/52) , Vascular or tubular 7/52 94 (49/52) , Edema or liquid 31/52 92 (48/52) , Hemorrhage 5/52 94 (49/52) Category D: classification of lesions (305/320) , Malignant 7/80 98 (78/80) , Benign tumorlike lesion 28/80 96 (77/80) , Benign inflammatory lesion 48/80 91 (73/80) , Benign traumatic lesion 6/80 96 (77/80) Note Values in parentheses are raw numbers. AJR:199, September

4 Mohrs et al. Fig year-old man with right-sided orchitis. Axial T2-weighted fat-suppressed (left), T1-weighted (center), and T1-weighted fat-suppressed gadopentetate dimeglumine enhanced (right) MR images show typical tiger skin like pattern (circle) of testicle, especially after contrast enhancement. Fig year-old man with right-sided seminoma (asterisk) correctly localized and characterized with MRI. Coronal T2-weighted fat-suppressed (left), T1-weighted (center), and T1-weighted fat-suppressed gadopentetate dimeglumine enhanced (right) MR images show multinodular configuration with enhancing septa of mass and infiltration of tunica albuginea, epididymis, and spermatic cord (solid arrow). Noninfiltrated part of testicle has necrotic nonperfused parenchyma after contrast enhancement (dashed arrow). cystic pattern (parameter 10) were correctly classified, but in another two cases (two patients with epididymitis), the categorization as cystic was a false-positive result. Mixed pattern (parameter 12) was rated as true-positive in three cases, true-negative in 46 cases, false-positive in one case (a patient with scrotal hematoma), and false-negative in two cases (two cases of epididymitis). Six cases were true-positive and 43 were true-negative as vascular or tubular lesions (parameter 13), whereas two lesions (in two patients with chronic epididymitis) were rated false-positive and one lesion (in a patient with varicocele), false-negative. Lesions with edema or a liquid pattern (parameter 14) were true-positive in 28 cases, true-negative in 20 cases, false-positive in one case (in a patient with varicocele), and false-negative in three cases (in one patient with scrotal hematoma and two patients with chronic epididymitis). Hemorrhage (parameter 15) was correctly identified in three cases but was classified as false-positive in one case (in a patient with epididymoorchitis) and as false-negative in two cases (in two patients with scrotal hematoma). Category D: Categorization of the Lesions All seven malignant lesions were identified correctly, but two cases were classified as falsepositive (in one patient with benign Leydig cell tumor and one patient with orchitis). Of the 73 benign lesions, 27 were rated correctly as tumorlike lesions, but two (in two patients with chronic epididymitis) were false-positive and one (in a patient with infarcted testicular cyst) was false-negative. Inflammatory lesions were categorized as true-positive in 44 cases, falsepositive in three cases (in one patient with infarcted testicular cyst and two patients with scrotal hematoma), and false-negative in four cases (in one patient with orchitis, one with epididymoorchitis, and two with chronic epididymitis). Traumatic lesions were true-positive in four cases but false-positive in one case (in a patient with epididymoorchitis) and false-negative in two cases (in two patients with scrotal hematoma). Discussion Our study showed the potential diagnostic value of MRI in everyday practice in the care of patients with suspected scrotal lesions. We found high diagnostic accuracy in identifying, localizing, and classifying these lesions in a consensus reading. High-resolution ultrasound with a combination of gray-scale and color-encoded techniques is the accepted clinical standard for imaging the scrotum. However, some studies 612 AJR:199, September 2012

5 Scrotal and Testicular MRI Fig year-old man with left-sided dilated rete testis (arrow) correctly diagnosed as benign tumorlike lesion with MRI. Coronal T2-weighted fat-suppressed (left), T1-weighted (center), and T1-weighted fat-suppressed gadopentetate dimeglumine enhanced (right) MR images show typically high signal intensity on T2-weighted image with tubular appearance without contrast enhancement. Fig year-old man with right-sided Leydig cell tumor (circle). Coronal T2-weighted fat-suppressed (left), T1-weighted (center), and T1-weighted fat-suppressed gadopentetate dimeglumine enhanced (right) MR images show testicular lesion falsely characterized as malignant testicular lesion with MRI. [21, 22] have shown that ultrasound can be hampered in evaluation of congenital disorders, such as undescended inguinal and intraabdominal testes, and that MRI can be used to solve this problem. This finding has great clinical importance because undescended testes are associated with substantially higher risk of development of malignancy [23]. MRI is the imaging modality of choice for these patients. For other patients, ultrasound findings may be inconclusive about the presence and precise classification of a scrotal lesion. Some tumorlike lesions may appear masslike on ultrasound images, for example, but have a typical presentation at MRI [6]. This discrepancy shows that there is a clinical need for MRI when ultrasound findings are inconclusive. In some countries, scrotal ultrasound is performed by urologists who refer the patients for MRI if the findings are inconclusive. It is not always possible, however, for radiologists to receive a detailed briefing about the differential diagnosis assessed with the urologic examination findings, especially in everyday practice. Radiologists therefore have to make the diagnosis solely on the basis of MRI findings. MRI has to be validated, however, and readers need a learning curve to improve their diagnostic accuracy. Our concept involved consecutively collecting records on patients undergoing routine care for various kinds of scrotal disorders, such as congenital, inflammatory, vascular, and traumatic lesions and masses. We analyzed the overall diagnostic value of MRI in the treatment of patients with suspected testicular and nontesticular scrotal lesions as encountered in routine clinical practice. In this interdisciplinary retrospective study, blinded observers analyzed the MRI findings on 165 scrotal compartments in 84 patients, most of whom had scrotal pain or painless enlargement, by comparing the findings with those in the clinical examination report, clinical history, surgical reports, ultrasound histopathologic correlation, laboratory test results, and follow-up examination findings. Risk factors for the development of testicular malignancy include a history of testicular cancer, cryptorchidism, infertility, testicular dysgenesis, and a family history of the condition. However, early diagnosis of scrotal and testicular masses remains difficult because testicular lesions often grow without causing symptoms and present clinically at a late stage in the form of painless scrotal enlargement. The first step in diagnosing an unclear testicular lesion is sonography. In our study, however, MRI was reliable for detecting or ruling out a scrotal or testicular disorder. The second step is to localize the culprit lesion. Our results show the high diagnostic accuracy of MRI for AJR:199, September

6 Mohrs et al. Limitations Our study design involved a retrospective analysis and included heterogeneous data, sometimes without histopathologic correlation. This meant that for 23 of 84 of our patients, findings were correlated with histopathologic or surgical results, whereas for the other 61, combined information from medical reports, scrotal ultrasound, follow-up, and laboratory tests served as the reference. We did not analyze single-reader results or kappa values to assess interobserver variability. This analysis was not part of the study design because we wanted to combine interdisciplinary radiologic and urologic knowledge to prevent insufficient diagnostic accuracy due to insufficient experience of a single reader in these rare cases. However, a single reader with less experience might have lower diagnostic accuracy compared with our results. Fig year-old man with epididymis. Coronal T2-weighted fat-suppressed (left) and T1-weighted (right) MR images show left-sided capsulated scrotal hematoma (circle) falsely classified as nontesticular malignant lesion. differentiating intratesticular and extratesticular disorders. In some patients, it can be difficult to evaluate the continuity of the echogenic line representing the tunica albuginea using ultrasound. The low signal intensity of the tunica albuginea is a notable sign of its integrity on MR images. MRI helps to rule out tunica albuginea rupture after traumatic scrotal events, for example [8, 17]. Furthermore, because it allows precise localization of nonmalignant testicular lesions, MRI can be helpful in guiding surgical management. MRI yields additional information on T1- and T2-weighted images, and perfusion can be estimated on contrast-enhanced images; these findings can improve the differential diagnosis. MRI can also show the precise number and location of nonmalignant testicular lesions. In certain patients with benign tumors, this additional information can help save the testis because tumor enucleation can be performed instead of orchiectomy [12, 13]. Compared with ultrasound, MRI has greater utility in characterizing certain lesions, such as lipoma and other fat-containing lesions, hematoma, fibrous pseudotumor, and focal testicular infarction [4 8]. Through the use of different sequence types and the administration of gadolinium, MRI can be used to characterize the pattern of scrotal disorders [24]. Correct classification of testicular lesions is crucial because the peak prevalence of testicular tumors occurs among 25- to 35-year-old men [6]. Nonneoplastic lesions such as orchitis, hemorrhage, and infarction can mimic a testicular mass because the ultrasound criteria for masses overlap those for neoplasms, which can lead to unnecessary orchiectomy [25]. We found that evaluation of testicular lesions has higher diagnostic accuracy than evaluation of nontesticular scrotal lesions. Chronic epididymitis and scrotal hematoma in particular can be challenging to characterize with MRI. A 2010 study [19] highlighted the utility of MRI in characterization and staging of testicular neoplasms. The authors discussed the limitations of MRI in differentiating chronic inflammatory testicular lesions from malignant lesions. In our study, we found that although the cases were analyzed by observers blinded to all clinical information, MRI seems to offer additional information that can be used to diagnose inflammation correctly. The diagnostic accuracy in classifying scrotal disorders in routine patient care will thus increase when clinical information is available. In characterization of scrotal lesions, the MRI findings may allow a decrease in the number of surgical procedures and reduce costs. Conclusion Our study showed that MRI is reliable in the detection of scrotal lesions, which can be precisely localized. Furthermore, with MRI, scrotal disorders can be differentiated into testicular or nontesticular scrotal lesions, and various lesion types can be characterized, including cysts and fluid, solid masses, fat, and fibrosis. In this study, the differential diagnosis was narrowed on the basis of the MRI findings, and some lesions, such as dilated rete testis and seminoma, were identified directly. The study findings are of clinical importance because more precise treatment strategies can be determined and orchiectomy avoided. Acknowledgments We thank Wilhelm Jungmann and Thomas Weissenfels of the Urological Center at Alice-Hospital for their support. References 1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, CA Cancer J Clin 2007; 57: Sloan JC, Beck SD, Bihrle R, Foster RS. Bilateral testicular epidermoid cysts managed by partial orchiectomy. J Urol 2002; 167: Heidenreich A, Engelmann UH, Vietsch HV, Derschum W. Organ preserving surgery in testicular epidermoid cysts. J Urol 1995; 153: Sohaib SA, Koh DM, Husband JE. The role of imaging in the diagnosis, staging, and management of testicular cancer. AJR 2008; 191: Cassidy FH, Ishioka KM, McMahon CJ, et al. MR imaging of scrotal tumors and pseudotumors. RadioGraphics 2010; 30: Woodward PJ, Sohaey R, O Donoghue MJ, Green DE. From the archives of the AFIP: tumors and tumorlike lesions of the testis: radiologic-pathologic correlation. RadioGraphics 2002; 22: Woodward PJ, Schwab CM, Sesterhenn IA. From the archives of the AFIP: extratesticular scrotal masses: radiologic-pathologic correlation. Radio- Graphics 2003; 23: Kim W, Rosen MA, Langer JE, Banner MP, Siegelman ES, Ramchandani PUS. MR imaging correlation in pathologic conditions of the scrotum. RadioGraphics 2007; 27: Tsili AC, Tsampoulas C, Giannakis D, et al. Case report: tuberculous epididymo-orchitis MRI findings. Br J Radiol 2008; 81:e166 e Gotto GT, Chang SD, Nigro MK. MRI in the diagnosis of incomplete testicular torsion. Br J Radiol 2010; 83:e105 e AJR:199, September 2012

7 Scrotal and Testicular MRI 11. Lam WW, Tam PK, Ai VH, Chan KL, Chan FL, Leong L. Using gadolinium-infusion MR venography to show the impalpable testis in pediatric patients. AJR 2001; 176: Watanabe Y, Nagayama M, Okumura A, et al. MR imaging of testicular torsion: features of testicular hemorrhagic necrosis and clinical outcomes. J Magn Reson Imaging 2007; 26: Fernández-Pérez GC, Tardáguila FM, Velasco M, et al. Radiologic findings of segmental testicular infarction. AJR 2005; 184: Johnson JO, Mattrey RF, Phillipson J. Differentiation of seminomatous from nonseminomatous testicular tumors with MR imaging. AJR 1990; 154: Tsili AC, Tsampoulas C, Giannakopoulos X, et al. MRI in the histologic characterization of testicular neoplasms. AJR 2007; 189:1473; [web]w331 W Terai A, Yoshimura K, Ichioka K, et al. Dynamic contrast-enhanced subtraction magnetic resonance imaging in diagnostics of testicular torsion. Urology 2006; 67: Baker LL, Hajek PC, Burkhard TK, et al. MR imaging of the scrotum: pathologic conditions. Radiology 1987; 163: Cramer BM, Schlegel EA, Thueroff JW. MR imaging in the differential diagnosis of scrotal and testicular disease. RadioGraphics 1991; 11: Tsili AC, Argyropoulou MI, Giannakis D, Sofikitis N, Tsampoulas K. MRI in the characterization and local staging of testicular neoplasms. AJR 2010; 194: Muglia V, Tucci S, Elias J, Trad CS, Bilbey J, Cooperberg PL. Magnetic resonance imaging of scrotal diseases: when it makes the difference. Urology 2002; 59: Stikkelbroeck NM, Suliman HM, Otten BJ, Hermus AR, Blickman JG, Jager GJ. Testicular adrenal rest tumours in postpubertal males with congenital adrenal hyperplasia: sonographic and MR features. Eur Radiol 2003; 13: Fritzsche PJ, Hricak H, Kogan BA, Winkler ML, Tanagho EA. Undescended testis: value of MR imaging. Radiology 1987; 164: Hörmann M, Balassy C, Philipp MO, Pumberger W. Imaging of the scrotum in children. Eur Radiol 2004; 14: Müller-Leisse C, Bohndorf K, Stargardt A, et al. Gadolinium-enhanced T1-weighted versus T2- weighted imaging of scrotal disorders: is there an indication for MR imaging? J Magn Reson Imaging 1994; 4: Trambert MA, Mattrey RF, Levine D, Berthoty DP. Subacute scrotal pain: evaluation of torsion versus epididymitis with MR imaging. Radiology 1990; 175:53 56 FOR YOUR INFORMATION The AJR has made getting the articles you really want really easy with an online tool, Really Simple Syndication, available at It s simple. Click the RSS button located in the menu on the right side of the page. You ll be on your way to syndicating your AJR content in no time. AJR:199, September