whether imaging features could permit distinction between benign and malignant lesions.

Transcription

1 Complex Renal Cysts: Findings on MR Imaging N. Cern Balci1 Richard C. Semelka1 Richard H. Patt2 3 David Dubois2 John A. Freeman4 Andres Gomez-Caminero5 John 1. Woosley6 OBJECTIVE. We retrospectively evaluated our experience with complex cystic renal masses on MR imaging, using Tl-weighted. T2-weighted, and gadolinium-enhanced images, to determine whether imaging features could permit distinction between benign and malignant lesions. MATERIALS AND METHODS. Thirty-seven patients with complex cystic renal lesions were included in this retrospective study. The patients selected had undergone TI-weighted, T2- weighted, and gadolinium-enhanced MR imaging examinations using l.5-t scanners, with at least one of the following findings: cyst fluid of heterogeneous signal intensity. mural irregularity, septa, mural masses or nodules, increased mural thickness, or intense mural enhancement. The diagnosis was established by histology in 19 patients and by follow-up studies in the remaining 18 patients. RESULTS. Fifty-five complex renal cystic lesions were present in the 37 patients. Among the 55 lesions, of 37 that contained fluid of a heterogeneous signal intensity, eight were malignant (22%); of 16 with irregular walls, 10 were malignant (63%); of four with septa, two were malignant (50%): offour with mural masses or nodules, three were malignant (75%); of 14 with a thick wall (>2 mm), 10 were malignant (71%); and of 32 with intense mural enhancement, 14 were malignant (44%). As independent variables, mural irregularity, mural masses or nodules, increased mural thickness, and intense mural enhancement each were highly associated with malignancy (p = ). The combination of mural irregularity and intense mural enhancement had the highest correlation with malignancy (p =.0002). CONCLUSION. The combination of mural irregularity and intense mural enhancement is a strong predictor of malignancy in renal cystic lesions. However, the appearance of benign and malignant lesions may overlap, suggesting that distinct separation of these entities is not currently possible in all cases with MR imaging. Received October 19, 1998; accepted after revision December 14, Department of Radiology, University of North Carolina, CB 7510, Chapel Hill, NC Address correspondence to R. C. Semelka. 2Department of Radiology, Georgetown Medical Center, Washington, DC Berlex Laboratories, Wayne, NJ Department of Surgery, University of North Carolina, Chapel Hill, NC School of Public Health, University of North Carolina, Chapel Hill, NC tdepartment of Pathology, University of North Carolina, Chapel Hill, NC AJR 1999;172: X/99/ American Roentgen Ray Society T he criteria for a simple renal cyst on CT include an imperceptible cyst wall, a smooth interface with the renal parenchyma, a round or ovoid cyst wall, absence of mural enhancement after IV injection of contrast material, and a simple fluid content [1-4]. Cystic renal lesions that do not fulfill these criteria are considered complex and indeterminate, and their treatment has been challenging [I, 3, 5]. A classification system ofrenal cysts develo_ by Bosniak [I] is based on sonographic and CT findings and has been used for the treatment ofcomplex cystic renal masses. MR imaging has been used to further evaluate some complex renal lesions, such as hyperattenuating cystic lesions seen on CT and hyperechoic cysts seen on sonography [2, 3, 6-8]. Cyst fluid with a varying signal intensity is more common on MR images than on comparative CT images, reflecting the high sensitivity of MR imaging for the paramagnetic properties of blood breakdown products, proteinaceous fluid, or both [4, 7]. A spectrum of appearances has been reported for complex and simple cystic lesions on unenhanced MR images [3]; however, current practice is to routinely use gadolinium to evaluate renal masses more definiuvely. Despite the reliance on MR imaging to further characterize renal lesions, a systematic description of complex renal lesions on MR imaging using gadolinium administration has not, to our knowledge, been established. The purpose of this study was to assess the spectrum of appearances of complex cystic renal lesions on TI-weighted, U-weighted, and gadolinium-enhanced MR imaging and to correlate their MR appearances with histopathologic findings or imaging follow-up findings. Materials and Methods Patients Thirty-seven patients (21 men and 16 women; years old: mean age. 56 years old) who had AJR:172, June

2 Balci et al. complex cystic renal lesions and had undergone MR examination between September 1992 and May 1998 were included in the study. The inclusion criteria for complex cystic lesions were a signal intensity different from that of simple fluid (a low signal intensity on TI-weighted images or a high signal intensity on T2-weighted images). mural irregularity. septa. mural masses or nodules. increased mural thickness, or intense mural enhancement equal to or greater than that of the adjacent renal cortex. Among the 37 patients. 19 underwent surgical excision of the cystic lesion and 18 underwent followup MR imaging or CT I year or more after the first MR examination. In 12 patients, the follow-up examination was I 2 months after the initial MR examination: in eight of these, follow-up was by MR imaging. and in four. by CT. Six patients underwent MR imaging more than 15 months after the initial MR examination. We included patients with imaging follow-up because studies have shown that stability of complex cystic renal lesions for at least I year is an acceptable criterion fbr establishing that they are benign 121. Seventeen patients underwent MR examinalion for further evaluation of a cystic lesion that had been revealed either by s()nography (five patients. three of whom had an elevated level of serum creatinine) or by CT ( I 2 patients. The 20 other patients had not been previously imaged and underwent MR imaging to investigate the kidneys. Twelve of these 20 had an elevated level of serum creatinine. MR Imaging MR examinations were performed on a I.5-T scanner (Vision [ii = 181 or SP 4000 [n = 19]: Siemens Medical Systems. lselin. NJ). The sequences included TI -weighted breath-hold spoiled gradientecho images (TR rangetl E range /4-6: flip angle o) ( patients). fat-suppressed spoiled gradient-echo images with the same parameters (20 patients). fat-suppressed TI-weighted spin-echo images ( /20-25) (three patients), and T2- weighted images acquired as fat-suppressed turbo (fast) spin-echo ( /80-90) (all patients), breathing-independent half-fourier acquisition single-shot fast spin-echo (T1TEeff infinitel60) (eight patients). and fat-suppressed half-fourier acquisition single-shot fast spin-echo (infinitei0) (two patients). Gadolinium chelate (Magnevist: Berlex, Wayne. NJ) was administered at a dosage of 0. 1 mmol/kg as a 5- sec hand-injected bolus followed by a rapid flush of 10 ml of normal saline. Spoiled gradient-echo images were repeated immediately after contrast administration (early phase) and at 45 sec and 90 sec (intermediate phase) and 5-7 mm (delayed phase) after completion of the normal saline flush. In 23 patients, the 90-sec gadolinium-enhanced spoiled gradientecho sequence was performed using fat saturation. All sequences were acquired using sections and a thickness of 7-10 mm. The matrix was x 256 (phase encoding range x frequency encoding). Image Interpretation Images were retrospectively evaluated by two experienced investigators in concert. The investigators were unaware ofclinical information, and their evaluation was based on the imaging findings for the cystic lesions. During each examination. the number, size, and location of complex cystic renal lesions were determined, and the following features were evaluated: the signal intensity of the cyst contents on 11- weighted and T2-weighted images. including layering of debris or hemorrhage: mural irregularity: septa: mural masses or nodules: increased mural thickness: and intense mural enhancement. For both Ti- and 12-weighted sequences. we used the signal intensities of the renal cortex. medulla, urine, pancreas, adipose tissue, and psoas muscle as references in determining the signal intensities of lesions. For TI-weighted images, the cystic lesion was considered hypointense when its signal was comparable to that of urine in the bladder. isointense when similar to that of the cortex, and hyperintense when higher than that of adipose tissue. For TI-weighted fat-suppressed images. the cystic lesion was considered hyperintense when its signal was similar to or higher than that of the pancreas. On T2-weighted images. the assessed area was considered hypointense when its signal was similar to or lower than that of the psoas muscle, isointense when similar to that of the renal cortex. and hyperintense when similar to that of urine. Individual variables were evaluated separately to determine their correlation with malignancy and in combination to determine if increased correlation could be established. Data were tested for significance using the Mantel-Haenszel. chi-square. and Fisher s exact tests for individual and combined variables. Combinations of variables were evaluated on the basis of the Bosniak classification [I]. Results Fifty-five complex renal lesions were identified in the 37 patients. Twenty-two patients had a solitary complex cystic renal lesion, and I 5 patients had more than one complex cystic lesion. In two of these 15, the complex cysts were bilateral. The associated diseases included autosomal dominant polycystic renal disease (three patients) and end-stage renal disease (seven patients). Lesion diameter ranged from 0.7 to 12 cm (mean, 3.6 cm). Thirty-four lesions were in the right kidney, and 21 were in the left kidney. Thirty-five cysts were superficial, cortical, and organ-deforming; 14 were deep-seated in the cortex and nondeforming: and six were parapelvic or medullary. Histology was available for 30 lesions, which had been surgically removed. Of the histologically examined lesions, 14 were malignant and 16 were benign. The malignant cystic lesions included 12 with cystic renal cell cancer, one (parapelvic) with transitional cell carcinoma, and one with metastatic adenocarcinoma. Eight were necrotic, two were unilocular cystic, and two arose from the cyst wall. The excised benign lesions revealed reactive macrophage infiltration in the cyst wall or adjacent renal parenchyma. Among the 16 surgically removed benign complex cystic lesions, nine were solitary and seven were multiple. One solitary cystic lesion was multiloculated. and histopathologic examination revealed a cystic nephroma. The nine solitary cystic lesions were not associated with underlying conditions, whereas the seven multiple cystic lesions were associated with either endstage renal disease with multiple cysts (n = 4) or autosomal dominant polycystic renal disease (n = 3). Among the 14 malignant cystic lesions, eight were solitary and six were multiple. The eight solitary cystic malignant lesions were not associated with underlying conditions, whereas the six multiple malignant cystic lesions were associated with eitherend-stage renal disease (n = 3)or autosomal dominantpolycystic renal disease (n = 3). Among the 25 cystic lesions for which imaging follow-up was available, five were solitary and 20 were multiple. None of the solitary cysts was associated with underlying conditions, whereas four of the 20 multiple cysts were associated with end-stage renal disease. The distribution of renal cystic lesions and associated conditions is summarized in Table I. Complex Cyst Fluid Thirty-seven lesions contained complex cyst fluid that had signal characteristics other than those of simple fluid (hypointensity on TI-weighted images and hyperintensity on 12-weighted images). The following signal characteristics were observed on TI - and T2- weighted images: hyperintensity on Tl-weighted images and hyperintensity on T2-weighted im AJR:172, June 1999

3 MR Imaging of Complex Renal Cysts ages (ii = 3). isointensity on TI-weighted images and hyperintensity on T2-weighted images ( 7 = 3). hyperintensity on TI-weighted images and isointensity on T2-weighted images (ii = 5). and hyperintensity on TI-weighted images and hypointensity on T2-weighted images (ii = 26). In 16 cystic lesions. two of which were malignant. a layering effect was present. Eight lesions with complex cyst fluid were malignant. whereas 29 were benign. No significant association was seen between complex cyst fluid alone and malignancy (p =.35). Benign cystic lesions had either no perceptible wall (ii = 8) (Fig. 1) or a wall with regular (ii = 19) or irregular (ii = 3) contours. All malignant lesions with complex cyst fluid were associated with intense mural enhancement, seven lesions had irregular enhancing walls, one lesion had a thick regular wall, and four lesions had septa or nodules. Mural Irregularity Sixteen lesions had an irregular wall. Irregularity was associated with variable thickness (ii = 3). a nonspheric mural contour (n = 2). or both (ii = I I). Ten lesions with an irregular cyst wall were malignant, whereas six were benign. A significant association was observed between mural irregularity alone and malignancy (p =.0004). An irregular wall was associated with masses or nodules in four lesions (Figs. 2 and 3) and with septa in four lesions (Fig. 4). In all malignant lesions with an irregular wall. the wall and its associated components enhanced intensely. whereas three of the benign lesions showed intense mural enhancement, and three showed negligible enhancement. Mural Thickness Fourteen lesions. 10 of which were malignant. had a thick wall (>2 mm; range. 2-5 mm). The association between a thick wall and malignancy was significant (p =.0(X)3). Of the 10 malignant lesions with a thick wall. all showed intense enhancement, four had a regular mural contour, and six had an irregular mural contour. A Fig. 1.-Complex renal cyst without perceptible wall in 43-year-old woman. A, Ti-weighted fat-suppressed spoiled gradient-echo MR image (TRITE. 180/4; flip angle, 80#{176}) reveals i-cm cartical cyst with hyperintense signal (large arrow). Note simple renal cyst arising from cortex (small arrow). B, 12-weighted breathing-averaged fat-suppressed turbo spin-echo MR image (4500/90) shows hypointense complex cyst (large arrow). Note that simple renal cyst is hyperintense (small arrow). B Mural Enhancement In 32 lesions. 25 of which underwent surgery and seven of which were followed up. the wall enhanced intensely. Among the surgically removed lesions with an enhancing wall, 14 were malignant and 1 1 were benign. Histopathologic examination of the malignant lesions revealed increased neovascularity in the mural tumor or the wall itself. Histopathologic examination of the benign lesions revealed reactive macrophage infiltration either of the cyst wall or of the adjacent renal parenchyma. Among the followed-up lesions. intense mural enhancement persisted on the repeated examinations. The association between intense mural enhancement and malignancy was significant (p =.0022). Fig. 2.-Renal cell carcinoma arising from cyst wall in 58-year-old woman. A, Ti-weighted spoiled gradient-echo (SGE) MR image (TRITE, 140/4; flip angle, 801 reveals 12-cm cystic lesion arising from left kidney. Cystfluid is hyperintense, and mass arises from cyst wall (arrow). B, T2-weighted fat-suppressed turbo spin-echo MR image (4500/90) reveals tumor to be moderate in signal intensity relative to hyperintense cyst fluid (arrow). C, On gadolinium-enhanced SGE MR image (140/4; flip angle, 80#{176}), central portion of mass shows moderate enhancement (arrow). AJR:172, June

4 The associations between each imaging feature and malignancy are listed in Table 2. A B Combined Complex Features Bosniak class 2 lesions (n = 18) were lesions with complex cyst fluid (n = I 8) that did not have a perceptible wall (ii = 8) or had a regular cyst wall ( i = 10). All these lesions were benign. Bosniak class 3 lesions (n = 3) were lesions that lacked intense mural enhancement and had an irregular wall (n = 3), a regular thick wall (ii = 2), complex cyst fluid (n = I), or calcification (n = 1 ). All were benign. Bosniak class 4 lesions (n = 32) were lesions with intense mural enhancement (ii = 32) associated with complex cyst fluid (n = I 8), enhancing septa (ii = 4), or calcification (n = 1). Fourteen of these were malignant. Among the lesions with intense mural enhancement, 19 had a regular contour and 13 had an irregular contour. Four of those with a regular contour were malignant. and 10 of those with an irregular contour were malignant. The association between combined mural irregularity and intense mural enhancement was significant (p =.0002). The associations between combinations of features and malignancy are listed in Table 3. C ariables Responsible for enal Cyst Complexity and ssoclation with MaIinancy Fig. 3.-Benign renal cyst with enhancing nodule in 36-year-old man. A, Ti-weighted spoiled gradient-echo (SGE) MR image (TRITE, i50/4; flip angle, 801 reveals 3-cm cystic lesion in right kidney, with hypointense cyst fluid and mildly hyperintense nodules (arrows). B, Corresponding 12-weighted half-fourier acquisition single-shot turbo spin-echo MR image (infinite/90) shows hyperintense cyst fluid and mildly hypointense nodules (arrows). C, On gadolinium-enhanced SGE image (150/4; flip angle, 800), nodules enhance (arrows). D, Enhancement of nodule (arrow) is also shown on gadolinium-enhanced fat-suppressed SGE coronal plane image (160/4; flip angle, 800). = benign lesions,i = malignantlesions. Fig. 4.-Benign renal cyst with thick, enhancing wall in 58-year-old man. A, Ti-weighted spoiled gradient-echo (SGE) MR image (TRITE, i40/4; flip angle, 80#{176}) reveals cystic lesion with dependent layering of hyperintense fluid (arrow). B, On corresponding breathing-averaged 12-weighted turbo spin-echo MR image (4500/90), layering fluid appears hypointense (arrow). C, On gadolinium-enhanced fat-suppressed SGE MR image (160/4; flip angle, 80#{176}), cyst wall enhances (arrow) AJR:172, June 1999

5 MR Imaging of Complex Renal Cysts Discussion losniak Classificadon (I] of ystlc Renal Lesions Sean on MR Imaging I swithhistopathologic.,cprrelation Malignant n % 0 -; 43 Most of the malignant cystic lesions in our study had intense mural enhancement, and the incidence of malignancy was higher in lesions with irregularly thickened enhancing walls. Most benign cystic lesions had either no perceptible wall or a perceptible regular wall without enhancement. These findings are consistent with two large series that have described the appearance ofcystic lesions on CT [5, 9]. Varying signal intensities reflecting different blood products have been previously observed within the cystic cavity on MR imaging [4]. In our series, cystic renal lesions with signal intensities other than that of simple fluid and with an imperceptible wall were considered complex and in need of follow-up. Although a varying signal intensity is relatively common in the fluid of renal cysts, only a few (three) lesions with this finding were included in our study because, for cystic lesions that have this finding alone, we generally do not perform either histologic confirmation or imaging follow-up. In no lesion with complex fluid alone was substantial change observed on repeated radiologic examinations. A simple nephrogenic renal cyst may become hemorrhagic as a result of trauma, varices in the cyst wall, bleeding diatheses, or unknown reasons [3]. Simple nephrogenic cysts with a high protein content have signal characteristics similar to those of hemorrhagic cysts and may also show a layering effect, especially when infected [10]. Variations in the signal intensity of cyst fluid can be explained by the changing paramagnetic properties of aging blood products or by changes in their concentration [4, 1 1]. Cyst fluid that contains proteinaceous material (e.g., infected cysts) may have a variable signal intensity based on the concentration of the proteinaceous material [1 2]. Hemorrhage can also occur in cystic neoplasms, with a higher incidence in patients with end-stage renal disease [13, 14]. For the eight cystic renal cell carcinomas in our series, hemorrhage into the cystic cavity was shown by MR imaging and confirmed by histopathology, and a thick, enhancing wall with enhancing mural nodules or masses was observed. Two of these cases were associated with end-stage renal disease. Mural morphology is a determinant of malignancy [2, 3]. An irregular or thick cyst wall according to the Bosniak classification system has been considered one of the most worrisome features of indeterminate cystic renal lesions [2, 3, 15]. Other associated features, such as the presence of mural nodules or intense enhancement of the wall, increase the likelihood of malignancy [2, 3]. Cystic lesions with a thick wall have an increased association with malignancy, and intense enhancement of the wall increases the likelihood of malignancy further [3]. In our patients, none of the lesions with an irregular wall alone, without intense mural enhancement, were malignant. Intense enhancement of the cyst wall is considered a criterion for malignancy I 1, 3]. Reported series of complex cystic lesions examined with CT revealed a high incidence of malignancy in cystic lesions with intense mural enhancement, independent of mural morphology [9]. In our series, renal cysts with an enhancing wall were surgically explored in most cases, whereas some patients were followed up if the cystic lesion had a regular thin wall and if the finding of a renal cyst on an abdominal screening examination was considered incidental. Fewer than half the cystic lesions (14/32, or 44%) with intense mural enhancement as an independent feature were associated with malignancy on histopathologic examination. Thick walls may also be observed for infected cysts [3, 15], and six of I 7 benign cysts in our patients exhibited a mactive macrophage infiltration. Cystic lesions with enhancing irregular walls have been shown to have a higher incidence of malignancy [2, 3]. In our series, the occurrence of malignancy in these types of cysts was high, occurring in 10 of 13 lesions. Cysts with a reactive macrophage infiltration were the most common benign lesions to have this appearance. Lesions with an enhancing mural nodule or mass were malignant in our series, except for one benign cyst with a mural nodule. Renal cell carcinomas can arise from the wall of a cyst [16, 17]. The growth patterns of cystic renal cell carcinomas are of four types: the multilocular cystic type, the cystic type, the unilocular cystic type, the cystic necrotic type, and the type arising within a simple cyst I 15-18]. In general, the cystic necrotic type is the most frequently observed, because larger renal cancers commonly undergo central necrosis I I 8]. Although obvious cancers with central necrosis were not included in our series, cystic necrotic was still the most commonly encountered type of cystic cancer in our study. The main limitation of using MR imaging for evaluating cystic lesions of the kidneys is poor visualization of mural calcification. Evidence of a calcified wall can be depicted on MR imaging by signal void foci or curvilinear structures, but the volume of calcium in our experience must be substantially greater than that needed to observe this finding on CT. Nonetheless, our clinical experience suggests that observation of soft-tissue thickening, irregularity, and enhancement, which are features of tumor, is more important than observation of calcification. In fact, the lack of interference from calcium in the visualization of soft tissue may be an advantage over CT or sonography [191. Reports from large series indicate that malignant lesions with calcified walls most often show mural masses or nodules on CT and intense mural enhancement, irregularity. or nodules may be better shown on MR images of these calcified cysts. Our data correlated well with the Bosniak classification system. Cystic lesions with complex cyst fluid, thin calcifications, or thin septa are Bosniak class 2 and considered benign [I]. In our series, all lesions with these features were benign. Cystic lesions with a thick or irregular wall, thick septa, and associated nonenhancing mural masses or nodules are Bosniak class 3. According to a large CT Series, this group of lesions has a 50% chance of being malignant [2]. In our series, six lesions, all of which were benign, had the imaging features of class 3 lesions. Our data differed slightly from prior studies in the mcidence of malignancy associated with intense mural enhancement. In prior studies, the mcidence of malignancy in this group of lesions has been high (90-100%) [2, 15]. In our series, malignancy occurred in 21% of cystic lesions with uniformly thickened enhancing walls and in 77% with irregularly thickened enhancing walls. Our findings suggest that coexistence of mural irregularity and intense mural enhancement is an important feature in the diagnosis of malignant lesions. In summary, the MR features of complex cystic lesions correlated with histopathologic findings and the findings of follow-up by CT or MR imaging. Cystic renal lesions with an AJR:172, June

6 Balci et al. enhancing, irregular wall had the highest association with malignancy, and such lesions should probably be surgically explored or excised. Cysts with a reactive macrophage infil- Irate were the most common benign lesions to have an appearance that simulates malignancy. Our results show that MR findings permit excellent characterization of complex cystic renal lesions as benign or malignant, but overlap between these entities still exists. References 1. Bosniak M. The current radiological approach to renal cysts. Radiology 1986:158: Bosniak M. Diagnosis and management of patients with complicated cystic lesions of the kidney. AiR 1997;169: DavidsonAi, Hartman DS, Choyke PL, Wagner BJ. Radiologic assessment of renal masses: imptications for patient care. Radiology 1997;202: Marotti M, Hricak H, Fritzsche P, Crooks LE, Hedgcock MW, Tanagho EA. Complex and simpie renal cysts: comparative evaluation with MR imaging. Radiology 1987;l62: Bosniak M. Difficulties in classifying cystic lesions of the kidney. Urn! Radio! 1991:13: Rofsky NM, Weinreb JC, Bosniak MA, Libes RB, Bimbaum BA. Renal lesion characterization with gadolinium-enhanced MR imaging: efficacy and safety in patients with renal insufficiency. Radio!ogy 1991;80: Semelka RC, Shoenut J1 Kroeker NA, MacMahon RG, Greenbei HM. Renal lesions: controlled com- _son between CT and 1.5 T MR imaging with nonenhanced and gadolinium-enhanced fat-suppressed spin-echo and breath-hold FLASH techniques. Radio!ogy 1992;182: Eilenberg SS, Lee JKT, Brown JJ, Mirowitz SA, Tartar VM. Renal masses: evaluation with gradient-echo Gd-DVtA-enhanced dynamic MR imaging. Radiology 1990:176: Siegel CL, McFarland EG, Brink JA, Fisher AJ, Humphrey P. Heiken JP. CT of cystic renal masses: analysis of diagnostic performance and interobserver variation. AiR 1997;l69:8l3-8l8 10. Frishman E, Orron DE, Heiman 7., Kessler A, Kaver I, Graif M. Infected renal cysts: sonographic diagnosis and management. J Ultrasound Med 1994;13:7-l0 II. Gomori JM, Grossman RI, Hackney DB, Goldberg HI, Zimmerman RA, Bilanuk LI. Variable appearance of subacute intracranial hematomas on high fieldecho MR.AJNR 1987:8: Fishman MC, Pollack HM, Arger PH, Banner MP. High protein content: another cause of CT hyperdense benign renal cysts. J Compu: Assist Tomogr 1983:7: John G, Semelka RC, Burdeny DA, et al. Renal cell cancer incidence of hemorrhage on MR images in patients with chronic renal insufficiency. J Magn Reson Imaging 1997:7: Mates AJ, Simmons Ri, Kjellstrand CM, et al. Increased incidence of malignancy during chronic renal failure. Lancet 1975:1: Levine E, Hartman DS, Meilstrup JW. Van Slyke MA, Edgar KA, Barth JC. Current concepts and controversies in imaging of renal cystic diseases. Urn! Clin North Am 1997:24: Yamashita Y, Watanabe 0, Miyazaki I, Yamomoto H, Harada M, Takahashi M. Cystic renal cell carcinoma: imaging findings and pathologic correlation. Acta Radio! 1994:35: Hartman DS, Weatherby E III, Laskin WB. Brody JM, Corse W, Baluch JD. Cystic renal cell carcinoma: CT findings simulating a benign hyperdense cyst. AJR 1992; 159: Hartman DS, Davis C. Johns JT, Woolner LB. Cystic renal cell carcinoma. Urn!ogv 1986:13: Semelka RC, Hricak H, Stevens SK, Fingold R, Tomei E, Caroll PR. Combined gadolinium-enhanced and fat saturation MR imaging of renal masses. Radio!ogs 1991:178: Amis ES Jr. Cronan JJ, Yoder IC, Pfister RC, Newhouse JH. Renal cysts: curios and caveats. Urn! Radio! 1982;4: Patterson J, Lohr D, Briscoe C, Briscoe G, Flanigan RC. Calcified renal masses. Umkig 1987:29: Hartman DS, Aronson S, Frazer H. Current status of imaging indeterminate renal masses. Radio! C!in North A,n 1991;29: AJR:172, June 1999