Differentiation of Subtypes of Renal Cell Carcinoma on Helical CT Scans

Transcription

1 Jeong Kon Kim 1 Tae Kyoung Kim 1 Han Jong hn 2 hung Soo Kim 2 Kyu-Rae Kim 3 Kyoung-Sik ho 1 Received September 21, 2001; accepted after revision December 28, Department of Radiology, san Medical enter, University of Ulsan, Poongnap-dong, Songpa-gu, Seoul, , South Korea. ddress correspondence to K.-S. ho. 2 Department of Urology, san Medical enter, University of Ulsan, Songpa-gu, Seoul, , South Korea. 3 Department of Pathology, san Medical enter, University of Ulsan, Songpa-gu, Seoul, , South Korea. JR 2002;178: X/02/ merican Roentgen Ray Society Differentiation of Subtypes of Renal ell arcinoma on Helical T Scans OJETIVE. The purpose of our study was to differentiate subtypes of renal cell carcinoma on helical T scans. MTERILS ND METHODS. We reviewed T scans of four subtypes of renal cell carcinoma: 76 conventional (clear cell), 19 papillary, 13 chromophobe, and two collecting duct. iphasic T scans (unenhanced, corticomedullary, and excretory phase scans) were obtained in 61 patients, and monophasic T scans (unenhanced and excretory phase scans) in 49. We compared patient age and sex; tumor size; degree and pattern (homogeneous, heterogeneous, predominantly peripheral) of enhancement; presence or absence of calcification; and tumor-spreading patterns including perinephric change, venous invasion, and lymphadenopathy in four subtypes. RESULTS. onventional renal carcinoma showed stronger enhancement than the other subtypes (p < 0.05): 106 ± 48 H (mean ± SD) in the corticomedullary phase and 62 ± 25 H in the excretory phase. The sensitivity and specificity for differentiating conventional renal carcinoma from the other subtypes were 74% and 100% when 84 H was used as the cutoff value in the corticomedullary phase and 84% and 91% when 44 H was used as the cutoff value in the excretory phase. onventional (84%), papillary (74%), and collecting duct (100%) renal carcinomas tended to show heterogeneous or predominantly peripheral enhancement, whereas chromophobe renal carcinoma (69%) usually showed homogeneous enhancement. alcification was more common in papillary (32%) and chromophobe (38%) renal carcinomas than in conventional renal carcinoma (11%) (p < 0.05). Perinephric change and venous invasion were not noted in chromophobe renal carcinoma, whereas both were common in collecting duct renal carcinoma. ONLUSION. For the differentiation of the subtypes of renal cell carcinoma, degree of enhancement is the most valuable parameter; enhancement pattern, the presence or absence of calcification, and tumor-spreading patterns can serve supplemental roles in the identification of the subtype of renal cell carcinoma. T he classification of renal cell carcinoma into subtypes has become of interest because of the association with prognosis. Investigators have also concluded that renal cell carcinoma is not a single disease but, rather, a group of several disease entities [1, 2]. ccording to the First International Workshop on Renal ell arcinoma held by the World Health Organization, renal cell carcinoma can be classified into conventional (i.e., clear cell) renal carcinoma, papillary renal carcinoma, chromophobe renal carcinoma, collecting duct renal carcinoma, and unclassified renal carcinoma [3, 4]. onventional renal carcinoma is the most common subtype, accounting for approximately 70% of renal cell carcinomas; the overall 5-year survival rate of patients with conventional renal carcinoma ranges from 55% to 60% [4 6]. Papillary renal carcinoma, the second most common subtype, comprises from 15% to 20% of renal cell carcinomas and is associated with a high 5-year survival rate (80 90%) [6 11]. hromophobe renal carcinoma accounts for 6 11% of the cases. Patients with this subtype have the best prognosis: the 5-year survival rate is approximately 90% [12 16]. ollecting duct carcinoma is a rare subtype, accounting for less than 1% of all renal cell carcinomas; however, patients with collecting duct carcinoma have the worst prognosis, with a 5-year survival rate of less than 5% [5]. The behavior of renal cell carcinoma apparently depends on its subtype, and, accordingly pre- JR:178, June

2 Kim et al. cise prediction of the subtype may be helpful for treatment planning, such as determining the degree of preoperative evaluation and the extent of surgery. T has been widely used for the evaluation of renal cell carcinoma because it can provide detailed information about the tumor itself and whether it has extended into perinephric fat or the renal vein. Furthermore, with the use of helical T, it is possible to analyze the dynamic enhancement pattern of the tumor [17]. lthough many reports about the T features of renal cell carcinoma have been published, to our knowledge only two previous studies have attempted to differentiate the subtypes of renal cell carcinoma [18, 19]. In both reports, researchers stated that stronger enhancement was a unique T finding for the differentiation of conventional renal carcinoma from the other subtypes [18, 19]. However, further investigation appears to be necessary because in previous studies the current World Health Organization classification system [3, 4] of renal cell carcinoma was not used, various T findings were not completely evaluated, and differentiation among the subtypes of nonconventional renal carcinomas was not attempted. In the current study, we compared contrast-enhanced helical T findings in four subtypes of renal cell carcinoma and investigated which T findings are helpful in differentiating among these subtypes. Materials and Methods Patients computerized search of our institution s medical records dated between January 1997 and February 2000 generated a list of 125 patients who had undergone nephrectomy for renal cell carcinoma. Of these 125 patients, the diagnoses for 23 patients with a pathologic diagnosis of granular renal carcinoma (n = 12) and sarcomatoid renal carcinoma (n = 11) were reclassified into a subtype according to the current classification system established by the World Health Organization [3, 4]. Fifteen patients were excluded because they did not undergo helical T. Therefore, a total of 110 patients (78 men and 32 women; age range, years; mean age, 56 years), each of whom had one tumor, were included in this study. These 110 patients consisted of 76 patients with conventional renal carcinoma, 19 with papillary renal carcinoma, 13 with chromophobe renal carcinoma, and two with collecting duct renal carcinoma. T Examination ll T examinations were performed using a helical T scanner (Somatom Plus-S; Siemens, Erlangen, Germany). T scans were obtained during patient breath-holding with the following parameters for imaging acquisition and reconstruction: 120 kvp; 210 m; section collimation, 5 mm; table feed, 7 mm/sec; and reconstruction interval, 5 mm. ll patients received ml of oral contrast material (E-Z T [2% barium sulfate suspension]; E-Z EM, Westbury, NY) 30 min before T and 120 ml of IV contrast material (Iopamiro 300 [iopamidol]; racco, Milano, Italy); the IV contrast material was injected into an antecubital vein using a mechanical injector at a rate of 3.0 ml/sec. Sixty-one patients (43 patients with conventional renal carcinoma, 10 with papillary renal carcinoma, seven with chromophobe renal carcinoma, and one with collecting duct renal carcinoma) underwent biphasic T, which included unenhanced, corticomedullary, and excretory phase scanning. The other 49 patients underwent monophasic T, which included unenhanced and excretory phase scanning. Scanning of the entire kidney was performed in every phase during approximately sec of patient breathholding. Immediately after the scanning for the excretory phase, scanning covering the lower abdomen and pelvis was performed. Scanning for the corticomedullary phase was started 30 sec after IV contrast injection, and scanning for the excretory phase was started 120 sec after contrast injection. Image nalysis Two experienced genitourinary radiologists who were not aware of the renal cell carcinoma subtype reviewed the T scans at a picture archiving and communication system (Radpia; Hyundai Information & Technology, Seoul, Korea), which made it possible to measure the length and T number of a lesion in a particular region of interest. In four subtypes of renal cell carcinoma, we compared the patient age and sex; tumor size; degree and pattern (homogeneous, heterogeneous, predominantly peripheral) of enhancement; the presence or absence of calcification; and the tumor-spreading patterns including perinephric change, venous invasion, and lymphadenopathy. Fig year-old man with conventional (clear cell) renal carcinoma., Unenhanced T scan obtained at level of renal hilum shows well-demarcated round mass (arrows) in right kidney. Diameter of mass is 2.5 cm, and attenuation value is 34 H. and, On contrast-enhanced T scans obtained at same level as, attenuation value of lesion (arrows) was measured as 173 H in corticomedullary phase () and 76 H in excretory phase (). Therefore, degree of enhancement was calculated as 139 H in corticomedullary phase and 42 H in excretory phase. This lesion shows heterogeneous enhancement JR:178, June 2002

3 Helical T of Renal ell arcinoma Fig year-old man with conventional (clear cell) renal carcinoma., Unenhanced T scan obtained at level of superior mesenteric artery shows lobulated mass (arrowheads) with 8-cm diameter and 35-H attenuation value. and, On contrast-enhanced T scans, enhancement pattern of lesion (arrowheads) is heterogeneous, and attenuation value was measured as 201 H in corticomedullary phase (degree of enhancement, 166 H) () and 128 H in excretory phase (degree of enhancement, 93 H) (). Perinephric change (arrows) is visible. To evaluate the degree of enhancement of a tumor, the first radiologist measured the attenuation of three separate regions of interest and calculated the mean of these three values. The locations of the regions of interest were decided by consensus of the two radiologists. One location for measuring the attenuation value was the solid enhancing area in the excretory phase. round or elliptic region-of-interest cursor was placed over an enhanced area, which was at least 1 cm2 and consistent in location during all phases of T. We tried to cover the enhanced area as much as possible in the region of interest and to exclude the area of calcification from the region of interest. The degree of enhancement was measured by calculating the difference in the mean attenuation values between the corticomedullary and unenhanced phase scans (n = 61) and between the excretory and unenhanced phase scans (n = 110). The enhancement pattern of the tumor was classified as homogeneous, heterogeneous, or predominantly peripheral. Homogenous enhancement was indicated when most areas in the tumor showed a uniform degree of enhancement. Predominantly peripheral enhancement was considered when most portions of the tumor were not enhanced and only the peripheral rim or septa showed enhancement. The remaining cases were considered to have heterogeneous enhancement. The enhancement pattern of a tumor is generally affected by its size; the larger a tumor grows, the more frequently intratumoral necrosis or hemorrhage occurs [20]. To eliminate the influence of tumor size on enhancement pattern, we divided tumors into the following groups according to maximum diameter: less than 3 cm, 3 7 cm, and greater than 7 cm. Thereafter, we compared the en- hancement pattern of the four subtypes of renal cell carcinoma in the three groups. Perinephric change was indicated when there was evidence of strands of soft-tissue attenuation or parasitized vessels in the perinephric area and thickening of Gerota s fascia. Venous invasion was indicated when the lumen of the renal vein or the inferior vena cava was replaced by the tumor. Lymphadenopathy was considered to be present when a lymph node was enlarged more than 1 cm in diameter. Statistical nalysis Statistical analysis was performed for comparison of the various T findings of conventional renal carcinoma, papillary renal carcinoma, and chromophobe renal carcinoma. ollecting duct renal carcinoma was excluded from the statistical analysis because the patient population was too small. We applied the one- Fig year-old man with papillary renal carcinoma., Unenhanced T scan obtained at level of renal hilum shows well-demarcated round mass (arrows) with 2.5-cm diameter and 38-H attenuation value in right kidney. and, ontrast-enhanced T scans obtained at same level as show homogeneous enhancement of lesion (arrows); attenuation value was measured as 82 H in corticomedullary phase () and 75 H in excretory phase (). Therefore, degree of enhancement was calculated as 44 H in corticomedullary phase and 37 H in excretory phase. JR:178, June

4 way analysis of variance test for comparison of patient age, tumor size, and the degree of tumor enhancement. To evaluate the diagnostic validity of the degree of enhancement in differentiating subtypes of renal cell carcinoma, we generated receiver operating characteristic curves and analyzed these curves to determine the cutoff value for differentiating the subtypes of renal cell carcinoma with highest accuracy. The chi-square test was used for comparison of the distribution of patient sex; the frequency of each enhancement pattern; the presence or absence of calcification; and tumor-spreading patterns including perinephric change, venous invasion, and lymphadenopathy in three subtypes of renal cell carcinoma. Results The mean patient age ± the standard deviation was 59 ± 11 years (range, years) in patients with conventional renal carcinoma, 52 ± 13 years (range, years) in Kim et al. TLE 1 ttenuation and Degree of ontrast Enhancement in Regions of Interest of Four Subtypes of Renal ell arcinoma Renal ell arcinoma Subtype Helical T Phase Unenhanced ttenuation orticomedullary ttenuation Enhancement b Excretory phase ttenuation Enhancement c onventional a Papillary hromophobe ollecting Duct 35 ± 9 37 ± 5 39 ± ± ± ± ± ± ± ± ± ± ± ± 9 67 ± 6 28 ± ± 3 30 ± 6 Note. ll data are means ± SD and are measured in Hounsfield units (H). a ttenuation value and degree of enhancement for conventional renal cell carcinoma are higher than for papillary carcinoma and chromophobe carcinoma in both the corticomedullary and excretory phases ( p < 0.05). b Difference in the attenuation value between the corticomedullary and unenhanced phases. c Difference in attenuation value between the excretory and unenhanced phases. Fig year-old woman with papillary renal carcinoma., Unenhanced T scan obtained at level of superior mesenteric artery shows lobulated mass (arrowheads) with 7-cm diameter and 37-H attenuation value. alcification (arrow) within tumor can be seen. and, On contrast-enhanced T scans, enhancement pattern of lesion (arrowheads) is predominantly peripheral, and attenuation value was measured as 94 H and degree of enhancement as 57 H in corticomedullary phase () and 71 H and 34 H, respectively, in excretory phase (). Fig year-old man with chromophobe renal carcinoma., Unenhanced T scan obtained at level of upper pole of right kidney shows lobulated mass (arrowheads) with 8-cm diameter and 37-H attenuation value. Multifocal calcifications (arrows) can be seen within tumor. and, On contrast-enhanced T scans obtained at same level as, mass (arrowheads) shows homogeneous enhancement that measured 51 H in corticomedullary phase () and 33 H in excretory phase (). ttenuation value was 88 H in corticomedullary phase and 70 H in excretory phase JR:178, June 2002

5 Helical T of Renal ell arcinoma those with papillary renal carcinoma, 51 ± 11 years (range, years) in those with chromophobe renal carcinoma, and 60 ± 6 years (range, years) in those with collecting duct renal carcinoma. The male-tofemale ratio was 5.9:1 in conventional renal carcinoma, 2.8:1 in papillary renal carcinoma, 2.3:1 in chromophobe renal carcinoma, and 1:1 in collecting duct renal carcinoma. Demographic distributions were not significantly different among conventional, papillary, and chromophobe renal carcinomas (p > 0.05). The mean tumor diameter ± the standard deviation was 5.7 ± 2.5 cm (range, 1 14 cm) in conventional renal carcinoma, 6.4 ± 4.2 cm (range, 1 14 cm) in papillary renal carcinoma, 7.2 ± 2.7 cm (range, 2 13 cm) in chromophobe renal carcinoma, and 10.9 ± 0.6 cm (range, cm) in collecting duct renal carcinoma. The tumor size was not significantly different in conventional, papillary, and chromophobe renal carcinomas (p > 0.05). The attenuation values and the degree of enhancement of the four subtypes of renal cell carcinoma are summarized in Table 1. The attenuation values of conventional, papillary, and chromophobe renal carcinomas did not differ on unenhanced scans (p > 0.05) (Figs. 1, 2, 3, 4, and 5). On the corticomedullary phase scans, both the attenuation value and degree of enhancement were higher in conventional renal carcinoma (Figs. 1 and 2) than in papillary renal carcinoma (p = for attenuation value and p = for degree of enhancement) (Figs. 3 and 4) and in chromophobe renal carcinoma (p = for attenuation value and p = for degree of enhancement) (Fig. 5). On the excretory phase scans, both attenuation value and degree of enhancement were also higher in conventional renal carcinoma (Figs. 1 and 2) than in papillary renal carcinoma (p = for attenuation value and p = for degree of enhancement) (Figs. 3 and 4) and in chromophobe renal carcinoma (p = for attenuation value and p = for degree of enhancement) (Fig. 5). No statistically significant difference between papillary and chromophobe renal carcinomas was detected in attenuation value or degree of enhancement in either the corticomedullary phase or the excretory phase (p > 0.05). lthough statistical analysis could not be performed, both the attenuation value and degree of enhancement of collecting duct renal carcinoma were similar to those of papillary and chromophobe renal carcinomas. Fig. 6. Graphs show receiver operating characteristics curves for degree of enhancement in differentiation of conventional (clear cell) renal carcinoma from nonconventional renal carcinomas (i.e., papillary, chromophobe, and collecting duct)., In corticomedullary phase, area under curve ( z value) is (95% confidence interval [I]: )., In excretory phase, z value is (95% I: ). TLE 2 Enhancement of the onventional Subtype Versus Nonconventional Subtypes of Renal ell arcinoma orticomedullary Phase (n = 61) Excretory Phase (n = 110) Subtype of Renal ell arcinoma >84 H 84 H >44 H 44 H onventional renal carcinomas (n) Nonconventional renal carcinomas (n) Variable Statistical nalyses of Enhancement Data (%) Sensitivity Specificity Positive predictive value Negative predictive value ccuracy Note. onventional renal carcinoma refers to the clear cell subtype of renal cell carcinoma. Nonconventional renal carcinoma includes papillary, chromophobe, and collecting duct carcinomas, three subtypes of renal cell carcinoma defined by the World Health Organization classification system [3, 4]. TLE 3 Enhancement Pattern Enhancement Patterns of Tumor in Four Subtypes of Renal ell arcinoma Renal ell arcinoma Subtype onventional Papillary hromophobe ollecting Duct No. % No. % No. % No. % Homogeneous a Heterogeneous b Predominantly peripheral Total a Frequency is higher in chromophobe carcinoma of the kidney than in conventional renal carcinomas or papillary renal carcinomas (p < 0.05). b Frequency is higher in conventional renal carcinomas than in chromophobe renal carcinomas ( p = 0.000). JR:178, June

6 Kim et al. D Fig year-old man with collecting duct renal carcinoma., Unenhanced T scan obtained at level of upper pole of right kidney shows lobulated mass (arrowheads) with 12-cm diameter and 37-H attenuation value. Perinephric change (arrows) can be seen. and, On contrast-enhanced T scans obtained at same level as, mass (arrowheads) shows heterogeneous enhancement with 43-H enhancement in corticomedullary phase () and 35-H enhancement in excretory phase (). ttenuation value was 80 H in corticomedullary phase and 72 H in excretory phase. Perinephric change (arrows) is shown. D, T scan obtained in excretory phase 8 cm above shows tumor invasion (arrows) of renal vein. The receiver operating characteristic curves for the degree of tumor enhancement in differentiating conventional renal carcinoma from nonconventional renal carcinomas are illustrated in Figure 6. The area under the curve (z value) for the corticomedullary phase was (95% confidence interval [I]: ) and that for the excretory phase was (95% I: ). In the receiver operating characteristic curve analysis, the cutoff value with the highest accuracy for the differentiation of conventional renal carcinoma from nonconventional renal carcinomas was 84 H in the corticomedullary phase and 44 H in the excretory phase. The diagnostic test statistics for those values are listed in Table The enhancement patterns of the four subtypes of renal cell carcinoma are summarized in Table 3. Heterogeneous or predominantly peripheral enhancement was frequent in conventional (84%), papillary (74%), and collecting duct (100%) renal carcinomas, whereas homogeneous enhancement was commonly noted in chromophobe renal carcinoma (69%). Statistically, the frequency of heterogeneous enhancement was significantly higher in conventional renal carcinoma (68%) than in chromophobe renal carcinoma (23%) (p = 0.000), whereas the frequency of homogeneous enhancement was higher in chromophobe renal carcinoma than in conventional renal carcinoma (p = 0.000) and papillary (p = 0.016) renal carcinomas. The frequency of predominantly peripheral enhance- ment was similar in conventional, papillary, and chromophobe renal carcinomas (p > 0.05). n analysis of the tumor enhancement patterns according to tumor size is summarized in Table 4. In tumors less than 3 cm in diameter, a heterogeneous or predominantly peripheral enhancement pattern was noted in 10 patients (56%) with conventional renal carcinoma (Fig. 1), whereas four of five papillary renal carcinomas (Fig. 3) and all chromophobe renal carcinomas showed homogeneous enhancement. In tumors 3 7 cm in diameter, conventional (93%) and papillary (88%) renal carcinomas showed predominantly heterogeneous or predominantly peripheral enhancement pattern, whereas chromophobe renal carcinomas (67%) usually showed homogeneous enhancement. In tumors greater than 7 cm in diameter, the frequency of homogeneous enhancement was higher in chromophobe renal carcinoma (63%) (Fig. 5) than in conventional renal carcinoma (6%) (p = 0.000) (Fig. 2) and papillary renal carcinoma (0%) (p = 0.029) (Fig. 4). t microscopic examination, all tumors with homogeneous enhancement were chiefly composed of solid elements, whereas all tumors with predominantly peripheral enhancement had extensive necrosis or hemorrhage. ll tumors with heterogeneous enhancement had both solid elements and necrosis or hemorrhage. alcification within the tumor was noted in eight patients with conventional renal carcinoma (11%), six with papillary renal carcinoma (32%), and five with chromophobe renal carcinoma (38%). The frequency was higher in papillary (Fig. 4) and chromophobe (Fig. 5) renal carcinomas than in conventional renal carcinoma ( p = for papillary renal carcinoma and p = for chromophobe renal carcinoma). Neither of the collecting duct carcinomas (n = 2) had calcifications (Fig. 7). In evaluating the tumor-spreading patterns, we observed perinephric change in 13 conventional (17%), four papillary (21%), and two collecting duct (100%) renal carcinomas. Venous invasion was noted in six conventional (8%) and two papillary (11%) renal carcinomas and in one collecting duct (50%) renal carcinoma; all of these cases were also confirmed at pathologic examination. Lymphadenopathy was noted in nine conventional (12%), two papillary (11%), and two chromophobe (15%) renal carcinomas and in one collecting duct (50%) renal carcinoma. In patients with chromophobe renal carcinoma, neither perinephric change nor venous invasion was observed. Statistical analysis revealed that the frequencies of perinephric change, venous JR:178, June 2002

7 Helical T of Renal ell arcinoma TLE 4 Enhancement Pattern Enhancement Pattern of Tumors ccording to Size in Four Subtypes of Renal ell arcinoma Size of Tumor (cm) onventional (n = 76) Note. Frequency of homogeneous enhancement is higher in chromophobe subtype than in conventional and papillary subtypes (p < 0.05). invasion, and lymphadenopathy in conventional, papillary, and chromophobe renal carcinomas did not differ significantly (p > 0.05). The T stage of renal cell carcinoma according to the merican Joint ommittee on ancer s system [21] was I (T1 N0 M0) in 58 patients (42 conventional, 11 papillary, and five chromophobe renal carcinomas), II (T2 N0 M0) in 30 (19 conventional, four papillary, and seven chromophobe renal carcinomas), III (T3a,b N0 1 M0) in 18 (13 conventional and four papillary renal carcinomas and one chromophobe renal carcinoma), and IV (T1 3b N2 M0) in four (two conventional and two collecting duct renal carcinomas). Discussion The classification of renal cell carcinoma is based mainly on the microscopic appearance of the tumor and genetic abnormalities. Each subtype is associated with a different prognosis and tumor behavior [2]. Patients with papillary renal carcinoma or with chromophobe renal carcinoma have a higher 5-year survival rate than those with conventional renal carcinoma of the same stage [2, 5]. To our knowledge, until recently, differentiation of the renal cell carcinoma subtypes has been attempted on T in only two series. Fujimoto et al. [18] analyzed the enhancement pattern of renal cell carcinomas greater than 5 cm in diameter on contrast-enhanced helical T. They reported that strong enhancement equal to the renal cortex was noted only in Renal ell arcinoma Subtype Papillary (n = 19) hromophobe (n = 13) ollecting Duct (n = 2) Total (n = 110) Homogeneous Heterogeneous <3 Predominantly peripheral Total Homogeneous Heterogeneous Predominantly peripheral Total Homogeneous Heterogeneous >7 Predominantly peripheral Total conventional renal carcinoma (75%) and not in the other subtypes of renal cell carcinoma. Wildberger et al. [19] evaluated T findings including tumor nodule, margin, shape, and enhancement to differentiate among the various subtypes of renal cell carcinoma. ccording to their data, Wildberger et al. found that the sensitivity for differentiating conventional renal carcinoma from nonconventional renal carcinomas was approximately 72%. However, in both of these studies, the subtypes of nonconventional renal carcinomas could not be predicted with the T criteria being evaluated. In our study, we found that enhancement pattern is the most useful parameter in differentiating subtypes of renal cell carcinoma, especially conventional renal carcinoma versus nonconventional renal carcinomas, with high validity ( z value > 0.8 in both the corticomedullary phase and the excretory phase). onventional renal carcinoma showed stronger enhancement than nonconventional renal carcinomas in both the corticomedullary and excretory phases, and the tumors that enhanced more than approximately 84 H in the corticomedullary phase and 44 H in the excretory phase were likely to be conventional renal carcinoma. lthough strong enhancement of conventional renal carcinoma has been observed in previous reports [18, 19], the actual values of enhancement for differentiating conventional renal carcinoma from nonconventional renal carcinomas have not, to our knowledge, been reported previously; our study is the first to present these data. Some investigators believe that the strong enhancement of conventional renal carcinoma is caused by its rich vascular network and alveolar architecture at histologic examination [2, 18]. Our study revealed that enhancement pattern, the presence or absence of calcification, and tumor-spreading pattern overlap among the subtypes of renal cell carcinoma. However, in certain circumstances, a combination of these characteristics may be helpful in predicting a specific subtype of renal cell carcinoma. tumor that is greater than 7 cm in diameter, exhibits homogeneously weak enhancement, and has calcifications is strongly suggestive of chromophobe renal carcinoma. In addition, a tumor with heterogeneously strong enhancement and a diameter of less than 3 cm may indicate conventional renal carcinoma. Different enhancement patterns of renal cell carcinoma according to subtype may be supported by findings at pathologic examination. hromophobe renal carcinoma, which tends to exhibit homogeneous enhancement on T, shows a homogenous cut surface without hemorrhage or necrosis and a solid growth pattern at pathologic examination. In contrast, conventional renal carcinoma, which usually shows heterogeneous or predominantly peripheral enhancement on T, commonly has hemorrhage or necrosis within the tumor at pathologic examination [2, 22]. With regard to the tumor-spreading pattern, chromophobe renal carcinoma did not show any perinephric change or venous invasion. This finding suggests that chromophobe renal carcinoma grows slowly and less aggressively, as has also been shown by other researchers [14, 15]. rotty et al. [14] reported that 43 (86%) of 50 chromophobe renal carcinomas were discovered in Robson stage I at presentation, and khtar et al. [15] reported that 20 of the 21 chromophobe renal carcinomas in their study were T1 2 N0 M0. In fact, in our study the overall frequencies of perinephric change, venous invasion, and lymphadenopathy were lower than those documented in previous reports [1, 5]. This difference in findings might be attributable to the fact that the recently increased use of T has resulted in earlier detection of renal cell carcinoma than in the past and that the use of contrast-enhanced helical T has been beneficial for differentiating malignant tumors from unenhanced benign cysts. Some T findings are known to be relevant to the prognosis of patients with renal cell carcinoma: hemorrhage and necrosis are predictors of poor prognosis, and calcification suggests a higher 5-year survival rate [1]. Our JR:178, June

8 Kim et al. results agree with these findings because hemorrhage and necrosis (heterogeneous or predominantly peripheral enhancement pattern) were more common in conventional and collecting duct renal carcinomas, both of which are associated with a poor prognosis, than in papillary and chromophobe renal carcinomas, which are associated with a good prognosis. alcification was more frequently seen in papillary and chromophobe renal carcinomas, which are associated with a better prognosis, than in the other subtypes. For a preoperative staging workup of a patient with renal cell carcinoma, examinations in addition to abdominal T such as chest radiography and bone scanning are necessary for metastasis survey. The prediction of the subtype of renal cell carcinoma may influence the degree of preoperative evaluation. For example, a patient with a subtype of renal cell carcinoma that tends to not metastasize, such as chromophobe renal carcinoma, may not need to undergo a complex metastasis survey. For patients with surgically treatable renal cell carcinoma, radical nephrectomy has been indicated as the standard treatment. In spite of improved surgical techniques, postoperative mortality and morbidity remain high. From this point of view, precise preoperative identification of the subtype of renal cell carcinoma may be helpful in determining the appropriate extent of surgery. For example, an unnecessarily wide resection may be avoided in patients with a subtype that is unlikely to recur or metastasize, thereby reducing postoperative morbidity and mortality. There are some limitations in this study. The major limitation is that the number of papillary, chromophobe, and collecting duct renal carcinomas was relatively small to analyze T features. In fact, this limitation arises from the significantly lower incidence of these three subtypes of renal cell carcinoma compared with conventional renal carcinoma. Therefore, further investigation with adequate numbers of these three subtypes will be necessary in the future. One possible criticism of this study is that enhancement of renal cell carcinoma may be variable according to the contrast injection parameter and scan delay time. In a study by irnbaum et al. [23], in which IV contrast material was administrated using a biphasic technique (injection rate of 3 ml/sec [50 ml] and 1 ml/sec [100 ml] for a total injection time of approximately 117 sec), the degree of enhancement was stronger in the excretory phase than in the corticomedullary phase; these findings dramatically differ from our results. Therefore, our criterion of tumor enhancement pattern is applicable only to those cases in which the contrast injection protocol and scan time are similar to those of our study. nother minor limitation of this study is that all the patients did not undergo biphasic T. However, the scanning protocol for the excretory phase was strictly constant in all patients, and the different degrees of enhancement among the subtypes of renal cell carcinoma were noted in the excretory phases as well as in the corticomedullary phase. In conclusion, the degree of enhancement is the most valuable parameter in differentiating among the subtypes of renal cell carcinoma, and enhancement pattern, calcification, and tumor-spreading patterns may play a supplemental role. cknowledgments We thank onnie Hami, of the Department of Radiology, University Hospitals Health System, leveland, OH, for editorial assistance and Wi hang Kang, of the san Medical enter, University of Ulsan, for statistical advice. References 1. Mclennan L, Deyoe L. The imaging evaluation of renal cell carcinoma: diagnosis and staging. Radiol lin North m 1994;32: Reuter VE, Presti J Jr. ontemporary approach to the classification of renal epithelial tumors. Semin Oncol 2000;27: ostwick DG, Eble JN, Murphy GP. onference summary: diagnosis and prognosis of renal cell carcinoma 1997 workshop. ancer 1997;80: ostwick DG, Eble JN. Diagnosis and classification of renal cell carcinoma. Urol lin North m 1999;26: osnib SM. Risk and prognosis in renal neoplasm: a pathologist s prospective. Urol lin North m 1999;26: Soyer P, Dufresne, Klein I, arbagelatta M, Herve JM, Scherrer. Renal cell carcinoma of clear type: correlation of T features with tumor size, architectural patterns, and pathologic staging. Eur Radiol 1997;7: min M, orless L, Renshaw, et al. Papillary (chromophil) renal cell carcinoma: histomorphologic characteristics and evaluation of conventional pathologic prognostic parameters in 62 cases. m J Surg Pathol 1997;21: Delahunt, Eble JN. Papillary renal cell carcinoma: a clinicopathologic and immunohistochemical study of 105 cases. Mod Pathol 1997;10: Renshaw, orless L. Papillary renal cell carcinoma: histology and immunohistochemistry. m J Surg Pathol 1995;19: hoyke PL, Walther MM, Glenn GM, et al. Imaging features of hereditary papillary renal cancers. J omput ssist Tomogr 1996;20: Roberts S III, Winick, Santi MR. Residents teaching files: papillary renal cell carcinoma diagnostic dilemma of a cystic renal mass. Radio- Graphics 1997;17: Renshaw, Henske EP, Loughlin KR, et al. ggressive variants of chromophobe renal cell carcinoma. ancer 1996;78: Megumi Y, Nishimura K. hromophobe cell renal carcinoma. Urol Int 1998;61: rotty T, Farrow GM, Lieber MM. hromophobe cell carcinoma: clinicopathologic features of 50 cases. J Urol 1995;154: khtar M, Kadar H, Linjawi T, Mclntock J, li M. hromophobe cell carcinoma of the kidney: a clinicopathologic study of 21 cases. m J Surg Pathol 1995;19: Thoenes W, Storkel S, Rumpelt HJ, Moll R, aum HP, Werner S. hromophobe cell renal carcinoma and its variants: a report on 32 cases. J Pathol 1988;155: Garant M, onaldi VM, Taourel P, Pinsky MF, ret PM. Enhancement patterns of renal masses during multiphase helical T acquisitions. bdom Imaging 1998;23:43l Fujimoto H, Wakao F, Moriyama N, Tobisu K, Sakamoto M, Kakizoe T. lveolar architecture of clear cell renal carcinomas (< or = 5.0 cm) show high attenuation on dynamic T scanning. Jpn J lin Oncol 1999;29: Wildberger JE, dam G, oeckmann W, et al. omputed tomography characterization of renal cell tumors in correlation with histopathology. Invest Radiol 1997;32: Russo P. Renal cell carcinoma: presentation, staging, and surgical treatment. Semin Oncol 2000;27: Fleming ID, ooper JS, Henson DE, et al., eds. ancer staging manual, 5th ed. Philadelphia: Lippincott-Raven, Storkel S, van den erg E. Morphological classification of renal cancer. World J Urol 1995;13: irnbaum, Jacobs JE, Ramchandani P. Multiphasic renal T: comparison of renal mass enhancement during the corticomedullary and nephrographic phases. Radiology 1996;200: JR:178, June 2002