Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2442060927
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Supplemental Figures and Tables
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zhang, J.
Right arrow Articles by Hricak, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zhang, J.
Right arrow Articles by Hricak, H.
(Radiology 2007;244:494-504.)
© RSNA, 2007

Genitourinary Imaging

Solid Renal Cortical Tumors: Differentiation with CT1

Jingbo Zhang, MD, Robert A. Lefkowitz, MD, Nicole M. Ishill, MS, Liang Wang, MD, Chaya S. Moskowitz, PhD, Paul Russo, MD, Halley Eisenberg, BS, and Hedvig Hricak, MD, PhD

1 From the Department of Radiology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, Room C278D, New York, NY 10021. Received May 26, 2006; revision requested July 27; revision received September 19; accepted October 5; final version accepted December 5. Address correspondence to J.Z. (e-mail: zhangj12{at}mskcc.org).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Purpose: To retrospectively determine if solid renal cortical tumors can be differentiated on computed tomographic (CT) images on the basis of their morphologic features and enhancement patterns.

Materials and Methods: Institutional review board approval was obtained and the informed consent requirement was waived for this HIPAA-compliant study. Between January 2004 and September 2005, 193 consecutive patients (age range, 19–95 years; 112 men, 81 women) with renal masses underwent total or partial nephrectomy and preoperative renal CT. Two radiologists retrospectively reviewed CT studies in an independent and blinded fashion. The pattern and degree of enhancement, lesion contour, presence of neovascularity, and calcifications were evaluated. Fisher exact tests, Pearson {chi}2 tests, multivariate logistic regression, and Wilcoxon rank sum tests were performed.

Results: Of the 198 renal tumors (median size, 3.4 cm; range, 1.1–20.0 cm) included in this study, 108 (55%) were clear cell renal cell carcinomas (RCCs); 30 (15%), papillary lesions; 24 (12%), chromophobe adenomas; 14 (7%), oncocytomas; six (3%), lipid-poor angiomyolipomas; and 16 (8%), other or unclassified renal tumors. Clear cell RCC most commonly manifested with a mixed enhancement pattern of both hypervascular soft-tissue components and low-attenuation areas that corresponded to necrotic or cystic changes (reader 1, 88% of clear cell tumors; reader 2, 79% of clear cell tumors). This pattern was highly predictive of clear cell RCC (odds ratio of 22 and 54 for readers 1 and 2, respectively, for comparison with homogeneous pattern), whereas the homogeneous and peripheral enhancing patterns were more predictive of less aggressive papillary and chromophobe lesions. Clear cell RCCs and oncocytomas tended to be hypervascular, chromophobe lesions and angiomyolipomas tended to enhance moderately, and papillary lesions were mostly hypovascular.

Conclusion: Certain imaging features and the degree of enhancement may be helpful in differentiating subtypes of renal cortical tumors.

Supplemental material: http://radiology.rsnajnls.org/cgi/content/full/244/2/494/DC1

© RSNA, 2007


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Renal cortical tumors are a family of neoplasms with distinct cytogenetic characteristics and molecular defects, unique histopathologic features, and varying malignant potential (1–4). Historically, malignant renal parenchymal tumors have been described as a single entity, such as hypernephroma, renal cancer, renal adenocarcinoma, or renal cell carcinoma (RCC). On the basis of conclusions drawn from a workshop (The Impact of Molecular Genetics on the Classification of Renal Cell Tumors), in the 1997 Heidelberg classification, renal cell tumors were subdivided into benign and malignant parenchymal neoplasms, and—when possible—each subcategory was limited to the most commonly documented genetic abnormalities (5). In this classification system, the most common subtypes of RCC include clear cell RCC (also known as common or conventional RCC), papillary RCC, and chromophobe RCC, as well as a small number of other unclassified tumors (5). Benign tumors account for approximately 20% of all renal cortical tumors, and renal oncocytoma is the most common tumor type.

Conventional clear cell RCCs, which account for approximately 65% of renal cortical tumors and 90% of metastases, have the greatest metastatic potential. Papillary and chromophobe RCCs, which account for approximately 25% of renal cortical tumors and 10% of metastases, are associated with less metastatic potential (6). The overall 5-year survival rate for patients with papillary and chromophobe RCCs (80%–90%) is higher than that for patients with conventional RCCs (50%–60%) (7–9). Renal oncocytomas are virtually benign (10). Since clinical implications and therapeutic strategies may differ for different subtypes of renal cortical tumors, preoperative identification of the subtype would be of great clinical interest.

The Bosniak classification system is used to assess the likelihood of malignancy in cystic renal masses on the basis of lesion complexity (11). This classification system has been used to guide clinical management of cystic renal masses. Although it is important to preoperatively differentiate a solid renal tumor for treatment planning and patient counseling, there are no well-established imaging criteria with which to classify the histologic subtypes of these lesions. Thus, assigning a definitive diagnosis on the basis of biopsy findings may be a challenge because it can be difficult to distinguish a chromophobe carcinoma from an oncocytoma or an RCC with sarcomatoid features from the spindle component of an angiomyolipoma (12–14). It has been suggested that certain imaging features may be associated with different subtypes of solid renal cortical tumors (15–20). However, diagnosing renal masses, especially those that are detected incidentally, remains problematic.

Thus, the purpose of our study was to retrospectively determine if solid renal cortical tumors depicted on computed tomographic (CT) images can be differentiated on the basis of their morphologic features and enhancement patterns.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Patients
Our institutional review board approved this study and waived the informed consent requirement. This study complied with the Health Insurance Portability and Accountability Act.

Between January 2004 and September 2005, 193 consecutive patients (age range, 19–95 years; mean age, 61 years; 112 men, 81 women) with renal masses underwent total or partial nephrectomy (excluding nephroureterectomy) and preoperative renal CT scanning. The median time from scanning to surgery was 35 days (range, 2 days to 14 months). Some patients underwent other examinations, including magnetic resonance (MR) imaging and ultrasonography, between initial CT scanning and surgery. Additionally, 11 patients with grossly cystic lesions without solid components were excluded (patients with benign unilocular cysts, n = 4; those with benign multilocular cysts, n = 3; those with multilocular cystic RCCs, n = 3; and those with benign cystic nephroma, n = 1).

Of the 193 patients included in this study, 112 were men (58%) and 81 were women (42%). Of these patients, 188 underwent unilateral nephrectomy and five underwent bilateral nephrectomy for bilateral renal masses. Of the 198 lesions, 141 (71%) were resected with partial nephrectomy and 57 (29%), with total nephrectomy.

CT Examination
All CT examinations were performed with four– or 16–detector row helical scanners (GE Medical Systems, Milwaukee, Wis) and with our standard renal mass protocol tailored to each scanner. CT images were obtained during patient breath holding with the following parameters: 120 kVp, 200–400 mA (depending on patient size), and section thickness and reconstruction interval of 2.5 mm through the kidneys and 5.0 mm through the rest of the abdomen. The pitch used with helical scanners varied for different scanners but ranged from 0.75 to 1.5. All patients received oral contrast material 30 minutes before CT. Unenhanced, parenchymal phase, and excretory phase images were obtained through the kidneys. The entire abdomen (and pelvis, if requested) was scanned during the parenchymal phase only. A 150-mL dose of nonionic intravenous contrast material (iohexol 300, Omnipaque; GE Healthcare, Milwaukee, Wis) was administered with a power injector at a rate of 2.5 mL/sec (or slower if mandated owing to suboptimal venous access). Time delay to scanning varied with the type of scanner used but was determined on the basis of the typical time to the renal parenchymal (70–85 seconds) and excretory (3 minutes) phases. All images were sent to our enterprise-wide picture archiving and communications system (Centricity; GE Healthcare) to be interpreted on workstations.

CT Image Analysis
The CT studies were independently reviewed by two fellowship-trained body radiologists who specialized in genitourinary imaging (J.Z. and R.A.L., 2 and 6 years of experience, respectively, interpreting genitourinary CT studies since completion of their fellowships). Readers were aware that patients were being evaluated for renal lesions, but they were blinded to any other clinical, pathologic, or imaging findings. Before image interpretation, the readers met and agreed on the CT features to be used to characterize renal masses and designed a data collection form.

For each of the 198 kidneys, the radiologists evaluated the largest lesion for several features. The first feature evaluated was pattern of enhancement. The enhancement pattern of the tumor was classified as homogeneous (Fig 1a) or heterogeneous. The heterogeneous tumors contained a mixture of solid enhancing soft-tissue components and low-attenuation areas that might have represented necrotic or cystic changes. Heterogeneous tumors were further categorized into three types on the basis of the relative proportion of solid areas to low-attenuation areas: solid or predominantly solid lesions with small areas of low attenuation (Fig 1b), lesions with mixed solid and low-attenuation areas (Fig 1c), and predominantly low-attenuation lesions with peripheral enhancement (Fig 1d).


Figure 1A
View larger version (177K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a: Transverse contrast material–enhanced CT images obtained during the renal parenchymal phase in four patients. (a) In a 63-year-old man with an incidentally discovered renal mass, surgical pathologic analysis revealed papillary RCC. The mass (arrow) demonstrates a homogeneous, relatively low-degree enhancement pattern (attenuation, 81 HU) in the renal parenchymal phase. (b) In a 50-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a predominantly solid enhancement pattern, with small low-attenuation foci that may represent necrosis or cystic changes. An avid degree of enhancement is present, with maximum attenuation of 239 HU during the renal parenchymal phase, which is nearly identical to that of adjacent renal parenchyma. (c) In a 54-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a mixture of hypervascular and low-attenuation areas. An avid degree of enhancement is present, with maximum attenuation of 209 HU, which is nearly identical to that of the adjacent renal cortex. (d) In a 48-year-old man with a large abdominal mass discovered at physical examination, surgical pathologic analysis revealed a 15-cm papillary RCC. The tumor (arrow) demonstrates a low degree of peripheral enhancement (maximum attenuation, 85 HU) and central low attenuation.

 

Figure 1B
View larger version (175K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b: Transverse contrast material–enhanced CT images obtained during the renal parenchymal phase in four patients. (a) In a 63-year-old man with an incidentally discovered renal mass, surgical pathologic analysis revealed papillary RCC. The mass (arrow) demonstrates a homogeneous, relatively low-degree enhancement pattern (attenuation, 81 HU) in the renal parenchymal phase. (b) In a 50-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a predominantly solid enhancement pattern, with small low-attenuation foci that may represent necrosis or cystic changes. An avid degree of enhancement is present, with maximum attenuation of 239 HU during the renal parenchymal phase, which is nearly identical to that of adjacent renal parenchyma. (c) In a 54-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a mixture of hypervascular and low-attenuation areas. An avid degree of enhancement is present, with maximum attenuation of 209 HU, which is nearly identical to that of the adjacent renal cortex. (d) In a 48-year-old man with a large abdominal mass discovered at physical examination, surgical pathologic analysis revealed a 15-cm papillary RCC. The tumor (arrow) demonstrates a low degree of peripheral enhancement (maximum attenuation, 85 HU) and central low attenuation.

 

Figure 1C
View larger version (174K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1c: Transverse contrast material–enhanced CT images obtained during the renal parenchymal phase in four patients. (a) In a 63-year-old man with an incidentally discovered renal mass, surgical pathologic analysis revealed papillary RCC. The mass (arrow) demonstrates a homogeneous, relatively low-degree enhancement pattern (attenuation, 81 HU) in the renal parenchymal phase. (b) In a 50-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a predominantly solid enhancement pattern, with small low-attenuation foci that may represent necrosis or cystic changes. An avid degree of enhancement is present, with maximum attenuation of 239 HU during the renal parenchymal phase, which is nearly identical to that of adjacent renal parenchyma. (c) In a 54-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a mixture of hypervascular and low-attenuation areas. An avid degree of enhancement is present, with maximum attenuation of 209 HU, which is nearly identical to that of the adjacent renal cortex. (d) In a 48-year-old man with a large abdominal mass discovered at physical examination, surgical pathologic analysis revealed a 15-cm papillary RCC. The tumor (arrow) demonstrates a low degree of peripheral enhancement (maximum attenuation, 85 HU) and central low attenuation.

 

Figure 1D
View larger version (189K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1d: Transverse contrast material–enhanced CT images obtained during the renal parenchymal phase in four patients. (a) In a 63-year-old man with an incidentally discovered renal mass, surgical pathologic analysis revealed papillary RCC. The mass (arrow) demonstrates a homogeneous, relatively low-degree enhancement pattern (attenuation, 81 HU) in the renal parenchymal phase. (b) In a 50-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a predominantly solid enhancement pattern, with small low-attenuation foci that may represent necrosis or cystic changes. An avid degree of enhancement is present, with maximum attenuation of 239 HU during the renal parenchymal phase, which is nearly identical to that of adjacent renal parenchyma. (c) In a 54-year-old woman with an incidentally discovered renal mass, surgical pathologic analysis revealed clear cell RCC. The tumor (arrow) demonstrates a mixture of hypervascular and low-attenuation areas. An avid degree of enhancement is present, with maximum attenuation of 209 HU, which is nearly identical to that of the adjacent renal cortex. (d) In a 48-year-old man with a large abdominal mass discovered at physical examination, surgical pathologic analysis revealed a 15-cm papillary RCC. The tumor (arrow) demonstrates a low degree of peripheral enhancement (maximum attenuation, 85 HU) and central low attenuation.

 
The second feature evaluated was degree of enhancement. For heterogeneous lesions, each radiologist selected the area that demonstrated the greatest degree of enhancement of the renal lesion on the parenchymal phase images. Matching round or elliptical regions of interest approximately 0.1 cm2 in size were placed in this area on unenhanced and parenchymal and delayed phase images (Fig 2). Regions of interest were also placed in the adjacent normal renal cortex and aorta for normalization. For homogeneous lesions, regions of interest were placed in the center of the mass. Areas containing calcifications or artifacts were avoided.


Figure 2A
View larger version (125K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a: (a) Unenhanced, (b) parenchymal phase, and (c) excretory phase transverse CT images obtained in a 48-year-old man with clear cell RCC. To assess the degree of enhancement, readers used visual assessment within the mass to evaluate renal parenchymal phase images to select the area that appeared to demonstrate the greatest degree of enhancement. Matching regions of interest were placed in this area (1), as well as in the adjacent renal cortex (2) and the aorta (3).

 

Figure 2B
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b: (a) Unenhanced, (b) parenchymal phase, and (c) excretory phase transverse CT images obtained in a 48-year-old man with clear cell RCC. To assess the degree of enhancement, readers used visual assessment within the mass to evaluate renal parenchymal phase images to select the area that appeared to demonstrate the greatest degree of enhancement. Matching regions of interest were placed in this area (1), as well as in the adjacent renal cortex (2) and the aorta (3).

 

Figure 2C
View larger version (137K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2c: (a) Unenhanced, (b) parenchymal phase, and (c) excretory phase transverse CT images obtained in a 48-year-old man with clear cell RCC. To assess the degree of enhancement, readers used visual assessment within the mass to evaluate renal parenchymal phase images to select the area that appeared to demonstrate the greatest degree of enhancement. Matching regions of interest were placed in this area (1), as well as in the adjacent renal cortex (2) and the aorta (3).

 
The third feature evaluated was neovascularity. Recruitment of tumor vessels—evidenced by increased and irregular vessels adjacent to the affected kidney within Gerota fascia, with frequent arteriovenous shunting—was present or absent (Fig E1, radiology.rsnajnls.org/cgi/content/full/244/2/494/DC1). The fourth feature evaluated was contour of renal masses. Lesion contours were classified as smooth, lobulated, or irregular. The fifth feature evaluated was calcification. When present, the location (rim or center) of calcification in the mass was recorded.

Each reader used a four-point rating scale to give an overall impression of the degree of malignancy of the renal mass, without stating a specific tumor type. A score of 1 indicated a benign lesion (eg, angiomyolipoma or oncocytoma); a score of 2, an indeterminate lesion; a score of 3, a less-aggressive malignant lesion (eg, papillary or chromophobe RCC); and a score of 4, an aggressive malignant lesion (conventional RCC).

Statistical Analysis
We used the Pearson {chi}2 test to compare the distribution of features across the three groups of disease (benign, less aggressive, and aggressive lesions, as assessed with pathologic analysis). The unclassified RCC tumors and malignant tumors other than RCC tumors were not analyzed because of the small number and heterogeneous nature of tumors in this group. An unweighted {kappa} statistic was used to assess interreader agreement for the features: A {kappa} value of less than 0.20 indicated poor agreement; a {kappa} value of 0.20–0.40, fair agreement; a {kappa} value of 0.41–0.60, moderate agreement; a {kappa} value of 0.61–0.80, good agreement; and a {kappa} value of more than 0.80, excellent agreement (21).

To summarize the potential utility of the features, we calculated sensitivity, specificity, positive and negative predictive values, and the corresponding 95% confidence intervals. Confidence intervals were constructed in accordance with guidelines set forth by Rao and Scott in 1981 (22) to account for clustered data.

We used multivariate logistic regression analysis to build prognostic models. First, features were assessed univariately as independent variables in a logistic regression model where the dependent variable was either disease status (benign vs malignant) or level of aggressiveness (high vs low). Features with a P value less than or equal to .10 at univariate analysis were assessed with multivariate analysis. Data from both readers were included in the models by using the Proc Genmod command with a repeated statement in SAS software (SAS Institute, Cary, NC). Stepwise methods were used to reduce the models until all remaining variables had P values less than or equal to .05. P values less than .05 were considered to indicate a statistically significant difference.

When we looked at the categories of heterogeneity and calcifications in greater detail, comparisons were made with the Fisher exact test. Tests for associations between degree of enhancement and either disease status or aggressiveness were performed with Wilcoxon rank sum tests.

To determine cutoff points between lesion groups with different degrees of enhancement, we used a minimum P value approach similar to that described by Mazumdar and Glassman (23). We evaluated the cutoff points by assessing agreement between the classification of lesions based on these cutoff points and the true histologic classification by using a bootstrapped estimate of the unweighted {kappa} statistic (24,25). For this bootstrapped estimate, we sampled patients with replacement—allowing patients to be represented more than once in each bootstrap sample, obtained cutoff points from the bootstrap sample by using these cutoff points to categorize the original data, calculated the {kappa} statistic, and then estimated the potential bias by subtracting this {kappa} statistic from the original observed {kappa} statistic (ie, the uncorrected {kappa} statistic from cutoff points obtained by using the original data). This process was repeated 100 times. Presented are the overfitting corrected {kappa} statistics, which are equal to the uncorrected {kappa} statistic minus the average bias. Analyses were performed with SAS, version 9.0, for Windows (SAS Institute) and R for Windows (http://www.r-project.org/).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
Lesions
Of 198 lesions, 177 (89%) were malignant and 21 (11%) were benign. Of the 177 malignant lesions, 108 (61%) were clear cell RCC; 30 (17%), papillary RCC; and 24, (14%) chromophobe type. Ten (6%) of the 177 malignant lesions were unclassified RCCs, and five (3%) were classified as "other malignant," including three tumors with multiple histologic findings, one metastatic colorectal adenocarcinoma, and one Wilm tumor. The 21 benign lesions included 14 oncocytomas (67%), six angiomyolipomas (29%), and a leiomyoma (5%). None of the angiomyolipomas demonstrated appreciable fat on CT images. Most of the patients had early-stage tumors at the time of presentation (T1a, 62%; T1b, 12%; T2, 8%; T3, 13%; and T4, <1%) (Table 1). The median lesion size was 3.4 cm (range, 1.1–20.0 cm).


View this table:
[in this window]
[in a new window]

  		Table 1. Characteristics of Patients and Renal Lesions

 
Features
There was a significant association between lesion type and (a) homogeneous (P < .001 for both readers) or heterogeneous (P < .001 for both readers) enhancement pattern, (b) type of contour (reader 1, P = .04; reader 2, P < .001), and (c) observed neovascularity (P = .024 for reader 1) (Table 2). Furthermore, there was moderate to good agreement between the readers regarding these features, as evidenced by {kappa} statistics that ranged from 0.56 to 0.71. While there was excellent agreement between the readers regarding calcification ({kappa} = 0.78), this finding was not significantly associated with lesion type for either reader.


View this table:
[in this window]
[in a new window]

  		Table 2. Features Identified by Each Reader and Summarized according to Histologic Type

 
For both readers, there was a significant difference in the enhancement patterns of different renal tumor types. Both readers found that clear cell RCCs most commonly manifested with a mixed enhancement pattern of both enhancing soft-tissue components and low-attenuation areas that may have represented necrotic or cystic changes (reader 1, 88% of clear cell carcinomas; reader 2, 79% of clear cell carcinomas); however, a relatively high proportion of oncocytomas manifested with a similar enhancement pattern (reader 1, 43%; reader 2, 64%) (Fig E2, radiology.rsnajnls.org/cgi/content/full/244/2/494/DC1). When grouped together, homogeneous and peripheral enhancement patterns were seen most commonly in papillary RCCs (reader 1, 86%; reader 2, 93%) (Figs 1a, 1d, 3). The readers almost always categorized fat-poor angiomyolipomas as homogeneous (five of six angiomyolipomas [83%] for both readers) (Fig E3, radiology.rsnajnls.org/cgi/content/full/244/2/494/DC1).


Figure 3
View larger version (167K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3: Contrast-enhanced transverse CT image obtained in the renal parenchymal phase in a 79-year-old woman with a left renal mass (arrow) discovered incidentally at chest CT. The mass has central low attenuation and a relatively low degree of peripheral enhancement (132 HU) compared with renal cortex enhancement (241 HU). A diagnosis of papillary RCC was assigned at surgical pathologic analysis.

 
Benign versus Malignant Lesions
The pretest probability of a renal lesion being malignant was 89% in this patient cohort (Table E1, radiology.rsnajnls.org/cgi/content/full/244/2/494/DC1). Because all of the features had positive predictive values relatively close to 89%, we inferred that the presence of a single feature did not substantially aid in predicting whether a lesion was malignant. The one possible exception was calcification, which had a comparably high positive predictive value (98%).

Aggressive versus Less Aggressive Tumors
We included only malignant lesions. We defined less aggressive tumors as those of papillary and chromophobe subtypes at histologic analysis (54 [33%] of 162) and aggressive tumors as those with clear cell RCC characteristics (108 [67%] of 162). Ten unclassified malignant lesions were excluded from this portion of the analysis. Thus, in this cohort (Table E2, radiology.rsnajnls.org/cgi/content/full/244/2/494/DC1), the pretest probability that a lesion was aggressive was 67%. The pretest probability that it was less aggressive was 33%. The features neovascularity presence and heterogeneous enhancement pattern have positive predictive values greater than 67%, indicating that the presence of a heterogeneous pattern may increase the probability that a malignant lesion is aggressive. In addition, the absence of a heterogeneous pattern has a high negative predictive value, indicating that the absence of this pattern or the presence of a homogeneous pattern is more indicative of less aggressive disease (Fig 4, P < .001). Less aggressive tumors were more likely to have smooth contours (P < .01).


Figure 4
View larger version (167K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4: Contrast-enhanced transverse CT image obtained in the renal parenchymal phase in a 60-year-old woman with a left renal mass (arrow) discovered incidentally on MR images obtained because of abnormal liver function test results. The homogeneous mass has a relatively low degree of enhancement (80 HU) compared with the renal cortex enhancement (265 HU). A diagnosis of chromophobe RCC was assigned at surgical pathologic analysis. D = descending colon.

 
The patterns of heterogeneity and calcification were then analyzed in greater detail (Table E3, radiology.rsnajnls.org/cgi/content/full/244/2/494/DC1). The pattern of heterogeneity identified was significantly associated with the aggressiveness of the renal tumor (P < .001). A heterogeneous and mixed pattern was more likely to represent aggressive disease (Fig 5), whereas peripheral enhancing lesions were more likely to represent less aggressive disease. The presence and location(s) of calcification were not significantly associated with aggressiveness of disease.


Figure 5
View larger version (196K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5: Contrast-enhanced transverse CT image obtained in the renal parenchymal phase in a 68-year-old woman with a right renal mass (arrow) discovered incidentally during evaluation of diverticulitis. The mass has a heterogeneous mixed enhancement pattern, and a diagnosis of clear cell RCC was assigned at surgical pathologic analysis.

 
We attempted to combine the features to build a predictive model that could be used to differentiate between aggressive and less aggressive tumors. Only the enhancement pattern remained a significant predictor of tumor aggressiveness and was more predictive than any other feature or combination of features (Table 3).


View this table:
[in this window]
[in a new window]

  		Table 3. Enhancement Pattern and Odds Ratio

 
Overall Impression of Lesion Type Based on Features
For the overall impression of lesion type based on imaging features, readers correctly classified a large number of clear cell RCCs and less aggressive malignant lesions (papillary and chromophobe RCCs, Table 4). None of the unclassified RCCs and none of the malignant renal tumors other than RCCs were classified as benign. However, the readers had difficulty identifying benign lesions.


View this table:
[in this window]
[in a new window]

  		Table 4. Overall Impression of Lesion Types Based on Imaging Features and Pathologic Results

 
Degree of Enhancement
During the parenchymal phase, clear cell RCCs and oncocytomas tended to enhance avidly, and both chromophobe carcinomas and fat-poor angiomyolipomas enhanced moderately (Table 5, Fig 6). Papillary carcinomas demonstrated the least degree of enhancement among renal cortical tumors. During excretory phases, there was more overlap in the degree of enhancement across the different histologic subtypes.


View this table:
[in this window]
[in a new window]

  		Table 5. Attenuation of Renal Lesions in the Most Common Carcinoma Groups

 

Figure 6
View larger version (44K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6: During the parenchymal phase, renal tumor subtypes demonstrate significantly different degrees of enhancement. Data in this graph were obtained by reader 2 and are the mean attenuation values (measured in Hounsfield units). Error bars are standard deviations.

 
On the basis of the finding that renal tumors enhance differently during the parenchymal phase, cutoff point analysis was performed to divide all renal lesions into three groups (ie, those with mild, moderate, or avid enhancement during the parenchymal phase). The cutoff points to separate tumors into mild, moderate, or avid enhancement groups were 97 and 140 HU. To normalize the variations due to individual patient factors and technical factors, the parenchymal phase measurements were divided by the cortex or aorta values measured at the same time, and cutoff points were found and evaluated to see if there was better information with the normalized values. The ratio of parenchymal phase lesion enhancement to cortical enhancement appeared to yield the greatest agreement between the histologic findings predicted with cutoff points and the true histologic findings (Table 6) ({kappa} = 0.62), compared with both the absolute parenchymal phase measurements and the measurements normalized with aorta values, although the difference was small ({kappa} = 0.52 for absolute measurements and {kappa} = 0.57 for measurements normalized with aorta values). For the parenchymal phase lesion enhancement to cortical enhancement ratio, we found cutoff points of 0.453 and 0.654. Table 7 demonstrates the predictive power of these cutoff points for separating tumors into three groups.


View this table:
[in this window]
[in a new window]

  		Table 6. Ratio of Renal Lesion Enhancement to Cortical Enhancement in the Most Common Carcinoma Groups

 

View this table:
[in this window]
[in a new window]

  		Table 7. Predictive Power of Degree of Enhancement in Differentiating Renal Tumor Subtypes

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 
The advent of multidetector CT has led to better detection and earlier diagnosis of renal cortical tumors (26–32). Currently, as many as 70% of tumors are discovered incidentally, with a median tumor size smaller than 5 cm (29,33). Among all renal cortical tumors detected, approximately 20% are benign lesions (including renal oncocytomas), and 25% are relatively indolent papillary or chromophobe carcinomas (7–10,34). It appears that although 80% of renal cortical tumors are malignant, not all malignant tumors undergo substantial active growth (26,27,30). Without tumor growth, the risk of metastasis may be limited as well. Thus, treatment strategy may differ according to the patient's age, general medical condition, and renal function status, as well as the aggressiveness of the detected renal tumor. Since different subtypes of renal tumors are associated with different clinical implications, it is clinically important to differentiate these lesions preoperatively.

The results of some studies indicate that certain imaging features may be associated with specific subtypes of renal cortical tumor (15–20). The most consistent finding (found by Herts et al [15], Jinzaki et al [16], Kim et al [17], Ruppert-Kohlmay et al [19], and Sheir et al [20]) was that degree of enhancement was the most valuable parameter for differentiation of RCC subtypes, as clear cell RCCs enhance to a greater degree than other subtypes of malignant lesions, especially papillary RCCs (15–17,19,20). In addition, Herts et al (15) found that papillary RCCs are typically homogeneous, and Sheir et al (20) found that cystic degeneration was more evident in the clear cell RCC subtype than in other RCC subtypes. However, Sheir et al (20) found that a hypervascular pattern was more prevalent in papillary RCC subtypes than in chromophobe RCC subtypes. This finding was inconsistent with the findings of the studies cited previously (15–17,19). In addition, most of these studies included only malignant lesions or, in some cases, subgroups of malignant lesions in their analyses. Although Jinzaki et al (16) included both malignant and benign tumors in their study, their patient cohort was relatively small (total of 40 renal tumors).

To our knowledge, our study is the largest series to include patients with benign lesions and those with malignant lesions. The large number of patients allowed us to analyze the less common renal tumor subtypes. In addition, we performed feature analysis and obtained quantitative attenuation measurements for all renal lesions.

Our findings show that certain imaging features are as important as the degree of enhancement in differentiating renal tumor subtypes. For example, clear cell RCC is strongly associated with a mixed enhancement pattern of both enhancing soft-tissue components and low-attenuation areas that may represent necrotic or cystic changes (P < .001). On the other hand, when homogeneous or peripheral enhancement patterns are present, RCC is a less likely diagnosis, and other cell types should be considered. Notably, a majority of papillary tumors were either homogeneous or demonstrated peripheral enhancement.

Our study results also showed that the presence of neovascularity was mildly associated with a more aggressive tumor and that smooth contour was associated with a less aggressive tumor, neither of which were unexpected findings. Another interesting finding in our study was that calcification was seen mostly in malignant renal tumors, although its presence or pattern was not predictive of the aggressiveness of disease. This is consistent with the findings of studies in the 1970s that showed that calcification in renal masses on radiographs indicated malignancy (35). However, none of these additional features (either alone or in combination) was nearly as useful as the enhancement pattern in predicting tumor subtypes.

Our results showed that clear cell RCCs and oncocytomas enhanced avidly during the parenchymal phase, whereas chromophobe carcinomas and lipid-poor angiomyolipomas enhanced moderately and papillary tumors enhanced the least. The degrees of enhancement of the tumor subtypes overlapped substantially more in the precontrast and excretory phases. As suggested previously (15), the measured attenuation of the renal lesions should be normalized by using the measured attenuation of either the renal cortex or the aorta to ensure that attenuation is independent of technical or patient variability. We compared direct measurements of the attenuation of renal lesions, the renal lesion–to-cortex attenuation ratios, and the renal lesion–to-aorta attenuation ratios, and it appears that renal lesion–to-cortex attenuation ratios yielded the highest predictive accuracy, although the difference was small. When the renal lesion–to-cortex ratios obtained during the parenchymal phase were used for evaluation, we found cutoff points of 0.453 and 0.654 for dividing the renal lesions into those with avid enhancement (suggestive of clear cell RCC and oncocytoma), moderate enhancement (suggestive of chromophobe RCC and lipid poor angiomyolipoma), and mild enhancement (suggestive of papillary RCC).

Since renal tumors are often heterogeneous, we decided to measure the areas of greatest enhancement in the lesion rather than in the entire tumor. We believe that measurements obtained with this approach minimize volume averaging effects from areas of cystic or necrotic changes and truly reflect the vascularity of the tumor. Because of the small size and heterogeneous nature of many renal lesions, as well as the limitations posed by the thickness of the normal renal cortex, the region of interest used to measure attenuation tended to be small in our study, and the selection of areas of greatest degree of enhancement was based on visual assessment and was therefore somewhat subjective. Nevertheless, there was a high degree of agreement between the two readers' independent measurements, indicating that this is a robust and reproducible method with which to evaluate the degree of renal tumor enhancement.

Oncocytomas accounted for only 7% of all renal lesions in our series, but they presented the greatest diagnostic challenge. They may overlap with clear cell RCCs in terms of imaging features and degree of enhancement. In our cohort, lipid-poor angiomyolipoma was the second most common benign renal tumor and accounted for 3% of all resected renal tumors. No fat-containing angiomyolipomas were seen, as angiomyolipomas with appreciable fat on CT images would not have been resected. Most of the lipid-poor angiomyolipomas in our study were homogeneous (five of six, as assessed by both readers) and had a moderate degree of enhancement that overlapped with that of chromophobe RCCs.

Our study was focused on characterization of renal tumors with solid soft-tissue components, as extensive research has been performed on cystic renal masses and findings have been published (11,36–39). It has been shown that cystic tumors with thin walls and septa and without solid components in adults may represent multilocular cystic nephroma, multilocular cystic RCC, or, rarely, cystic hamartoma of the renal pelvis. However, these cystic tumors are similar in gross appearance and cannot be differentiated by using preoperative imaging studies (40). Thus, these lesions were excluded from our study.

In our study, most lesions were early stage; 62% of renal tumors were stage T1a lesions smaller than 4 cm. This is consistent with the downward migration in the size and stage of renal cortical tumors at presentation that has occurred over the last decade. Despite the generally small sizes of these lesions, their characterization based on imaging features and degree of enhancement could be achieved with reasonable interobserver agreement.

A limitation of our study was that the percentage of malignant lesions (89%) may have been slightly higher than that in the general population, since our institution is a tertiary referral center for oncology patients. However, even in the general population the incidence of malignant renal lesions is much higher than that of benign lesions. Another limitation of the study was its retrospective nature.

On the basis of our findings, we conclude that certain imaging features and the degree of enhancement on CT images are helpful in differentiating renal cortical tumor subtypes, despite some overlap. Thus, imaging findings may contribute incremental value to clinical parameters in providing prognostic information, consequently improving the quality of the data used in therapeutic planning.


   ADVANCES IN KNOWLEDGE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
IMPLICATION FOR PATIENT CARE
 References
 


   IMPLICATION FOR PATIENT CARE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
References
 


   ACKNOWLEDGMENTS
 
We thank Ada Muellner, BA, for her help in editing this manuscript.


   FOOTNOTES
 

Abbreviations: RCC = renal cell carcinoma

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, J.Z., R.A.L., P.R.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, J.Z., P.R.; clinical studies, J.Z., R.A.L., P.R., H.E.; statistical analysis, N.M.I., L.W., C.S.M.; and manuscript editing, J.Z., R.A.L., N.M.I., P.R., H.H.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 IMPLICATION FOR PATIENT CARE
 References
 

  1. Reuter VE, Presti JC Jr. Contemporary approach to the classification of renal epithelial tumors. Semin Oncol 2000;27:124–137.[Medline]
  2. Weiss LM, Gelb AB, Medeiros LJ. Adult renal epithelial neoplasms. Am J Clin Pathol 1995;103:624–635.[Medline]
  3. Storkel S, van den Berg E. Morphological classification of renal cancer. World J Urol 1995;13:153–158.[Medline]
  4. Russo P. Evolving understanding and surgical management of renal cortical tumors. Mayo Clin Proc 2000;75:1233–1235.[Medline]
  5. Kovacs G, Akhtar M, Beckwith BJ, et al. The Heidelberg classification of renal cell tumours. J Pathol 1997;183:131–133.[CrossRef][Medline]
  6. Motzer RJ, Bacik J, Mariani T, Russo P, Mazumdar M, Reuter V. Treatment outcome and survival associated with metastatic renal cell carcinoma of non-clear-cell histology. J Clin Oncol 2002;20:2376–2381.[Abstract/Free Full Text]
  7. Amin MB, Corless CL, Renshaw AA, Tickoo SK, Kubus J, Schultz DS. Papillary (chromophil) renal cell carcinoma: histomorphologic characteristics and evaluation of conventional pathologic prognostic parameters in 62 cases. Am J Surg Pathol 1997;21:621–635.[CrossRef][Medline]
  8. Renshaw AA, Henske EP, Loughlin KR, Shapiro C, Weinberg DS. Aggressive variants of chromophobe renal cell carcinoma. Cancer 1996;78:1756–1761.[CrossRef][Medline]
  9. Crotty TB, Farrow GM, Lieber MM. Chromophobe cell renal carcinoma: clinicopathological features of 50 cases. J Urol 1995;154:964–967.[CrossRef][Medline]
  10. Beck SD, Patel MI, Snyder ME, et al. Effect of papillary and chromophobe cell type on disease-free survival after nephrectomy for renal cell carcinoma. Ann Surg Oncol 2004;11:71–77.[Abstract/Free Full Text]
  11. Bosniak MA. The current radiological approach to renal cysts. Radiology 1986;158:1–10.[Abstract/Free Full Text]
  12. Dechet CB, Zincke H, Sebo TJ, et al. Prospective analysis of computerized tomography and needle biopsy with permanent sectioning to determine the nature of solid renal masses in adults. J Urol 2003;169:71–74.[CrossRef][Medline]
  13. Dechet CB, Sebo T, Farrow G, Blute ML, Engen DE, Zincke H. Prospective analysis of intraoperative frozen needle biopsy of solid renal masses in adults. J Urol 1999;162:1282–1284.[CrossRef][Medline]
  14. Lechevallier E, Andre M, Barriol D, et al. Fine-needle percutaneous biopsy of renal masses with helical CT guidance. Radiology 2000;216:506–510.[Abstract/Free Full Text]
  15. Herts BR, Coll DM, Novick AC, et al. Enhancement characteristics of papillary renal neoplasms revealed on triphasic helical CT of the kidneys. AJR Am J Roentgenol 2002;178:367–372.[Abstract/Free Full Text]
  16. Jinzaki M, Tanimoto A, Mukai M, et al. Double-phase helical CT of small renal parenchymal neoplasms: correlation with pathologic findings and tumor angiogenesis. J Comput Assist Tomogr 2000;24:835–842.[CrossRef][Medline]
  17. Kim JK, Kim TK, Ahn HJ, Kim CS, Kim KR, Cho KS. Differentiation of subtypes of renal cell carcinoma on helical CT scans. AJR Am J Roentgenol 2002;178:1499–1506.[Abstract/Free Full Text]
  18. Press GA, McClennan BL, Melson GL, Weyman PJ, Mauro MA, Lee JK. Papillary renal cell carcinoma: CT and sonographic evaluation. AJR Am J Roentgenol 1984;143:1005–1009.[Abstract/Free Full Text]
  19. Ruppert-Kohlmayr AJ, Uggowitzer M, Meissnitzer T, Ruppert G. Differentiation of renal clear cell carcinoma and renal papillary carcinoma using quantitative CT enhancement parameters. AJR Am J Roentgenol 2004;183:1387–1391.[Abstract/Free Full Text]
  20. Sheir KZ, El-Azab M, Mosbah A, El-Baz M, Shaaban AA. Differentiation of renal cell carcinoma subtypes by multislice computerized tomography. J Urol 2005;174:451–455.[CrossRef][Medline]
  21. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–174.[CrossRef][Medline]
  22. Rao JNK, Scott AJ. The analysis of categorical data from complex sample surveys. J Am Stat Assoc 1981;76:221–230.[CrossRef]
  23. Mazumdar M, Glassman JR. Categorizing a prognostic variable: review of methods, code for easy implementation and applications to decision-making about cancer treatments. Stat Med 2000;19:113–132.[CrossRef][Medline]
  24. Efron B. Estimating the error rate of a prediction rule: improvement on cross-validation. J Am Stat Assoc 1983;78:316–331.[CrossRef]
  25. Harrell FE. Regression modeling strategies. New York, NY: Springer-Verlag, 2000.
  26. Bosniak MA. Observation of small incidentally detected renal masses. Semin Urol Oncol 1995;13:267–272.[Medline]
  27. Bosniak MA, Birnbaum BA, Krinsky GA, Waisman J. Small renal parenchymal neoplasms: further observations on growth. Radiology 1995;197:589–597.[Abstract/Free Full Text]
  28. Bosniak MA. Problems in the radiologic diagnosis of renal parenchymal tumors. Urol Clin North Am 1993;20:217–230.[Medline]
  29. Russo P. Renal cell carcinoma: presentation, staging, and surgical treatment. Semin Oncol 2000;27:160–176.[Medline]
  30. Kassouf W, Aprikian AG, Laplante M, Tanguay S. Natural history of renal masses followed expectantly. J Urol 2004;171:111–113.[CrossRef][Medline]
  31. Motzer RJ, Russo P, Nanus DM, Berg WJ. Renal cell carcinoma. Curr Probl Cancer 1997;21:185–232.[CrossRef][Medline]
  32. Smith SJ, Bosniak MA, Megibow AJ, Hulnick DH, Horii SC, Raghavendra BN. Renal cell carcinoma: earlier discovery and increased detection. Radiology 1989;170:699–703.[Abstract/Free Full Text]
  33. Russo P. Localized renal cell carcinoma. Curr Treat Options Oncol 2001;2:447–455.[Medline]
  34. Bonsib SM. Risk and prognosis in renal neoplasms: a pathologist's prospective. Urol Clin North Am 1999;26:643–660, viii.[CrossRef][Medline]
  35. Daniel WW, Hartman GW, Witten DM, Farrow GM, Kelalis PP. Calcified renal mass: a review of ten years experience at the Mayo Clinic. Radiology 1972;103:503–508.[Medline]
  36. Israel GM, Bosniak MA. Follow-up CT of moderately complex cystic lesions of the kidney (Bosniak category IIF). AJR Am J Roentgenol 2003;181:627–633.[Abstract/Free Full Text]
  37. Israel GM, Bosniak MA. Calcification in cystic renal masses: is it important in diagnosis? Radiology 2003;226:47–52.[Abstract/Free Full Text]
  38. Israel GM, Bosniak MA. An update of the Bosniak renal cyst classification system. Urology 2005;66:484–488.[CrossRef][Medline]
  39. Israel GM, Hindman N, Bosniak MA. Evaluation of cystic renal masses: comparison of CT and MR imaging by using the Bosniak classification system. Radiology 2004;231:365–371.[Abstract/Free Full Text]
  40. Eble JN, Bonsib SM. Extensively cystic renal neoplasms: cystic nephroma, cystic partially differentiated nephroblastoma, multilocular cystic renal cell carcinoma, and cystic hamartoma of renal pelvis. Semin Diagn Pathol 1998;15:2–20.[Medline]




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Supplemental Figures and Tables
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zhang, J.
Right arrow Articles by Hricak, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zhang, J.
Right arrow Articles by Hricak, H.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE