Original Research
Genitourinary Imaging
March 22, 2013

Combined Late Gadolinium-Enhanced and Double-Echo Chemical-Shift MRI Help to Differentiate Renal Oncocytomas With High Central T2 Signal Intensity From Renal Cell Carcinomas

Abstract

OBJECTIVE. The purpose of our study was to evaluate the combination of dynamic contrast-enhanced T1-weighted and double-echo gradient-echo MR imaging to distinguish renal oncocytoma with high T2 signal intensity centrally from renal cell carcinoma (RCC).
MATERIALS AND METHODS. Between 2006 and 2011, 63 renal tumors (59 patients) presenting with a high signal intensity (SI) central area on T2-weighted sequences were imaged with dynamic contrast-enhanced sequences, including phases later than 5 minutes after contrast injection, and double-echo chemical-shift sequences were selected from our institutional database. Two experienced radiologists visually assessed presence and distribution of signal enhancement of central areas after injection and measured SI changes on opposed-phase images for calculation of the SI index and tumor-to-spleen ratio. Cutoff values were derived from the receiver operating characteristic (ROC) curve.
RESULTS. There were 19 oncocytomas (16 patients), 43 RCCs (42 patients), and one leiomyoma. Complete late enhancement of the central area was observed in 14 oncocytomas (74%) and in five RCCs (12%) (p = 0.05). The combination of complete enhancement and SI index lower than 2% (p = 0.02) or tumor-to-spleen ratio higher than −6% (p = 0.001) provided sensitivity of 36% and 55%, specificity of 95% and 97%, positive predictive value of 67% and 86%, and negative predictive value of 84% and 88%, respectively, for diagnosis of oncocytomas.
CONCLUSION. Absence of central area SI inversion or presence of a signal drop on chemical-shift imaging may rule out the diagnosis of oncocytoma.
Renal oncocytoma is considered a benign tumor, representing 3–7% of all renal tumors [1]. Apart from exceptional cases of renal carcinomas coexisting as hybrids or separate tumors, the prognosis is considered excellent [13]. On the basis of these considerations, conservative treatment can be recommended using nephron-sparing surgery [2, 4] when technically feasible. The importance of preoperative diagnosis is thus crucial, and, although the literature on characteristic features of renal oncocytomas abounds, until now, no imaging techniques have been identified that can accurately separate oncocytomas from malignant lesions [3, 5, 6]. Thus, diagnosis of oncocytoma remains based on histopathologic examination of surgical or biopsy specimens [7, 8] showing large eosinophilic cells with small round nuclei with large nucleoli, apical Hale coloration, and no expression of cytokeratin 7 and vimentin [9].
Recent studies have attempted to identify these benign tumors with contrast-enhanced CT according to an evaluation of the gradual enhancement after contrast injection [10] and the wash-in and washout of the contrast agent [11], which offer only low sensitivity and specificity. More recently, a segmental enhancement inversion was described within some renal tumors [12, 13] as a tumor segment with a lower density at the arterial phase and a higher density at the early excretory phase. This last feature was observed in indeterminate small renal masses without a scar and is associated, according to Kim et al. [12], with differences in hypocellular hyalinized stroma content within the tumor parenchyma. However, these results have not been confirmed by others [10, 13]. Using MRI, Rosenkrantz et al. [14] observed a comparable contrast inversion in 28.6% of oncocytomas and 13.3% of renal cell carcinomas (RCCs) but without differentiating inversions within the tumor parenchyma or with contrast enhancement within the central scar itself.
Morphologically, a central scar is found in 54% [5] to 80% [2] of renal oncocytomas. It was classically considered as evocative of this tumor type [15], but it is also observed in 30–40% of chromophobe carcinomas [14]. However, when a central area with high signal intensity (SI) on T2-weighted images is present within a tumor, the question of central necrosis within classic renal cell carcinoma still arises [16]. Whereas fibrous scars of hepatic focal nodular hyperplasia show delayed enhancement on gadolinium-enhanced T1-weighted sequences [17], such patterns in renal oncocytomas have received only little attention.
Because of signal loss on opposed-phase images, double-echo chemical-shift MRI sequences can detect small amounts of fat in atypical angiomyolipomas or intracellular lipids in clear cell and papillary carcinomas [18]. According to Rosenkrantz et al. [14], no change of signal intensity occurred in oncocytomas and chromophobe carcinomas.
The objective of this study was to evaluate the combination of dynamic contrast-enhanced T1-weighted and double-echo gradient-echo MRI in tumors presenting with a central area compatible with central scar or central necrosis to distinguish renal oncocytoma with high T2 SI centrally from RCC.

Materials and Methods

Patient Selection

Institutional review board approval was obtained for this retrospective study; informed patient consent was not required according to local laws. All patients examined with MRI for renal lesions in our institution between January 2006 and August 2011 were retrospectively selected from our prospectively maintained institutional database (Fig. 1). All cases identified were reviewed by two genitourinary radiologists with 5 and 1 years of experience to select only renal tumors visually presenting high SI central areas on T2-weighted sequences compatible with central scarring or necrosis and imaged with contrast-enhanced T1-weighted MRI sequences later than 5 minutes after contrast injection. In addition, a review of the institutional pathology database was completed to identify any additional cases. Any cases without histopathologic analysis available were excluded.
Fig. 1 —Flowchart shows study profile on basis of recommended standards for reporting diagnostic accuracy.

MRI Examination

All MRI examinations were performed on the same 1.5-T system (ACS-NT, Philips Healthcare) with a phased-array body coil. The MRI protocol is reported in Table 1. A delayed T1-weighted water-select gradient-echo sequence was systematically added in our institution (mean, 11 minutes; SD, 3.7 minutes; range, 6–26 minutes) after injection for all tumors to provide an adequate evaluation of their relationships with the pyelocaliceal system before treatment.
TABLE 1: MRI Protocol
MRI ProtocolMRI Sequence
T1-Weighted Gradient-EchoT2-Weighted Fast Spin-EchoDynamic Gradient-EchoLate T1-Weighted Water-Select Gradient-Echo
PlaneAxialAxialAxial or coronalAxial
Fat saturationNoNoYes or noYes or no
TR (ms)18221123.9235
TE (ms)4.6–2.31001.85
Angulation (°)70901080
Thickness (mm)5545
FOV (mm)375325384365
Matrix (mm × mm)220 × 284268 × 344184 × 184180 × 224
Scanning time (s)201803022
Delay (s)  40, 120, 170, 250> 360

Image Analysis

Two genitourinary radiologists with 5 and 20 years of experience blinded to the pathology results reviewed the MR images in consensus on a PACS workstation. Enhancement within the central area of these tumors was assessed visually. An SI inversion was considered to be present when, on late images, the central area presented an SI greater than the SI on peripheral tumoral tissues. This SI inversion was qualified as complete when including the entire central area and partial when only one part of this area showed enhancement.
The quantitative measurement of intratumoral lipid content was performed after drawing a region of interest (ROI) including the whole tumor and the central area, matched on in phase and opposed phase images. The change of SI was quantified using the SI index and the tumor-to-spleen ratio as previously proposed [19]: SI index = [(TuSIinTuSIopp) / (TuSIin)] × 100, where TuSIin is tumor SI on in phase images and TuSIopp is tumor SI on opposed phase images; tumor-to-spleen ratio = {[(TuSIopp / SpSIopp) / (TuSIin / SpSIin)] − 1} × 100, where SpSIin is spleen SI on in phase images and SpSIopp is spleen SI on opposed phase images.

Histopathology Analysis

Histopathologic diagnosis was available for all tumors after a percutaneous biopsy using an 18-gauge automated side-cutting needle (Monopty Disposable Core Biopsy Instrument, Bard) or surgical excision, which was reviewed retrospectively by a uropathologist with more than 30 years of experience. Tumors removed surgically were reanalyzed to determine the exact composition of the central areas. All tumors were fixed in acetic-formaldehyde, and cut tumor fragments were embedded in paraffin after dehydration. H and E was used to stain 3- or 4-μm sections, followed by immunostaining using vimentin, cytokeratin 7, CD10, CD117, and Hale colloidal iron.

Statistical Analysis

All patient data were acquired by one of the authors through review of medical records and imaging and pathology reports. Statistical significance to compare SI inversion in renal oncocytomas and RCCs was calculated using the Fisher exact test (Excel, Microsoft) and p values less than 0.05 were considered significant. SI index and tumor-to-spleen ratio were compared between oncocytomas and RCCs using the Student t test. To assess the diagnostic performance of SI inversion for differentiation of oncocytomas from RCCs, we calculated sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). ROC analysis was performed (ROCKIT, version 1.9B, Charles E. Metz, University of Chicago), and optimal cutoff values were derived from results.

Results

Patients

A total of 59 patients (35 men and 24 women; mean age, 63 years; range, 25–83 years) with 63 tumors were selected for the study from the institutional radiologic database. No further patients were identified in the pathology database. Histopathologic diagnosis was obtained after surgical excision of the tumors in 50 patients and exclusive percutaneous biopsy in nine. Overall, 26 tumors were percutaneously biopsied before any therapeutic decision. No change in the initial pathologic diagnosis in those patients who first underwent biopsy and then had surgical removal was reported.

Lesions

The mean size ± SD (range) of tumors was 39.9 ± 21.8 mm (18–103 mm) for oncocytomas, 51.4 ± 23 mm (17–101 mm) for RCC, and 96 mm for the leiomyoma. Overall, 19 tumors were oncocytomas in 16 patients (one patient had an oncocytomatosis with six tumors but only four had a central scar on T2-weighted images), 43 were RCCs (42 patients), and one was a leiomyoma (Fig. 2). Among the RCCs, 32 were clear cell carcinomas, seven were papillary carcinomas, and four were chromophobe carcinomas.
Fig. 2 —39-year-old woman with renal tumor.
A, Axial T2-weighted image (TR/TE, 2112/100; flip angle, 90°) shows left renal tumor isointense to renal cortex with central area of high signal intensity (arrow).
B and C, Axial in phase (B) and opposed phase (C) T1-weighted gradient-echo images (TR/TE range, 182/4.6–2.3; flip angle, 70°) show no variation of signal intensity.
D, Dynamic axial T1-weighted images (TR/TE, 3.9/1.8; flip angle, 10°) without fat suppression obtained 40 seconds (upper left), 2 minutes (upper right), 2.5 minutes (lower left), and 4.1 minutes (lower right) after gadolinium injection. Peripheral tumor component shows fast enhancement, whereas central area shows no enhancement.
E, Delayed (10 minutes) water-select image (TR/TE, 235/5; flip angle, 80°) shows no central enhancement.
F, Macroscopic photograph shows brown tumor with whitish stellate central scar (arrow). Microscopic examination (not shown) showed scar is fibrous with edematous rearrangements in which there were few capillaries without tumor cells, corresponding to leiomyoma.

Enhancement Analysis

The SI inversion analysis observed on late contrast-enhanced T1-weighted images is reported in Tables 2 and 3. A complete SI inversion was observed in 19 cases overall and, in all cases but one, only on late contrast-enhanced images (> 6 minutes after injection). Except in the case of oncocytomas (Fig. 3), the tumor type in this situation was a clear cell carcinoma in four cases (Fig. 4) and a chromophobe in one. When SI inversion was partial (n = 38), it always appeared at the periphery of the central area as a thin rim along the junction with the tumor component (Fig. 5).
TABLE 2: Late Contrast-Enhanced MRI Analysis
Enhancement of Central Area (Signal Inversion)Type
Oncocytoma (n = 19)Renal Cell Carcinoma (n = 43)Leiomyoma (n = 1)p by Fisher Exact Test
None05 (12)1 
Incomplete5 (26)33 (77)00.57
Complete14 (74)5 (12)00.05

Note—Data in parentheses are range.

TABLE 3: Accuracy for Differentiation of Oncocytomas From Renal Cell Carcinomas for All Criteria
CriteriaDifferentiationSensitivity (%)Specificity (%)Positive Predictive Value (%)Negative Predictive Value (%)
Signal intensity inversion analysisIncomplete26231342
 Complete74887488
 Incomplete or complete100933100
Chemical shift imaging analysisSignal intensity index < 2%50507921
 Tumor-to-spleen ratio > −6%70734190
Combination of both criteriaComplete signal intensity inversion and signal intensity index < 2%36956784
 Complete signal intensity inversion and tumor-to-spleen ratio > −6%55978688
Fig. 3 —62-year-old woman with renal tumor.
A, Axial T2-weighted image (TR/TE, 2112/100; flip angle, 90°) shows 56-mm renal tumor isointense to renal cortex with central area of high signal intensity.
B and C, Dual-shift images (TR/TE range, 182/4.6–2.3; flip angle, 70°) show no signal variation on in phase (B) or opposed phase (C) images with signal intensity index of 2% and tumor-to-spleen ratio of −5%.
D, Dynamic axial T1-weighted fat-suppressed images (TR/TE, 3.9/1.8; flip angle, 10°) obtained 40 seconds (upper left), 2 minutes (upper right), 2.5 minutes (lower left), and 4.1 minutes (lower right) after gadolinium injection. Peripheral tumor component shows early enhancement, whereas central area shows slow and progressive enhancement.
E, Delayed (15 minutes) T1-weighted water-select image (TR/TE, 235/5; flip angle, 80°) shows signal intensity inversion between tumor and central area is complete, without discernible washout.
F, Macroscopic photograph shows mahogany-brown tumor with whitish stellate central scar (arrow).
G, Photomicrographs show fibrous scar (top) corresponding microscopically with edematous and fibrohyaline tissue containing myofibroblasts, capillaries, and thin collagen fibers as well as low cellularity (bottom). Tumor was identified as oncocytoma.
Fig. 4 —74-year-old woman with renal tumor.
A, Axial T2-weighted (TR/TE, 2112/100; flip angle, 90°) image shows right renal tumor, slightly hyperintense to renal cortex with well-marked central area of high signal intensity.
B and C, Axial in phase (B) and opposed phase (C) T1-weighted gradient-echo images (TR/TE range, 182/4.6–2.3; flip angle, 70°) show low signal intensity of central area and slight signal intensity drop on opposed phase images with signal intensity index of 24% and tumor-to-spleen ratio of −24%.
D, Dynamic axial T1-weighted fat-suppressed images (TR/TE, 3.9/1.8; flip angle, 10°) obtained 40 seconds (upper left), 2 minutes (upper right), 2.5 minutes (lower left), and 4.10 minutes (lower right) after gadolinium injection show early enhancement of tumor component, with progressive enhancement of scar.
E, Delayed (11 minutes) axial T1-weighted water-select image (TR/TE, 235/5; flip angle, 80°) shows complete enhancement of central area (arrow).
F, Macroscopic (F) and microscopic (G) photographs show stellate hemorrhagic central scar (arrow, G) with peripheral edema, few capillaries, and no persistent tumor cells, corresponding to clear cell carcinoma with Fuhrman grade 2.
F and G, Macroscopic (F) and microscopic (G) photographs show stellate hemorrhagic central scar (arrow, G) with peripheral edema, few capillaries, and no persistent tumor cells, corresponding to clear cell carcinoma with Fuhrman grade 2.
Fig. 5 —61-year-old woman with renal tumor.
A, Axial T2-weighted image (TR/TE, 2112/100; flip angle, 90°) shows left renal tumor isointense to renal cortex with central area of high signal intensity.
B and C, Axial in phase (B) and opposed phase (C) T1-weighted gradient-echo images (TR/TE range, 182/4.6–2.3; flip angle, 70°) show variation of signal intensity on whole tumor with signal intensity index of 26% and tumor-to-spleen ratio of −25%.
D, Dynamic axial T1-weighted fat-suppressed images (TR/TE, 3.9/1.8; flip angle, 10°) obtained 40 seconds (upper left), 2 minutes (upper right), 2.5 minutes (lower left), and 4.1 minutes (lower right) after gadolinium injection show early enhancement of tumor component, with progressive enhancement of central area except in inner portion.
E, Delayed (14 minutes) water-select image (TR/TE, 235/5; flip angle, 80°) shows enhancement appears incomplete, with partial contrast reversal at interface between tumor and central area. Note thin rim of enhancement at junction between tumor and central area components (arrow).
F, Macroscopic photograph shows stellate central scar (arrow) corresponding to fibrotic component, with few necrosis and capillaries and without persistent tumor cells. Tumor was identified as clear cell carcinoma with Fuhrman grade 2.

Chemical-Shift Imaging Analysis Between Oncocytomas and Renal Cell Carcinoma Groups

Results of the chemical-shift imaging analysis between oncocytomas and RCC groups are summarized in Tables 3 and 4 and Figure 6. Optimal cutoff values of SI index of 2% and for tumor-to-spleen ratio of −6% were extracted from the ROC curve analysis (Fig. 7) for differentiation of the two groups. This means that a tumor with an SI index lower than 2% or a tumor with a tumor-to-spleen ratio greater than −6% was considered an oncocytoma.
TABLE 4: Chemical Shift MRI Analysis
Chemical Shift MRITypep
OncocytomaRenal Cell Carcinoma
Signal intensity index (mean ± SD) (%)−1.5 ± 9.1 (−19.1 to 11.7)10.1 ± 15.5 (−19.6 to 50.5)0.02
Tumor-to-spleen ratio (mean ± SD) (%)1.5 ± 11.6 (−15.2 to 20.3)−14.5 ± 14.2 (−54 to 8.9)0.001

Note—Data in parentheses are range.

Fig. 6A —Scatterplots of signal intensity index and tumor-to-spleen ratio for tumors of oncocytoma and renal cell carcinoma (RCC) groups.
A, Scatterplot of signal intensity index. With 2% as cutoff value (dashed line), sensitivity, specificity, and positive and negative predictive values for differentiation of oncocytomas from RCCs were 50%, 50%, 79%, and 21%, respectively.
Fig. 6B —Scatterplots of signal intensity index and tumor-to-spleen ratio for tumors of oncocytoma and renal cell carcinoma (RCC) groups.
B, Scatterplot of tumor-to-spleen ratio. Ratio of −6% was chosen as cutoff value (dashed line), and sensitivity, specificity, and positive and negative predictive values for differentiation of oncocytomas between two groups were 70%, 73%, 41%, and 90%, respectively.
Fig. 7 —Receiver operating characteristic (ROC) curve for signal intensity index and tumor-to-spleen ratio for differentiation of oncocytomas from renal cell carcinoma. Area under ROC curve was 0.70 (95% CI, 0.53–0.87) for signal intensity index and 0.81 (95% CI, 0.68–0.95) for tumor-to-spleen ratio.

Combination of Enhancement Analysis and Chemical-Shift Imaging for the Differentiation of Oncocytomas From Renal Cell Carcinomas

For the differentiation of oncocytomas from RCCs, the combination of a complete central SI inversion on gadolinium-enhanced images and a low SI index (cutoff value of 2%) provided sensitivity of 36%, specificity of 95%, PPV of 67%, and NPV of 84%. The combination of a complete central SI inversion on gadolinium-enhanced images and a smaller tumor-to-spleen ratio (cutoff value of −6%) gave sensitivity of 55%, specificity of 97%, PPV of 86%, and NPV of 88% (Table 3).

Pathology

Among the 19 oncocytomas, 13 were managed nonoperatively after percutaneous biopsy with basic imaging follow-up for 11 cases and percutaneous radiofrequency ablation for two. For these cases, no histopathologic information was available about central scarring.
Gross pathology was available in the six operated oncocytomas. Only five presented with a complete central SI inversion: the scar always appeared macroscopically as a whit-ish area (Fig. 3F), corresponding microscopically to an edematous and fibrohyaline tissue containing myofibroblasts, capillaries, and thin collagen fibers and persistent oncocytic cells (Fig. 3G). In the resected lesion with an incomplete inversion, pathologic analysis showed macroscopically a more yellowish appearance, corresponding with hyaline tissue with lower cellularity (less fibroblast, less collagen, and no oncocytic cells).
Gross pathology was available after surgical resection for the leiomyoma and for all RCCs. For the central area of the leiomyoma, gross pathology showed a fibrous scar with edematous rearrangements in which there were few capillaries without tumor cells.
The five carcinomas with complete SI inversion of the central area were all resected: four clear cell carcinomas and one chromophobe carcinoma. Pathologic analysis showed that the central area corresponded with a fibrotic component with a low level of necrosis and a few capillaries without persistent tumor cells (Figs. 4E and 4F). In the 32 cancer lesions without complete SI inversion, only necrosis was present centrally.

Discussion

Characterizing tumors before deciding on the most appropriate treatment is a general rule in oncology. However, pretherapeutic characterization of renal tumors is still problematic in clinical practice. The high prevalence of RCC and the lack of reliable imaging criteria for recognition of benign tumors other than typical angiomyolipoma justify the growing role of pretherapeutic imaging-guided biopsy. Diagnosis of oncocytoma and low-fat-content angiomyolipoma remains a challenge [7, 8]. Validation of new imaging criteria would be useful before deciding on any therapeutic option and to avoid total nephrectomy when possible.
In this retrospective series of renal masses on MRI, the high prevalence of oncocytomas (16/63, 25.4%) has three main causes. The first is that, in our clinical practice, MRI was often indicated when this tumor type was suspected on CT, as proposed by Kim et al. [12], and to look for microscopic fat content enabling us to rule out the hypothesis of oncocytoma where applicable [14, 18]. The second is that if a renal tumor presented a central area and any typical malignant patterns on CT, as described previously in some cases of RCC [12, 14], this mass was usually not imaged with MRI and, in consequence, the rate of benign tumors increased. Finally, our study only evaluated tumors with high central T2 signal intensity, which probably increased the rate of oncocytomas relative to RCCs by excluding all cancers that did not have central necrosis or fibrosis.
Our study provides some important results. First, all oncocytomas with central areas presented late central SI inversions within the central scar, responsible for a contrast inversion that was either complete or incomplete. Second, examples of all subtypes of RCC, not only chromophobes, showed complete or partial central SI inversions. Third, a signal loss on opposed phase images, especially after the calculation of SI index and tumor-to-spleen ratio, was significantly more frequently observed in all subtypes of RCC than oncocytoma. In other words, absence of central area SI inversion or presence of a signal drop on chemical-shift imaging can be used to rule out the diagnosis of oncocytoma.
We decided to retrospectively focus on the late gadolinium enhancement of central areas of renal tumors on the basis of experience with focal nodular hyperplasia, which shows central fibrous tissue with delayed enhancement [17]. Our study showed that most central fibrous scars enhance slowly after injection of gadolinium, justifying performing late T1-weighted sequences after injection, later than 5 minutes, when an oncocytoma was suspected on T2-weighted imaging or CT.
This late SI inversion explains why it has never been described with CT, the late excretory phase usually being performed only 3 minutes after injection. The segmental SI inversion between the corticomedullary phase and the early excretory phase recently described by Kim et al. [12] using MDCT corresponded to reversing tumor segments in small homogeneous tumors without a central scar.
We believe that this delayed contrast inversion between the central scar and the peripheral tumor parenchyma observed in oncocytomas is probably due more to the late SI inversion of the central tissue than to the high level of washout of contrast material from the periphery, as reported by Bird et al. [11] because it increased progressively on dynamic images in a centripetal manner.
Rosenkrantz et al. [14] evaluated a series of 28 oncocytomas and 15 chromophobe RCCs, with dynamic contrast-enhanced T1-weighted images that were acquired earlier (no later than 3 minutes after injection). They described a segmental SI inversion in 28.6–42.9% of oncocytomas and in 13.3–26.7% of chromophobe RCCs according to each reader. However, no distinction was made in the results between segmental SI inversion within tumor compartments in homogeneous lesions, as defined by Kim et al. [12], and the SI inversion within scars. Moreover, only chromophobe types of RCC were evaluated, whereas our study showed that enhancing fibrous central scars could also be observed in clear cell and papillary RCC.
Renal oncocytomas are composed of two independent components: oncocytic cells and pronounced hyaline and a few fibrotic stromas, which consist of proliferating myofibro-blasts with edema and deposition of extra-cellular matrix (collagen I, collagen III, and fibronectin) [20]. In our series, differences in the degree of SI inversion in central areas of renal masses were related to heterogeneity in vascularization of the tissue and the presence or absence of residual tumor cells. In oncocytomas, enhancing areas corresponded with edematous and fibrohyaline tissue, with poor but preserved cellularity and perfusion, whereas nonenhancing areas corresponded with noncellular and nonperfused hyaline compartments. In carcinomas, enhancing areas corresponded with fibrotic tissue, preserved perfusion without residual tumor cells, and nonenhancing areas due to necrosis, whether hemorrhagic or not.
The microscopic fat content of renal tumors evaluated in vivo with MRI using chemical-shift double-echo sequences [19] is typically present in tumor cell cytoplasm in the clear cell subtype of RCC [15, 21] and in spumous histiocytes in papillary RCC [18]. Such content was reported in only one case of a chromophobe RCC [14], but the decrease in SI was assessed only visually. In the same study, none of the oncocytomas showed any signal decrease on opposed phase images. These results support the findings of Krishnan and Truong [22] who noted that only minimal lipid could be detected with electron microscopic examination in oncocytoma and chromophobe RCC. In our study, quantification of SI changes using SI index and tumor-to-spleen ratio showed some overlap between oncocytomas and RCCs and provided cutoff values that, when associated with the analysis of central SI inversion, provided high levels of specificity and NPV.
Our study had some limitations. The first is that the series is retrospective and may have selection bias, as mentioned earlier, explaining the high prevalence of oncocytomas in our series. Second, the sample size is small and this study evaluated only a subset of renal tumors. We did not address oncocytomas or carcinomas with more homogeneous appearance. Third, the SI inversion was assessed qualitatively, although consensus of two experienced radiologists was obtained. Fourth, given the relatively low frequency of oncocytomas, we included MR images acquired over a period longer than 6 years, making the time delay for late MR images quite heterogeneous. Fifth, this was a single institution study, and only one MRI scanner was used. This could flaw the data, especially when calculating precise cutoff values based on MRI-generated signal intensity values as has been proposed.
In summary, according to our results, renal masses presenting with a central heterogeneous area, compatible with a central scar on contrast-enhanced CT or T2-weighted MR images, should be further studied with a delayed (more than 5 minutes) gadolinium-enhanced T1-weighted and a chemical-shift series to better characterize these lesions. Absence of any visual SI inversion of this central area or decrease in SI on opposed phase images should rule out the diagnosis of oncocytoma. In this series, on delayed images all oncocytomas with a scar showed a complete or partial SI inversion, higher than for other areas of the tumor. Some RCCs can also have late-enhancing central areas, but in these cases, some degree of signal drop in opposed phase MR images can assist in the differential diagnosis.

Acknowledgments

We thank Pippa McKelvie-Sebileau for medical editorial assistance in English.

References

1.
Romis L, Cindolo L, Patard JJ, et al. Frequency, clinical presentation and evolution of renal oncocytomas: multicentric experience from a European database. Eur Urol 2004; 45:53–57; discussion, 57
2.
Chao DH, Zisman A, Pantuck AJ, Freedland SJ, Said JW, Belldegrun AS. Changing concepts in the management of renal oncocytoma. Urology 2002; 59:635–642
3.
Newhouse JH, Wagner BJ. Renal oncocytomas. Abdom Imaging 1998; 23:249–255
4.
De Carli P, Vidiri A, Lamanna L, Cantiani R. Renal oncocytoma: image diagnostics and therapeutic aspects. J Exp Clin Cancer Res 2000; 19:287–290
5.
Marciano S, Petit P, Lechevallier E, et al. Renal oncocytic adenoma [in French]. J Radiol 2001; 82:455–461
6.
Bandhu S, Mukhopadhyaya S, Aggarwal S. Spoke-wheel pattern in renal oncocytoma seen on double-phase helical CT. Australas Radiol 2003; 47:298–301
7.
Lechevallier E, Andre M, Barriol D, et al. Fine-needle percutaneous biopsy of renal masses with helical CT guidance. Radiology 2000; 216:506–510
8.
Eshed I, Elias S, Sidi AA. Diagnostic value of CT-guided biopsy of indeterminate renal masses. Clin Radiol 2004; 59:262–267
9.
Cochand-Priollet B, Molinie V, Bougaran J, et al. Renal chromophobe cell carcinoma and oncocytoma: a comparative morphologic, histochemical, and immunohistochemical study of 124 cases. Arch Pathol Lab Med 1997; 121:1081–1086
10.
McGahan JP, Lamba R, Fisher J, et al. Is segmental enhancement inversion on enhanced biphasic MDCT a reliable sign for the noninvasive diagnosis of renal oncocytomas? AJR 2011; 197:916; [web]W674–W679
11.
Bird VG, Kanagarajah P, Morillo G, et al. Differentiation of oncocytoma and renal cell carcinoma in small renal masses (< 4 cm): the role of 4-phase computerized tomography. World J Urol 2011; 29:787–792
12.
Kim JI, Cho JY, Moon KC, Lee HJ, Kim SH. Segmental enhancement inversion at biphasic multi-detector CT: characteristic finding of small renal oncocytoma. Radiology 2009; 252:441–448
13.
Millet I, Doyon FC, Hoa D, et al. Characterization of small solid renal lesions: can benign and malignant tumors be differentiated with CT? AJR 2011; 197:887–896
14.
Rosenkrantz AB, Hindman N, Fitzgerald EF, Niver BE, Melamed J, Babb JS. MRI features of renal oncocytoma and chromophobe renal cell carcinoma. AJR 2010; 195:1338; [web]W421–W427
15.
Pedrosa I, Alsop DC, Rofsky NM. Magnetic resonance imaging as a biomarker in renal cell carcinoma. Cancer 2009; 115:2334–2345
16.
Ball DS, Friedman AC, Hartman DS, Radecki PD, Caroline DF. Scar sign of renal oncocytoma: magnetic resonance imaging appearance and lack of specificity. Urol Radiol 1986; 8:46–48
17.
Mortelé KJ, Praet M, Van Vlierberghe H, Kunnen M, Ros PR. CT and MR imaging findings in focal nodular hyperplasia of the liver: radiologic-pathologic correlation. AJR 2000; 175:687–692
18.
Yoshimitsu K, Irie H, Tajima T, et al. MR imaging of renal cell carcinoma: its role in determining cell type. Radiat Med 2004; 22:371–376
19.
Kim JK, Kim SH, Jang YJ, et al. Renal angiomyolipoma with minimal fat: differentiation from other neoplasms at double-echo chemical shift FLASH MR imaging. Radiology 2006; 239:174–180
20.
Yen TH, Chen Y, Fu JF, et al. Proliferation of myofibroblasts in the stroma of renal oncocytoma. Cell Prolif 2010; 43:287–296
21.
Pedrosa I, Chou MT, Ngo L, et al. MR classification of renal masses with pathologic correlation. Eur Radiol 2008; 18:365–375
22.
Krishnan B, Truong LD. Renal epithelial neoplasms: the diagnostic implications of electron microscopic study in 55 cases. Hum Pathol 2002; 33:68–79

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: 830 - 838
PubMed: 23521457

History

Submitted: April 24, 2012
Accepted: June 28, 2012

Keywords

  1. kidney
  2. MRI
  3. oncocytoma
  4. tumors

Authors

Affiliations

FranÇois Cornelis
Department of Radiology, Pellegrin Hospital, Pl Amélie Raba Léon, 33076 Bordeaux, France.
Department of Radiology, Bergonié Institute, Bordeaux, France.
Anne-Sophie Lasserre
Department of Radiology, Pellegrin Hospital, Pl Amélie Raba Léon, 33076 Bordeaux, France.
Thomas Tourdias
Department of Radiology, Pellegrin Hospital, Pl Amélie Raba Léon, 33076 Bordeaux, France.
Colette Deminière
Department of Pathology, Pellegrin Hospital, Bordeaux, France.
Jean-Marie Ferrière
Department of Urology, Pellegrin Hospital, Bordeaux, France.
Yann Le Bras
Department of Radiology, Pellegrin Hospital, Pl Amélie Raba Léon, 33076 Bordeaux, France.
Nicolas Grenier
Department of Radiology, Pellegrin Hospital, Pl Amélie Raba Léon, 33076 Bordeaux, France.

Notes

Address correspondence to F. Cornelis ([email protected]).

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media