Diagnostic Pathology Sub-typing of Renal Cell Tumours; Contribution of Ancillary Techniques

Background: Adult renal epithelial neoplasms are a heterogeneous group with varying prognosis and outcome requiring sub-classification.


Background
Renal cell carcinoma (RCC), accounts for 2-3% of all new cancers diagnosed and 85% of all primary renal neoplasms in adults [1]. Adult renal epithelial neoplasms are a heterogeneous group comprised of subtypes that have distinct gross, histologic, ultrastructural, and immunohistochemical features. These morphologic distinctions are amply supported by unique cytogenetic and chromosomal aberrations for many of the subtypes [2,3]. Thus classification of renal cell carcinoma is important from the treatment and prognosis point of view as well as for understanding its histogenesis, molecular and cytogenetic behaviour for further improvement in its management approach.
Keeping this in view, many classification systems have been made till date. Mainz classification [4] and subsequently AFIP classification [5] sub-classified these tumours purely based on morphological grounds. However, the first classification based on molecular and cytogenetic studies and on the cell of origin of each entity came into being in 1997 as the Heidelberg classification [6]. Subsequently, WHO classified renal cell tumors into clear cell RCC, multilocular clear cell RCC, papillary RCC, chromophobe RCC, carcinoma of the collecting ducts of Bellini, renal medullary carcinoma, Xp11 translocation carcinoma, carcinoma associated with neuroblastoma, mucinous tubular and spindle cell carcinoma, renal cell carcinoma-unclassified, papillary adenoma and oncocytoma [7].
Whilst a majority of renal cell carcinomas seen in routine practice of surgical pathology, are easy to diagnose based on haematoxylin and eosin preparation alone, there is insufficient information on how to deal with the differential diagnoses regarding subtypes of RCC. The present study aims at classifying adult renal tumors based on the cell of origin, by histomorphology, immunohistochemistry, and ultrastructural studies.

Methods
The material of this study was derived from cases of renal cell carcinoma received in the Department of Histopathology, Post Graduate Institute of Medical Education and Research PGIMER, Chandigarh from July, 1997 to June, 2007 (ten years period).
Patients aged more than 16 years were included in the study. All the cases were routinely fixed in 10% buffered formalin. 4 μ sections were cut and haematoxylin and eosin staining was performed in routine using the standardized methods. Haematoxylin and eosin stained sections were reviewed by independent pathologists in a blinded manner and the diagnosis of the four pathologists were compared. All cases with discordant diagnosis or in which a definitive diagnosis on the basis of morphology was not possible were taken up for further study which included histochemical, immunohistochemical, and ultrastructural analysis.

Hale's colloidal iron stain
Cases in which chromophobe cell carcinoma was kept as possibility by any of the three observers were taken up for Hale's colloidal iron stain. The method of Hale's colloidal iron stain used was a modification published by Tickoo et al [8].

Immunohistochemistry
All the cases with discordant diagnoses were taken up for immunohistochemistry; vimentin (Dako Cytomation, 1:50 dilution), Cytokeratin 7(Dako Cytomation, 1:50 dilution), and vinculin (Santa Cruz Biotechnology, 1:40 dilution) according to the differential diagnosis kept in a given case and with a minimum of 5 cases in each subgroup and including all 5 oncocytomas.
Immunostaining was carried out on 4 μ paraffin sections after antigen retrieval using pressure cooker method. The sections were incubated with appropriately diluted primary antibody, washed in PBS (3 × 5 min) and then incubated with peroxidase conjugated secondary antibody. Color reaction was developed by DAB and counterstained by hematoxylin. Appropriate positive and negative controls were taken. Immunoreactivity was evaluated by taking into account the percentage positivity of tumor cells. Positivity was taken as a brown reaction product staining the cytoplasm. The cells were scored as negative or positive and the percentage of positive tumor cells were recorded, which ranged from 0-100%. The percentage positivity was graded from 1+ to 3+ as follows:-

Electron Microscopy
Ultrastructural analysis of selected cases was done to confirm the diagnosis, when required. The tissue for electron microscopy was taken from formalin fixed tissue of all cases of oncocytoma and representative cases from discordant groups. Tissue blocks were fixed in 3% buffered glutaraldehyde and processed for electron microscopy by routine methods. The grids were examined under Zeiss 906 electron microscope, and representative photographs were taken under suitable magnification.

Results
A total of 278 cases of renal cell carcinoma were included in this study. Age of the patients ranged from 16-78 years, the average being 52 years (mean = 51.45). The peak incidence was in the fourth and fifth decades and the male to female ratio of 2.3:1. Out of these 278 renal carcinomas, concordant diagnoses were obtained in 250 cases whereas in the remaining 28 cases, diagnosis varied amongst the four pathologists. The discordance was resolved by histochemistry, immunohistochemistry and electron microscopy as detailed in Table 1. Table 2 shows incidence of sub-types of RCCs amongst the total of 278 cases.
Clear Cell RCC was the most common kidney tumor accounting for 74.8% (208/278) of all adult renal tumors. They exhibited a male preponderance and mean age at presentation of 56.2 years. Clear cell RCC were mostly sol-itary and bilaterality was noted in a single case. Fuhrman grading revealed that 85% of clear cell renal cell carcinomas were grade 1 and 2, and less than 5% were grade 4. Multicystic clear cell renal cell carcinoma was noted in 5 cases, 4 of which were nuclear grade 1 and one case was nuclear grade 2. Cytoplasmic inclusions were noted in 78 cases. Well formed psammoma bodies and fibro-calcific bodies were seen in 2 cases each; however calcification, cholesterol clefts, hemosiderin pigment and hyaline globules were seen in many cases. Areas with rhabdoid differentiation were noted in 2 cases while granulomatous inflammation was noted in the adjoining lymph node in another 2 cases. Sarcomatoid change was seen in 5 cases. One case revealed membranous glomerulonephritis in the adjoining renal parenchyma. Invasion of peri-renal fat and extension into the renal vein with thrombosis was noted in two cases.
Papillary (or chromophilic) renal cell carcinomas comprised 12.2% (n = 34) of cases. The mean age at presentation was 52.4 years and the sex ratio was M: F = 2.1:1. Based upon morphology two types of PRCC were categorized; Type 1 tumours (27/34 cases) had papillae covered by small cells with scanty basophilic cytoplasm, arranged  Renal oncocytoma comprised 1.8% (n = 5) of all renal cell carcinomas. The mean age at presentation was 48 years with a slight female preponderance. The cells were roundto polygonal with densely granular eosinophilic cytoplasm.

Others
Only one case of collecting duct carcinoma in a 70 year male was seen. The cells of collecting duct carcinoma displayed high nuclear grade (Fuhrman 4). Sarcomatoid RCC was noted in 2.9%(n = 8) of all adult renal cell carcinomas. The mean age at presentation was 66.2 years with a male preponderance. As the sarcomatoid element overshadowed the original antecedent carcinoma to the extent that it could not be recognized, so it was kept under Sarcomatoid RCC or unclassified category (according to the recent WHO).

Discordant cases (Table 1)
Most (12/28 cases) of the discordance was between clear cell RCC and chromophobe RCC. This was solved by Hale's colloidal iron stain which yielded a diffuse blue granular cytoplasmic staining in chromophobe renal cell carcinoma (Fig. 1) and pink periodic acid Schiff positivity in cytoplasm of clear cell renal cell carcinoma cells. Other subtypes of RCC were negative. This stain was very helpful and statistically significant (p < 0.0001) in differentiating chromophobe renal cell carcinoma from other subtypes of RCC. CK7 was also positive in all cases of chromophobe RCC and negative in clear cell RCC. Next group of discordance was between clear cell RCC, with papillary areas and papillary RCC (10/28) which was resolved by CK7 immunostain which was positive in most of the cases (90%) of papillary RCC and was negative in all cases of clear cell RCC. Third group of discordance was between eosinophilic variant of chromophobe RCC and oncocytoma. This discordance was resolved by CK7 immunostain which was positive in all cases of chromophobe RCC and was negative in oncocytoma. Electron microscopy was of diagnostic importance in oncocytomas by showing cytoplasm packed with mitochondria to the exclusion of other organelles.
CK 7 (Fig. 2 and 3): was positive in 90%(9/10) cases of papillary renal cell carcinoma (p < 0.001) and was more frequently observed in type 1 (100%) than type 2 (75%) tumours. All the cases of chromophobe renal cell carcinoma (10/10) and collecting duct carcinoma ( Vinculin (Fig. 4) was weakly but distinctly positive on the cell membrane of only 2/15 cases of chromophobe RCC and 1 case of collecting duct RCC. It was negative in all cases of clear cell renal cell carcinoma (0/13), papillary renal cell carcinoma (0/12) and renal oncocytoma (0/5). It was not found to be useful in resolving the differential diagnosis because of its low sensitivity in cases of chromophobe RCC.

Electron Microscopy
Ultrastructural examination was done in renal oncocytoma(5 cases), chromophobe renal cell carcinoma (7 cases), and 2 cases each of eosinophilic variants of clear cell renal cell carcinoma and type 2 papillary RCC. Special attention was paid to mitochondria and microvesicles and interrelations thereof. The cytoplasm of all the oncocytomas were packed with abundant mitochondria (Fig. 5).
Although abundant microvesicles were present in all the chromophobe renal cell carcinomas, but scant numbers of microvesicles were also present in renal oncocytomas and in the eosinophilic variant of clear cell renal cell carcinoma. The mitochondria in all three types of renal neoplasms studied differed in morphology, being predominantly uniform and round with prominent lamellar cristae packing the cytoplasm in renal oncocytoma. Ultrastructural examination of papillary RCC showed basal infoldings and increased mitochondria with glycogen granules.

Discussion
Classification of renal cell carcinoma is important from the treatment and prognosis point of view as well as for understanding of histogenesis. The main objective of this study was to establish the degree of inter-observer variation in sub-classification of renal cell carcinomas and to find out means of subtyping renal cell carcinoma short of cytogenetic studies. In the present study, cases were seen by four independent pathologists in a blinded manner and their diagnoses were compared. Out of these 278 renal tumors, concordant diagnosis was obtained in 250 cases whereas in the remaining 28 cases, diagnoses varied amongst the four pathologists. This data emphasizes the fact that haematoxylin and eosin histology alone is sufficient to classify a majority of renal cell carcinoma. However help from ancillary techniques may be required in approximately 10% of the cases.
The peak incidence of RCC in this study was in the fourth and fifth decades, in contrast to other studies in the western world, where the majority of cases were in their sixth and seventh decades [9][10][11]. The data in the present study also showed that the mean age of patients with chromophobe RCC and renal oncocytoma were less than the other sub-types. Literature from the western world show two to three fold male predominance in renal tumors. The data in the present study also showed similar results with a sex ratio, M: F = 2.3:1. However, the sex ratio was altered in chromophobe RCC and renal Oncocytoma where a marginal female predominance was seen. This data also varied with the world literature which shows male predominance in oncocytoma and equal sex incidence in chromophobe RCC [9][10][11].
Fuhrman grading of most (85%) cases of clear cell renal cell carcinomas in our study was grade 1 and 2, and less than 5% was nuclear grade 4. This data, however varied from the study by Srivastava et al [16] who showed 14.6% grade 1 tumor, 38.2% grade 2 tumor, 35.4% grade 3 tumor and 11.8% grade 4 tumor. Sarcomatoid change was seen in 5 cases of clear cell renal cell carcinoma, and in one case of papillary renal cell carcinoma. This is consistent with literature which shows less than 5% clear cell renal cell carcinoma with sarcomatoid dedifferentiation [10].
Vimentin was positive in 53.9% cases of clear cell renal cell carcinoma, 80% cases of papillary renal cell carcinoma, focally positive in one case of chromophobe RCC and diffusely positive in another case of chromophobe RCC with sarcomatoid areas. One case of collecting duct RCC showed focal positivity in stromal cells, however, oncocytomas (0/5) showed negative vimentin staining. The result was comparable to the study by Dierick et al [17] who also found 53.3% positivity in cases of renal cell carcinoma. Similar findings have been reported by Waldherr et al [18] and by Holthofer et al [19]. However, vimentin immunostain did not prove to be useful in differential diagnosis of subtypes of renal cell carcinoma.
In this study, CK 7 was positive in 90%(9/10) cases of papillary renal cell carcinoma, 100% cases of chromophobe renal cell carcinoma, and the single case of collecting duct carcinoma. It was negative in all cases of clear cell  The tinctorial characteristics of different renal epithelial neoplasms appear to be dependent on cytoplasmic contents, including various organelles [24]. Differential diagnoses between eosinophilic variant of clear cell renal cell carcinoma, oxyphil variant of chromophobe renal cell carcinoma and renal oncocytoma was resolved by electron microscopy. The mitochondria in these three types of renal neoplasms differed in morphology, being predominantly uniform and round with predominantly lamellar cristae in renal oncocytoma, variable in shape and size with predominantly tubulocystic cristae in chromophobe renal cell carcinoma, and swollen and pleomorphic with rarefied matrix and attenuated cristae in the eosinophilic variant of conventional (clear cell) renal cell carcinoma.

Conclusion
It can be concluded that incidence of different subtypes of renal tumours is similar to western and other Asian countries, however the age of presentation is one decade earlier than western population. In order to accurately subclassify renal cell carcinomas, the ancillary techniques required in a good surgical pathology set up consist of Hale's colloidal iron stain, immunostain CK7 and electron microscopy. With the help of these three ancillary techniques it would be possible to resolve all difficult cases that one may be faced with.