Renal oncocytomas (ROs) comprise approximately 3-7% of renal epithelial neoplasms 123. Common origin of RO and chromophobe renal cell carcinoma (ChRCC) from the intercalated cells of collecting ducts explains the histomorphologic, immunophenotypical, ultrastructural and molecular similarities of both neoplasms 234567. The benign nature of ROs, compared with the potential of ChRCCs for metastases and sarcomatoid transformation 89 has led to search for a confirmatory diagnostic modality; a broad array of ancillary studies have been employed in the differential diagnosis of the two neoplasms, including electron microscopy, flow cytometry, immunohistochemistry, cytogenetics and fluorescent in situ hybridization (FISH).

Cytogenetic studies of RO revealed genetic heterogeneity of these neoplasms; three distinct subsets were identified: 1) translocation between 11q13 and other chromosomes, 2) loss of 1/1p followed by loss of chromosome Y or 14 and 3) non-recurrent or no detectable aberrations 10. Molecular and fluorescent in-situ hybridization (FISH) studies of ROs described partial or total loss of chromosome 1 to be the most consistently reported clonal chromosomal abnormality in RO 111213 with a possible tumor suppressor gene expected on chromosome 1p 14. In contrast, ChRCC was consistently found to display multiple losses of chromosomes 1, 2, 6, 10, 17 and 21 12151617. FISH was considered to be a potentially useful tool for distinguishing RO from ChRCC, with the former showing no abnormality or loss of chromosome 1 and the latter usually multiple additional chromosomal aberrations 1218.

Based on the current data on genetic abnormalities occurring in ROs, we have employed an abbreviated FISH panel for chromosomes 1, 2, 7 and 17 in our institution that would aid in their differential diagnosis. In this study, we present our experience with this FISH panel and its contribution in the differential diagnosis along with survival data of 73 patients with RO.

Renal oncocytomas are relatively uncommon neoplasms accounting for 3-7% of primary renal epithelial neoplasms 123. The majority of RO has distinctive cytoarchitectural features that permit accurate diagnosis; however, the occasional differential diagnostic challenge between RO and ChRCC has been repeatedly acknowledged in the literature 231819. Both renal oncocytoma and chromophobe renal cell carcinoma originate form the collecting duct intercalated cells and can therefore share similar histomorphologic, immunophenotypical and ultrastructural characteristics 2345, as well as gene expression profiles 6720. Renal oncocytomas can contain foci of atypical cells and show vascular space invasion or extracapsular extension not adversely affecting the overall excellent prognosis of this neoplasm 23; the metastatic potential of RO remains controversial 21.

Although ChRCC was found to have low metastatic potential and its prognosis and long term survival is considered to be excellent 22, ChRCC was also found to have a potential to recur, metastasize and undergo sarcomatoid transformation 89. The difficulty to distinguish the two neoplasms has been demonstrated by the broad array of ancillary studies employed so far routinely in the differential diagnosis such as electron microscopy, DNA flow cytometry, immunohistochemistry, cytogenetics, and FISH. A small proportion of renal epithelial neoplasms with eosinophilic cytoplasm cannot be definitively classified based on histology alone 19.

Based on published cytogenetic studies, ROs are genetically heterogeneous tumors. Fűzesi et.al proposed subclassification into 3 groups: the first one defined by translocation between 11q13 and other chromosomes, the second with a loss of 1/1p followed by loss of chromosome Y or 14, and the third with non-recurrent or no detectable aberrations 10. Prior FISH studies have detected a varying incidence of partial or total loss of chromosome 1 in RO ranging from 10 - 59% 111213, whereas loss of other autosomes was not detected. In contrast, ChRCCs were consistently found to have multiple recurrent chromosomal abnormalities, such as losses of chromosomes 1, 2, 6, 10, 13, 17 and 21 by FISH, conventional cytogenetics and comparative genomic hybridization 1215161822. No significant difference in chromosomal abnormalities was found between the classic and eosinophilic variant of ChRCC 12. Since both RO and ChRCC can show loss of chromosomes 1 and Y, it was hypothesized that a subset of ROs may progress to ChRCC with subsequent loss of chromosomes 2, 6, 10, 13, 17 and 21 23.

Based on the above data, we have employed an abbreviated FISH panel for chromosomes 1, 2, 7 and 17 to guide us in the differential diagnosis of renal epithelial neoplasms with eosinophilic cytoplasm. Whereas an isolated loss of chromosome 1 would favor the diagnosis of RO, multiple losses of chromosomes 1, 2 and 17 were expected in ChRCC cases. Although abnormalities involving chromosome 7 are not specific and have been reported in other types of malignancies, we included this chromosome since its trisomy (together with trisomy 17) was observed in papillary renal cell carcinoma 2425 that can rarely contribute to the pool of unclassifiable cases of renal epithelial neoplasms with eosinophilic cytoplasm.

In our current study, the incidence of chromosomal abnormalities in ROs (60/73 cases, 82%) was significantly higher than in the previous reports 101218. Brunelli et al. 12 analyzed 10 RO by FISH using centromeric probes for chromosomes 1, 2, 6, 10 and 17; only one case showed loss of signal for chromosome 1, and no other aberrations were detected. The criteria for chromosomal losses were similar (22 - 30% of cells with one signal vs. 30% in the current study); however, the number of nuclei examined in our study was lower (60 vs. 100 - 200), but sufficient for overall assessment. Only one similar case of an incidentally discovered RO with multiple chromosomal losses involving chromosomes 1, 2, 3, 6, 8, 9, 15, 17, 21 and 22 in a 55-year-old male was identified in the literature 26. The second most common genetic alteration in our subset of RO was the loss of signal for chromosome 17 (32/73, 44%); a common finding reported in ChRCCs and also observed in the control ChRCC subset in our study. Recently, a tumor suppressor gene associated with Birt-Hogg-Dubé syndrome was mapped to the pericentromeric region of 17p chromosome 27282930. Birt-Hogg-Dubé syndrome is a rare genodermatosis characterized by the development of dermatologic lesions (fibrofolliculomas, trichodiscomas, acrochordons), lung cysts with spontaneous pneumothorax, colonic polyps and also renal tumors, mainly hybrid oncocytic neoplasms with both ChRCC and RO morphology, ChRCCs, clear cell carcinomas, oncocytomas and papillary renal cell carcinomas 31.

A total of 3/73 RO cases showed loss of chromosome 1 and gain of chromosome 7 and/or 17; morphologically, 2 cases displayed characteristic alternating nested and organoid growth patterns; the third case had predominantly tubulocystic arrangement of typical oncocytes. All three were initially diagnosed as RO without any ancillary studies. Gains of chromosome 7 and/or 17 without any accompanying chromosomal loss was observed in 4/73 cases; histologic review confirmed the initial diagnosis of RO in all four cases; however, only some of the original histologic sections were available for review in 3 cases precluding a definitive diagnosis. RO cases with extracapsular or vascular space invasion usually showed multiple chromosomal losses or gains; however, one of the cases with perirenal soft tissue extension was normal by FISH. Long term follow-up of patients included in our study confirmed the excellent prognosis of RO as none of the patients died of disease, developed metastases or local recurrence. No association was found between overall patient survival and the extent of chromosomal abnormalities detected by FISH.

The incidence of chromosomal abnormalities in ChRCC was in agreement with previous studies 121516; 19/20 (95%) cases showed multiple losses involving chromosomes 1, 2 and 17 in different combinations; in addition, loss of chromosome 7 was observed in 6/20(30%) cases and 6/20 (30%) cases showed loss of signals for all chromosomes examined (Figure 2). A total of 3/20 (all classic ChRCCs positive for colloidal iron) had chromosomal losses associated with gains of chromosome 7 and/or 17. Only one case did not show any gains or losses; review of the slides revealed an eosinophilic variant of ChRCC with a strong cytoplasmic positivity for colloidal iron (Figure 3). Rare cases of ChRCCs with no chromosomal abnormality have been reported 1216. One patient in the control ChRCC group developed local recurrence 25 months after partial nephrectomy; the reminder of the patients was disease free or died of other causes.

Our FISH study of ROs, which is largest to date, was based on a series of well-defined oncocytomas diagnosed according to the current morphologic WHO classification criteria 1. Our results show that chromosomal abnormalities in ROs are common and include frequent losses of chromosomes 1 and 17, yet another feature that overlaps with ChRCCs. FISH was not found to be a useful diagnostic modality in the differential diagnosis of RO and ChRCC and did not alter the initial primarily morphology-based diagnosis of renal oncocytoma in any of the 73 cases. Our results suggest that RO and ChRCC are genetically related tumors and support the hypothesis that they might rather represent a spectrum of neoplasia than two distinct neoplasms. This finding is further supported by the concurrent RO and ChRCC as well as their hybrid form in patients with Birt-Hogg-Dubé syndrome. The significance of specific combinations of chromosomal abnormalities as well as the impact on prognosis, propensity to metastasize, recur and undergo sarcomatoid transformation remains to be further investigated.

The authors declare that they have no competing interests.

MD carried out the research studies, participated in the drafting of the manuscript. KMC, CRS and TLM carried out the FISH studies. MBA carried out the immunohistochemical studies. RD and SIB participated in the design of the study. AVP conceived of the study, helped draft the manuscript and participated in its design and coordination. All authors read and approved the final manuscript.