Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Renal oncocytomas are classified as benign renal neoplasms in the 2004 World Health Organization classification of renal tumours.1 Cases of metastatic tumours have been documented,2–4 but subsequent reports have identified subtypes of renal tumours having similar histological morphology but having malignant potential. These tumours have been classified as either eosinophilic variants of renal cell carcinoma (RCC) or chromophobe carcinomas. Differentiating these tumours from oncocytomas has proved to be difficult and relies on a combination of histological morphology, immunohistochemistry and electron microscopy. There is still a presumption that if metastasis occurs then the tumour should not be classified as a renal oncocytoma and should be called an eosinophilic RCC.
We present a patient who presented with liver metastasis 9 years after the removal of a renal oncocytoma.
A woman in her 70s presented with left-sided abdominal pain. She was noted to have a mobile mass in the left upper quadrant. Baseline blood examinations were normal and she had no other significant medical history. An abdominal ultrasound was suggestive of a left RCC. A CT scan showed a 12 cm renal mass extending just across the midline, arising from the upper anterior part of the left kidney (fig 1). The left renal vein, liver and lymph nodes appeared unaffected. The conclusion was that this was most probably an RCC, but the pattern of enhancement was suggestive of a renal oncocytoma. A chest x ray was clear.
She underwent a left radical nephrectomy and adrenalectomy en bloc. There was no involvement of adjacent organs. She made an uneventful recovery. She remained well on follow-up, with normal renal function and full blood counts. She was discharged from follow-up 18 months after surgery.
After 9 years she re-presented with left-sided abdominal pain, bloating and nausea, but no vomiting. A chest x ray was normal and an abdominal x ray showed faecal loading; this was thought to be diverticular disease. An abdominal CT scan showed multiple liver masses.
An ultrasound-guided liver biopsy was performed.
A CT scan 4 months later showed multiple lesions throughout the liver, which enhanced in the arterial phase, consistent with multiple hypervascular metastases (fig 1). The tumours appeared much the same as on the previous scan. Her liver function tests were normal. She had few symptoms, consisting of some tiredness, minor weight loss but a good appetite until 18 months after she re-presented when she developed confusion. A CT scan of her head showed acute left parietotemporal haemorrhage with intraventricular extension, but there was no evidence of metastatic tumour. She remained an inpatient for 4 months during which her renal and liver function tests were normal. She was discharged to a nursing home where she died a week later. A postmortem examination was not carried out.
The renal tumour was in the upper and mid-zones of the kidney and measured up to 12 cm in diameter. The tumour was yellow–brown in colour with a central white stellate scar, 1.5 cm in diameter. The perinephric fat, renal vein and adrenal gland were uninvolved. Histology showed packets of regular cells with eosinophilic cytoplasm and occasional highly cellular trabecular areas (fig 2). There was no cellular atypia, and mitotic figures were not seen (0 per 1 mm2). MIB-1 labelling index was <1%. Hales colloidal iron was negative. The conclusion was that this was a renal oncocytoma.
The liver biopsy specimen showed cores of liver parenchyma with well defined islands of oncocytic cells arranged in nests (fig 3). The original renal tumour was reviewed in light of the patient presenting with metastatic disease, and further immunohistochemistry was performed both by our laboratory and by Professor Susani in Vienna.
The renal tumour expressed vimentin and pan-cytokeratin marker (AE1/AE3), and there was focal positivity for epithelial membrane antigen (EMA; fig 2) and Ber EP4. CD10 showed very weak patchy positivity and the tumour was negative for c-kit (CD117) and cytokeratin 7. Limited tissue was present in the liver cores, but the tumour was also AE1/AE3 and EMA positive (fig 4). Professor Susani performed immunohistochemical analysis for the novel marker kidney-specific cadherin (Ksp cadherin), and both the renal tumour and the metastasis showed <10% marking, which is consistent with a renal oncocytoma.5
Electron microscopy on the renal tumour showed cells containing large numbers of mitochondria (fig 2), a finding characteristic of oncocytomas. By contrast, ultrastructural examination of chromophobe carcinomas would typically reveal numerous vesicles. Liver biopsy tissue was insufficient to perform electron microscopy.
There have been extensive reports on the biological behaviour and histological characteristics of renal oncocytomas since they were first described over 50 years ago.3,4 Since then, only rare cases of metastatic oncocytomas have been described and none has been studied with the modern panel of immunohistochemistry utilised in this case report.
Carcinomas and benign epithelial tumours are characterised by the expression of cytokeratins. Currently, 20 distinct subtypes of cytokeratins are known, and the expression profile of these can help classify tumours. However, there is often a degree of overlap between tumours—for example, Skinnider et al6 found that 10% of renal oncocytomas expressed cytokeratin 7 as opposed to 73% of chromophobe carcinomas. Another marker studied is c-kit, which encodes the membrane-bound tyrosine kinase KIT, and in a recent study was found to be expressed by both chromophobe and oncytomas, but was not detected in most other subtypes.7
Pan et al8 studied a variety of markers and found EMA to be expressed similarly by oncocytomas and chromophobe carcinomas, and BerEP4 to be negative in all but one renal oncocytoma while positive in all but two chromophobe carcinomas. CD10 was negative in all the renal oncocytomas they tested, but was negative in 68% of the chromophobe carcinomas. Vimentin was negative in all the oncocytomas and in 79% of the chromophobe carcinomas. This study was limited as only 7 oncocytomas and 28 chromophobe carcinomas were examined.
A larger study was carried out by Mazal et al5 examined the novel marker kidney-specific cadherin. This is a cell adhesion molecule found on the basolateral membrane of renal tubular epithelial cells and collecting duct cells. This study involved a variety of renal tumours but included 31 oncocytomas and 30 chromophobe tumours. The conclusion was that kidney-specific cadherin was expressed almost exclusively in chromophobe tumours.
The limitation of these immunohistochemical studies is that the tumours have to be classified on the basis of morphology before the immunopanel is performed, but there is often disagreement among pathologists when chromophobe carcinomas are examined.9
Adequate sampling of renal tumours is essential to exclude any heterogeneity, and in this case eight blocks from the renal tumour were examined. This may mean that an atypical area was missed but the metastatic tumour had the same morphology and immunoprofile as the tumour areas sampled, suggesting that this was not the case. DNA microsatellite analysis could be used to confirm this, but liver biopsy tissue was insufficient for this to be performed.
With an ever-expanding repertoire of immunohistochemical markers available, these have been applied to renal tumours in order to differentiate renal oncocytomas from malignant renal neoplasms. Examining the tumours at the ultrastructural level and looking for numerous mitochondria still remains the key to the diagnosis of renal oncocytomas. This case highlights that using these techniques can help classify but cannot predict malignant potential. With this in mind, renal oncocytomas should be considered as having a very low rather than no malignant potential.
We thank Professor Martin Susani (Professor of Pathology, Department of Clinical Pathology, AKH–Wien, Vienna, Austria), and his staff for reviewing and performing immunohistochemistry on this case.
Competing interests: None declared.