Papillary renal cell carcinoma (RCC), defined histologically as a malignant epithelial tumor of the kidney with a minimum of 50% papillary architecture1 and, more recently, by a combination of histologic and cytogenetic features,2 represents between 7% and 15% of renal carcinomas.1, 3 Compared with other subtypes of RCC, papillary RCC has distinctive pathologic and clinical features.1, 3-7 Tumors with low nuclear grade and low stage are associated with an excellent prognosis;1, 3, 6, 8 however, patients with high grade tumors have a more guarded prognosis. Compared with clear cell RCC, papillary RCCs are more often multifocal,9 yet this feature is predicted radiologically in less than half of cases.10 Accurate cytologic diagnosis may still be important for clinical management, particularly in considering a partial versus a radical nephrectomy.
There have been few reports regarding the fine-needle aspiration (FNA) diagnosis of papillary RCC.11, 12 Several criteria have been emphasized, including psammoma bodies, nuclear grooves, intracytoplasmic hemosiderin, foamy macrophages, and papillary tissue fragments. However, the relative sensitivity and specificity of these criteria are unknown. Herein, in an attempt to define further the most sensitive and specific cytologic criteria of this tumor, we report the cytologic features of 17 papillary RCCs and 52 nonpapillary tumors.
MATERIALS AND METHODS
- Top of page
- MATERIALS AND METHODS
Seventeen cytologic specimens from patients with papillary RCC who had their tumors resected were identified in the files of Brigham & Women's Hospital and Children's Hospital, Boston. None of the 8 cases previously reviewed are included.13 All the resected primary tumors were composed of greater than 50% papillary architecture or were otherwise histologically and cytogenetically characteristic of papillary RCC.14 Cytologic material from 15 FNAs performed under radiologic guidance for a renal mass or during intraoperative consultation for a renal mass was reviewed. Two specimens were touch preparations performed during intraoperative consultation. Alcohol-fixed, Papanicolaou-stained direct smear preparation (Papanicolaou-stained ThinPrep, Cytyc Corporation, Boxborough, MA), air-dried, Diff-Quik-stained smears, and hematoxylin and eosin-stained cell block and direct smear material were examined. Of the 15 FNAs, both direct smears and ThinPrep preparations were made for 10, ThinPrep preparations were made for 3, and direct smears were made for 2. The following cytologic features were examined: cellularity, presence of clusters, single cells, papillary fragments, fibrovascular cores, spherules, necrosis, blood, intracytoplasmic hemosiderin (within tumor cells), macrophages, psammoma bodies, nuclear pleomorphism, nuclear pseudoinclusions, nuclear grooves, nucleoli, and cytoplasmic vacuoles. For comparison, the same cytologic features were examined in cytologic specimens (intraoperative aspirations and touch preparations) from histologically confirmed clear cell renal RCC (40 cases), oncocytoma (3 cases), chromophobe RCC (2 cases), metanephric adenoma (1 case), and angiomyolipoma (6 cases). Again, none of these cases previously reviewed are included.13
- Top of page
- MATERIALS AND METHODS
Papillary RCC represents between 7% and 15% of renal carcinomas.1, 3 By histologic definition, papillary RCC must be composed of at least 50% papillary architecture.1 More recently, a combination of histologic and cytogenetic criteria have been used to define this tumor.14 The most compelling evidence that papillary RCC is a distinct variant of RCC has been the demonstration of unique chromosomal abnormalities. Deletions in the short arm of chromosome 3 are identified in the majority of clear and granular cell RCCs, but are not observed in papillary RCC.4 Instead, a characteristic set of chromosomal trisomies, most commonly involving 7, 16, and 17, have been documented by karyotyping.5 Papillary RCCs are also distinguished by their immunohistochemical phenotype. Like other subtypes of RCC, papillary RCCs are immunoreactive for epithelial membrane antigen (EMA) and low molecular weight keratins. Unlike other RCCs, however, papillary RCC routinely exhibits strong immunoreactivity with antibodies to high-molecular-weight callus keratins (DAKO)7 and CK7.15 This observation applies to both low grade and high grade papillary RCCs. Papillary RCCs also have characteristic radiologic findings; they are often cystic and necrotic, and may appear hypodense on CT scan. Angiographically, papillary RCC often appears avascular or hypovascular.3
In 1976, Mancilla-Jimenez et al. published the first large series of papillary RCCs (34 cases) and found they were predominantly low stage tumors (85% Robson Stage I [AJCC Stage I and II]), and the overall 5-year survival was 84%.1 Mydlo and Bard studied 39 cases of papillary RCC and reported similar 5-year survival rates for both Stage I (87%) and Stage II (84%) patients, but a lower rate for Stage III patients (50%).3 More recently, Lager et al. found that tumors with low nuclear grade were associated with 100% disease specific 5-year survival, irrespective of stage; however, survival for patients with with high grade tumors was only 50% at a mean of 4 years of follow-up.8 In the largest series to date, which contained 62 papillary RCCs, pathologic stage was the only independently significant predictor of clinical outcome. However, there was a reasonable correlation between stage and nuclear grade.6 Based on these data, the prognosis of patients with low grade/low stage papillary RCC at 5 years of follow-up is excellent-probably even better than that for patients with comparable clear cell tumors. Patients with high grade/high stage papillary RCCs have a much more guarded prognosis.
Despite the widespread acceptance of papillary RCC as a distinct variant of RCC, very few reports about its cytologic features have been published.11, 12 In one of these series, there was no surgical pathology follow-up to confirm the FNA diagnosis.11 Furthermore, the sensitivities and specificities for different cytologic characteristics were not defined. In these studies, the authors concluded that papillary RCCs have distinctive cytologic features that allow for "confident diagnosis." Our experience, prior to examining this series, contrasted with these previous reports. In a study of accuracy of FNA in distinguishing subtypes of RCC, we found that the most common pitfall was misclassifying a papillary RCC, usually a high grade tumor, as a clear cell or sarcomatoid RCC.12 We undertook this survey of additional papillary RCCs to examine the cytologic spectrum of papillary RCC critically and define the most sensitive and specific criteria for its diagnosis.
The cytologic features emphasized in cytology texts and frequently referred to as "classic features" of papillary RCC were often absent in this series. For example, in this study, psammoma bodies were present in only one case and intranuclear pseudoinclusions were present in only three cases. Dekmezian et al. concluded that "the most distinguishing features, however, are nuclear; nuclear grooves are frequently evident, nuclear pleomorphism is unusual, and nucleoli are inconspicuous."12 Nuclear grooves, one of the most touted features of papillary RCC, were common in only 2 of our papillary RCC cases (they were rare in 6 and altogether absent in 7 cases) but were seen in 25% of other tumors. Despite the conspicuous absence of some "classic features" of papillary RCC, the diagnosis of low grade papillary RCC was not difficult when the entire spectrum of changes was evaluated. The malignant cells were often arranged in papillary groups with fibrovascular cores. The nuclei were bland and contained fine chromatin. Foamy macrophages and prominent intracytoplasmic hemosiderin were often seen. In this series, the most sensitive features for the diagnosis of low grade RCC were foamy macrophages and intracytoplasmic hemosiderin, with sensitivities of 82% and 76%, respectively. Both of these criteria were 96% specific. Although very helpful when present, nuclear grooves and psammoma bodies appear to be less sensitive and specific criteria for the diagnosis of papillary RCC.
Compared with low grade papillary RCC, the diagnosis of high grade papillary RCC is more challenging. Seven of 17 tumors in this series were high grade tumors. In the previous two cytologic series, Furhman grading was not reported, and these studies neither described nor illustrated high grade papillary tumors.11, 12 Papillary fragments with fibrovascular cores, which are perhaps the most specific cytologic feature of papillary RCC, were seen in only one high grade tumor in our series. Together with pleomorphism and frequent, prominent nucleoli, these tumors might easily be mistaken for a high grade clear cell variant of RCC (Fig. 5 (26K)). The most sensitive features for the diagnosis of high grade papillary RCC in this series were again intracytoplasmic hemosiderin deposition (present in 6 of 7 high grade tumors) and foamy macrophages (present in 6 of 7 cases). However, one case, a Grade 4 tumor, lacked papillary fragments, hemosiderin, and macrophages. This case serves to illustrate that accurate subtyping of some high grade papillary RCCs is probably not possible. However, it should be pointed out that making a distinction between high grade papillary RCC and clear cell RCC may not be as important as with low grade tumors, because the prognosis may not differ significantly.
Figure 5. Some high grade papillary renal cell carcinomas (RCCs) are difficult to distinguish from the clear cell variant of RCC. We found the presence of foamy macrophages and intracytoplasmic hemosiderin to be the most sensitive diagnostic features of papillary RCC (Papanicolaou stain, original magnification x800).
Download figure to PowerPoint
The presence of hemosiderin is fairly sensitive for the diagnosis of papillary RCC. We have also demonstrated that hemosiderin is relatively specific as well. Only 5% of clear cell RCCs showed intracytoplasmic hemosiderin. That intracytoplasmic hemosiderin is seen only rarely in clear cell RCC is also reflected in the scarcity of reported clear cell RCCs with this feature.16 Interestingly, there is cytogenetic evidence raising the possibility that these tumors are a distinctive subset of RCC and may be more closely related to papillary tumors than clear cell tumors at the cytogenetic level.16 It is important to emphasize that the intracytoplasmic hemosiderin specifically relates to tumoral hemosiderin, not hemosiderin within macrophages. Tumors that have intracytoplasmic (tumoral) hemosiderin are usually associated with hemosiderin within macrophages. We consider hemosiderin within macrophages less specific for the diagnosis of papillary RCC. It should also be kept in mind that pigment that closely resembles neuromelanin has been observed in a case of clear cell RCC and may be mistaken for hemosiderin and inadvertently lead to a misdiagnosis of papillary RCC.17 One criterion deserves further comment. As illustrated in Figure 6 (79K), the papillarylike structures with ragged irregular outlines were seen in 27.5% clear cell RCCs and are indistinguishable from similar structures seen in true papillary RCC. We believe that the presence of fingerlike papillae with bulbous, rounded ends and spherules are a more reliable criteria of papillary RCC. Spherules are simply the detached ends of these blunt-ended papillae. We did not observe these structures in any of the 40 clear cell RCCs in this study.
In conclusion, low grade papillary RCC has distinctive cytologic features that facilitate a specific diagnosis based on FNA. High grade papillary RCCs, however, may be difficult or impossible to distinguish from high grade clear cell RCCs by FNA, even with adequate cellularity. Intracellular hemosiderin and foamy macrophages are the two most helpful features in making this distinction. In difficult cases, if cell block material is available, immunohistochemical studies may be helpful in distinguishing papillary RCC from clear cell RCC.