Skip to main content

Variants of Urothelial Carcinoma

By Bahram Robert Oliai, M.D.


Although most of us are familiar with the features, behavior and treatment of the “usual” type urothelial (so-called transitional cell)carcinoma, what do we do when the pathology report comes back with an unexpected and unusual diagnosis? For example, what are the connotations/nuances behind a diagnosis of “nested variant of urothelial carcinoma,”“micropapillary carcinoma,” and “lymphoepithelioma – like carcinoma,” among others? Are these variants just curiosities among pathologists that we throw around to boost our egos and make ourselves look smart to our clinical colleagues like peacocks strutting our vividly colored diagnostic feathers, or are these indeed important variants with real clinical significance maybe even pertaining to prognosis and treatment?

In this newsletter I will review the salient points of these peculiar entities. Please join me on this tour of the “zebra park,” and hopefully in the end we will feel more comfortable when confronted by these most unusual and often unexpected diagnoses.

Nested variant of urothelial carcinoma

Nested variant of urothelial carcinoma is the true “wolf in sheep’s clothing” of bladder tumors. Much like the murderous, but mild mannered neighbor who everyone describes as a“regular guy,” this tumor is composed of remarkably bland “nests” of urothelial cells with an appearance mimicking von Brunn nests (Fig 1a). These nests can also form tubules and microcystic structures and the main differential diagnoses for this tumor would include such entities as cystitis cystica/glandularis, inverted urothelial papilloma, nephrogenic adenoma, carcinoid tumor, and paraganglioma. The diagnosis of carcinoma is usually suggested or made not by observing frank cytologic atypism, but by noting that this proliferation is very florid, consists of urothelial nests which are smaller and more irregular than von Brunn nests or cystitis cystica, and appears to “drop off the bottom” of the tissue invading often deeply into muscularis propria, a place no benign urothelial proliferation would dare be seen (Fig 1b, 1c). Some nests do show cytologic atypism, but this is often found very deeply within the tumor. The key point is that often a definitive diagnosis is very difficult to render unless muscularis propria invasion is noted in the biopsy, stressing the importance of deep biopsy in these cases. In cases showing a marked nested urothelial proliferation in a superficial biopsy, I may comment on the findings, suggest the possibility of nested variant of urothelial carcinoma and recommend additional, deeper, tissue sampling.

Clinically these tumors show a marked male predominance, are “bad actors” and should be treated aggressively. In one reported series, 70% of patients with this diagnosis eventually died of their disease, despite aggressive treatment (6).

In summary, from the pathologist’s point of view, nested variant of urothelial carcinoma can be a very difficult diagnosis to make, especially if deep biopsies with underlying muscularis propria are not obtained. Clinically, this variant needs aggressive surgical intervention and can be a killer.

Micropapillary carcinoma

This diagnosis refers to a histologic pattern of both in situ and invasive carcinoma. On the bladder surface the in situ component consists of slender, delicate, filiform structures like seagrass waving in azure (or perhaps yellow in this case) waters (Fig 2a). Invasive micropapillary carcinoma is composed of tight clusters of micropapillary structures often within lacunae infiltrating deep within the bladder wall and often showing lymphovascular invasion (Fig 2b). Micropapillary carcinoma is commonly associated with more typical papillary and nonpapillary urothelial tumors. The key differential diagnosis in these cases (especially when a more common variant of urothelial carcinoma is not present in the background) is that of metastasis especially from ovarian (serous) carcinoma, breast carcinoma, and mesothelioma, all of which can show micropapillary features.

Clinically, there is a male predominance (M:F ratio 5:1), and the mean age at diagnosis is 67 years. This is considered an aggressive or high grade carcinoma. In one series 8/18 patients had metastatic disease on presentation, 1/18 had local recurrence after therapy, and tumor invasion into pelvic structures was noted in 3/18 cases. On followup (6-96 month interval), 7 patients died of their disease, 4 were alive with their disease, and 7 were alive with no evidence of disease (1).

In summary, from a diagnostic standpoint metastastic involvement of the bladder, especially from ovarian carcinoma, breast carcinoma, and mesothelioma needs to be ruled out. If more than 10% of a bladder carcinoma shows a micropapillary pattern, this should be mentioned in the report as more than 10% micropapillary pattern imparts a poor prognosis. Clinically these patients often present at high stage and this is an aggressive variant with poor prognostic implications.

Lymphoepithelioma – like carcinoma (LELC)

LELC is a rare variant of urothelial carcinoma that bears a striking resemblance to undifferentiated nasopharyngeal carcinoma, so-called “lymphoepithelioma.” As such, the microscopic picture is that of nests, sheets, cords, or single malignant cells with large nuclei with vesicular chromatin and prominent nucleoli in a dense lymphoid background (host reaction) which can sometimes obscure the carcinomatous nature of the lesion (Fig 3). High power examination gives the impression of an undifferentiated malignant neoplasm, however clues to the epithelial/urothelial nature of this tumor include a background of carcinoma in situ or conventional uorthelial carcinoma. If in doubt, one can use cytokeratin immunoperoxidase stains to confirm the epithelial nature of the tumor. Unlike nasopharyngeal carcinoma, these bladder carcinomas have been consistently found to be EBV negative. The main differential diagnosis in LELC would include metastatic nasopharyngeal carcinoma, lymphoma, usual urothelial carcinoma or poorly differentiated squamous carcinoma with dense lymphoid stroma.

Clinically, these carcinomas are more common in men generally occurring in late adulthood and tend to involve the bladder with rare cases involving the renal pelvis and urethra. Of the 30 cases reviewed in the largest series reported, 17 (57%) were composed solely of LELC, with the remaining associated with more conventional types of urothelial carcinoma including flat carcinoma in situ, high grade papillary non-invasive carcinomas, adenocarcinoma (in situ and invasive), squamous carcinoma, and typical invasive urothelial carcinoma (8).

Although it was originally thought that in pure form these tumors behaved better than conventional urothelial carcinomas and may be chemoresponsive allowing for the potential to salvage bladder function, most of the original studies were based on small series and more recent data suggests that in patients treated by cystectomy, the prognosis is quite similar to that of conventional urothelial carcinoma and did not differ significantly between pure and mixed cases. Of the 3 pure cases treated by chemotherapy, 2 were free of disease at 4 and 65 months and one had recurrent disease at 17 months. Additionally, given that 47% of cases of LELC in the largest series were associated with more typical urothelial carcinoma, partial cystectomy and/or chemotherapy may not be adequate treatment for a proportion of patients with LELC (8).

Sarcomatoid carcinoma (Carcinosarcoma) A.K.A. “Spindle cell carcinoma” A.K.A. “Metaplastic carcinoma”

This is the prototype biphasic tumor consisting of both malignant appearing spindle cell (sarcoma like) and epithelial (carcinoma) components. As such, any combination of sarcoma (in order of frequency – osteosarcoma, chondrosarcoma, rhabdomyosarcoma, leiomyosarcoma, liposarcoma, angiosarcoma) and carcinoma (i.e. conventional urothelial, squamous cell carcinoma, even small cell carcinoma) may be seen (Fig 4). Not uncommonly, the surface may show carcinoma in situ, which may be the only epithelial component present. The key to the diagnosis of sarcomatoid carcinoma is the identification of an epithelial component, to this end one can utilize various cytokeratin immunostains. The main differential diagnoses would include: bladder sarcoma (i.e. leiomyosarcoma, etc.), urothelial carcinoma with reactive stroma, Inflammatory Myofibroblastic Tumor (ALK-1 positive), and even nerve sheath tumors of the bladder (S100, Sox-10 positive, markers of neural crest differentiation).


Clinically, these patients often have a history of carcinoma which has been treated by cyclophospamide and/or radiation. Cystoscopically, these often form polypoid masses. Although these tumors do not demonstrate a difference in survival when compared to usual urothelial carcinoma stage for stage, virtually all sarcomatoid carcinomas present at high stage – thus having a poor prognosis.

When I make this diagnosis I call these tumors “Sarcomatoid carcinoma (carcinosarcoma).” Additionally, I mention the various carcinomatous and sarcomatoid components describing any heterologous differentiation (i.e. osteosarcoma, chondrosarcoma, etc.) so that others reviewing future specimens (especially metastases) may be alerted to the components of the original lesion – as one or both components can metastasize.

Small cell carcinoma

Just as pulmonary small cell carcinoma is notorious in the world of oncology for its aggressive behavior, small cell carcinoma of the bladder should connote the same aggressive behavior. It shows the same morphology and behavior as its more common pulmonary “cousin,” representing the “stone cold felon” of bladder tumors. The diagnosis is based mainly on cytomorphology often backed up with Immunohistochemistry. Small cell carcinoma is composed of an infiltrate of small relatively uniform cells exhibiting nuclear molding, stippled chromatin, and inconspicuous nuclei. Many of the cells are proliferating and frequent mitotic figures are noted. Often there is a background of carcinoma in situ, adenocarcinoma in situ, a high grade papillary tumor, and/or conventional infiltrating urothelial carcinoma (Fig 5). Crush artifact can be prominent sometimes making the diagnosis difficult, since the cells may masquerade as crushed lymphocytes. Small cell carcinoma of the bladder is almost always positive for neuroendocrine markers such as synpatophysin, CD56, sometimes chromogranin, stains with CAM5.2/low molecular weight cytokeratins in a characteristic perinuclear dotlike pattern, and even TTF1 (pulmonary maker) focally. The main differential of small cell carcinoma of the bladder includes metastatic pulmonary small cell carcinoma (diffuse TTF1 positivity, history of lung lesion), lymphoma (can be a very difficult differential on morphology alone; LCA, CD3 or CD20 positive), metastatic Merkel cell carcinoma (perinuclear dot like cytokeratin 20 staining), and crushed/cauterized lymphocytes.


Clinically, these patients tend to be older males (M:F 3.3:1, mean age 66) and as noted above small cell carcinoma is extremely aggressive – of 64 cases reported in one series, all but 1 had muscularis propria invasion at presentation. Of the patients who underwent cystectomy, 66% had metastasis (5). There is data suggesting that cystectomy alone may not confer a survival advantage in these patients. Overall 5 year survival is quite poor ranging from 16-40% varying among several series. As in pulmonary small cell carcinoma, small cell carcinoma of the bladder may manifest various paraneoplastic syndromes including hypercalemia, hypophosphatemia, and ectopic ACTH secretion.

Optimal treatment may include adjuvant chemotherapy with a platinum based regimen, especially for high stage patients (3). A paper out of M.D. Anderson does offer some hope, reporting that a preoperative “neuroendocrine type” regimen (4 cycles of etoposide/cisplatin or isofamide/doxorubicin) was successful in downstaging 12 of 21 patients. In this study a 78% 5 year disease specific survival was achieved in 21 patients treated with a regimen of preoperative chemotherapy followed by cystectomy (7).

Giant cell/Large-cell undifferentiated carcinoma

The histology of giant cell carcinoma is that of an undifferentiated high grade malignant neoplasm, often barely recognizable as carcinoma. These are the kind of cases that pathologists look at and think “it’s definitely bad, but could be anything from anywhere.” Often the diagnosis of carcinoma is made only after cytokeratin immunoperoxidase stains are shown to be positive. The prototype tumor with this morphology is the so-called giant cell carcinoma of the lung. The cells are extremely large, cohesive, in some cases multinucleate, anaplastic cells with abundant amphophilic or eosinophilic cytoplasm (Fig 6). When the picture is that of nondescript undifferentiated carcinomatous cells without giant forms the diagnosis of large-cell undifferentiated carcinoma is made.


The main differential diagnosis would include metastatic giant cell carcinoma (i.e. lung, etc.), melanoma, and even direct extension from a prostatic giant cell carcinoma (which can be an extremely difficult differential diagnosis in the bladder neck or in a TUR specimen). The presence of a background conventional urothelial carcinoma or even another variant of urothelial carcinoma (squamous carcinoma, LELC, and micropapillary carcinoma have all been reported associated with pleomorphic giant cell carcinoma) can aid in confirming the impression of a bladder primary.

Little is known about this tumor, especially in its pure form. However, in a recent review of 8 cases representing the largest experience with pleomorphic giant cell carcinoma, the outcome was quite poor with 87.5% (or 7 of 8 patients) dying of disease or developing metastases within 2 years of diagnosis and only 1 patient surviving long term (74 months, 5).


Carcinomas showing unique cytoplasmic features:

Plasmacytoid urothelial carcinoma

Recently cases of urothelial carcinoma bearing a striking resemblance to plasma cells have been described (Fig 7). The main pitfall/differential diagnosis for pathologists in these cases is with hematolymphoid neoplasms such as myeloma/plasmacytoma, so-called MALT lymphoma and lymphoplasmacytic lymphoma. The key is to be aware of this variant and to use Immunohistochemistry to help arrive at a correct diagnosis. Plasmacytoid variant of urothelial carcinoma will be positive for various cytokeratins and negative with stains such as LCA (CD45), CD20, kappa, and lambda which are generally positive in myeloma/ plasmacytoma/lymphoma. Although CD138 is a marker often helpful in the diagnosis of plasma cell tumors, it can also stain epithelial tumors such as plasmacytoid urothelial carcinoma, if used results should be interpreted with caution and only in the context of the results of the other markers mentioned.


Patients tend to be males (M:F 2:1) with a mean age of 58 years. In one series 5 of 6 patients died with a mean survival of 23 months (6).

In summary, the key points are not to misdiagnose this variant as myeloma as the treatment is vastly different. Additionally, these tumors may behave very aggressively.

Clear cell (glycogen rich)/lipoid cell carcinoma variants

Clear cell urothelial carcinoma is similar to classic urothelial carcinoma, but shows striking clear cell change due to cytoplasmic glycogenation (Fig 8). The important point is not so much clinical, but for pathologists to be aware of this pattern and to avoid misdiagnoses such as metastatic renal carcinoma, clear cell adenocarcinoma of the urethra/bladder (generally a diagnosis of females), and paraganglioma.


Lipoid cell urothelial carcinoma is a rare variant exhibiting a transition from more classic urothelial carcinoma to cells that resemble lipoblasts (Fig 9). Of the few reported cases most patients are elderly men presenting with gross haematuria. The main differential diagnoses in these cases include carcinosarcoma with liposarcoma like elements and metastatic/primary liposarcoma involving the bladder (a very unusual diagnosis). The “lipoid cells” in lipoid cell urothelial carcinoma tend to stain strongly and diffusely with cytokeratin immunoperoxidase stains.


Urothelial carcinoma with “osteoclast like giant cells”

This rare variant of urothelial carcinoma shows a striking osteoclast like giant cell infiltrate which is intimately associated with the invasive carcinoma. The carcinoma is composed of relatively bland appearing ovoid or spindled mononuclear cells with evenly spaced osteoclast like giant cells. These multi-nucleated cells have been shown to be of histiocytic lineage like osteoclasts (most likely representing some strange host reaction) while the mononuclear cells have been shown to be cytokeratin positive representing the carcinomatous component of this lesion (Fig 10). Many of the reported cases have been associated with more typical papillary tumors or in situ carcinoma. Although little is known of the clinical significance of this tumor, one recent study of 6 cases (3 in the bladder, 3 renal pelvis) suggested aggressive behavior in the five cases with available followup (1 patient alive with disease, 4 succumbing within 15 months, 2).


Rhabdoid urothelial carcinoma

Rhabdoid urothelial carcinoma is an extremely rare variant showing eccentric nuclei with pink globular cytoplasmic densities (so-called rhabdoid features, Fig 11). In general tumors with rhabdoid features behave quite poorly regardless of site. The main differential diagnoses includes rhabdomyosarcoma, true rhabdoid tumor (generally only seen in the pediatric population), and rhabdoid areas within a carcinosarcoma. Cytokeratin immunostains are often necessary to confirm the carcinomatous nature of the tumor.


Urothelial carcinoma with syncytiotrophoblastic giant cells

At times, otherwise standard appearing urothelial carcinomas can be associated with syncytiotrophoblastic giant cells, which stain positively for B-HCG (Fig 12). As a general rule in these cases the number of syncytiotrophoblastic giant cells are inversely correlated with grade. These tumors may even secrete B-HCG into the serum. The main differential diagnosis would include choriocarcinoma (a rather unusual diagnosis in the bladder). In fact it may well be that many of the previously reported cases of choriocarcinoma of the bladder really represented urothelial carcinomas with syncytiotrophoblastic giant cells.


In conclusion there are an increasing number of urothelial carcinoma variants being described. Many of these have diagnostic points salient to pathologists and may impart important clinical information to urologists/oncologists. Table 1 is a quick review of these variants and their clinical behavior in relation to “usual type” urothelial carcinoma.



1. Amin M, Ro JY. El-Sharkawy T, Lee KM, Troncoso P, Silva EG, Ordonez NG, Ayala AG. Micropapillary variant of transitional carcinoma of the urinary bladder. Histologic pattern resembling ovarian papillary serous carcinoma. American Journal of Surgical Pathology 1994 Dec; 18 (12): 1224-1232.

2. Baydar D, Amin MB, Epstein JI. Osteoclast-rich undifferentiated carcinomas of the urinary tract. Modern Pathology 2006 Feb; 19(2):161-71.

3. Choong NW, Quevedo JF, Kaur JS. Small cell carcinoma of the urinary bladder. The Mayo Clinic experience. Cancer 2005 March; 103 (6): 1172-1178.

4. Epstein JI, Amin MB, Reuter VR. Bladder biopsy interpretation, Philadelphia: Lippincott, Williams & Wilkins, 2004.

5. Lopez-Beltran A, Blanca A, Montironi R, Cheng L, Regueiro JC. Pleomorphic giant cell carcinoma of the urinary bladder. Human Pathology. 2009 40: 1461-1466.

6. Lopez-Beltran A, Cheng L. Histologic variants of urothelial carcinoma: differential diagnosis and clinical implications. Modern Pathology. 2006 November; 19(11): 1395-1401.

7. Siefker-Radtke AO, Dinney CP, Abrahams NA, Moran C, Shen Y, Pisters LL, Grossman HB, Swanson DA, Millikan RE. Evidence supporting preoperative chemotherapy for small cell carcinoma of the bladder: a retrospective review of the M.D. Anderson cancer experience. The Journal of Urology. 2004 August; 172: 481-484.

8. Tamas EF, Neilsen ME, Schoenberg MP, Epstein JI. Lymphoepithelioma-like carcinoma of the urinary tract: a clinicopathologic study of 30 pure and mixed cases. Modern Pathology. 2007 20: 828-834. conventional urothelial carcinoma and did not differ significantly between pure and mixed cases. Of the 3 pure cases treated by chemotherapy, 2 were free of disease at 4 and 65 months and one had recurrent disease at 17 months. Additionally, given that 47% of cases of LELC in the largest series were associated with more typical urothelial carcinoma, partial cystectomy and/or chemotherapy may not be adequate treatment for a proportion of patients with LELC (8).


Special thanks to Dr. Jonathan Epstein for his permission in using images from Epstein JI, Amin MB, Reuter VE. Bladder Biopsy Interpretation, Philadelphia: Lippincott, Williams & Wilkins, 2004.


Date of last revision: February 2010


Using Fluorescent in Situ Hybridization (UroVysion™ FISH)


Variants of Prostate Carcinoma