Skip to main content

Variants of Prostate Carcinoma

By Bahram Robert Oliai, M.D.


Although the overwhelming majority of prostate cancers are acinar adenocarcinomas, behaving in accordance to their various Gleason scores, peculiar variants account for 5-10% of prostatic carcinomas (4,9). Additionally, there are some variants of acinar adenocarcinoma (psuedohyperplastic and foamy gland carcinoma) which tend to confuse both clinicians and pathologists for various reasons. Given the previously mentioned statistics, we are bound to run into these unusual cases at some time in our career. In this focus newsletter we examine these variant prostate tumors with an eye to their prognostic and therapeutic implications.

Pseudohyperplastic and Foamy Gland Carcinoma

These variants of acinar adenocarcinoma are best characterized by a striking tendency to mimic benign processes such as adenosis and/or foci of crowded benign glands. The issue here is not so much one of bizarre or unusual histology, but of not completely missing a malignant diagnosis. Pseudohyperplastic carcinoma is composed of malignant glands masquerading as benign glands architecturally. At low power, the glands show features commonly associated with benignity including large, branched glands with papillary infoldings and even corpora amylacea (almost never seen in malignant processes, Fig 1a-b). At higher power, however, prominent nucleoli and enlarged nuclei reveal the true nature of the beast (Fig 1c). In these cases the differential diagnosis could also include high grade prostatic intraepithelial neoplasia (HGPIN). While pseudohyperplastic carcinomas tend to be Gleason score 3+3=6 cancers, some can be higher grade. In their review, Levi and Epstein reported 20 cases diagnosed on needle biopsy. Of the seven cases with subsequent radical prostatectomies, one was Gleason score 7 and three cases showed extraprostatic extension.

In contrast, so-called foamy gland carcinoma often has a more worrisome, infiltrative pattern with rather bland cytology. Foamy gland cancers are characterized by xanthomatous cytoplasm (reminiscent of the head on a Guinness Draft, Fig 2a-c) with small dark nuclei showing little in the way of cytologic atypism or nucleolar prominence. It is the infiltrative pattern that really gives foamy gland cancer away. Again, while the majority of these tumors tend to be Gleason score 3+3=6, there are high grade foamy gland cancers and ultimately behavior will be true to the Gleason score. In their review of foamy gland carcinomas (defined as cancers showing 20% or more foamy gland features), Nelson and Epstein studied 38 cases diagnosed on biopsy. Of the 15 with radical prostatectomy followup, 43% were Gleason score 7 with 11 overall showing capsular penetration, four with positive margins, and two with pelvic lymph node/soft tissue metastases (6).


Given that both these variants mimic benign processes it is not only helpful, but perhaps also prudent to perform ancillary immunoperoxidase stains (high molecular weight cytokeratin, p63, racemase) to confirm the diagnosis.


Ductal Adenocarcinoma

Previously known as “endometrioid adenocarcinoma” (due to a superficial resemblance to uterine endometrioid carcinoma), prostatic duct carcinoma is perhaps the most common prostate cancer variant. Histologically these tumors may well be associated with conventional acinar adenocarcinoma and tend to show two main patterns or a mix thereof. One consists of true papillary fronds with well established fibrovascular cores lined by high columnar cells exhibiting a variable degree of cytologic atypism and nucleolar prominence (Fig 3a-b). Sometimes these lining cells can be extremely bland causing diagnostic confusion with a benign process. Due to their delicate and exophytic nature, these fronds tend to fragment out of the core at biopsy, a clue to the diagnosis. A second pattern consists of an intraductal proliferation of large, back to back glands imparting a somewhat cribriform appearance (Fig 4). In both cases the surrounding stroma often appears altered or fibrotic.


Often the main differential consideration is high grade prostatic intraepithelial neoplasia (HGPIN). In contrast to HGPIN, duct carcinoma shows true fibrovascular stalks, fragments on core biopsy, markedly dilates the involved ducts, and shows a fibrotic stroma (1). At times this differential can be extremely difficult and basal cell markers only help when “dead negative,” as both duct carcinoma and HGPIN can show patchy staining.


Clinically, patients tend to be elderly (63-72 years) often presenting with obstruction, hematuria, and abnormal digital rectal exam (a function of the often extensive involvement of the gland at presentation). Classically situated centrally in the periurethral glands (but can originate in the peripheral and transition zones) the cysto/urethroscopic picture of duct carcinoma has been described as exophytic or wormlike, almost like a malignant octopus waving its tentacles in the verumontanum (4). These are aggressive tumors, correlating to a Gleason score 4+4=8. In an excellent review by Brinker et al., of 20 cases originally diagnosed on needle biopsy and treated with radical prostatectomy; 63% showed extraprostatic extension, 20% had positive margins, and 10% had seminal vesicle involvement. Additionally, these patients had large tumors and a shortened time to progression (1). Others have confirmed the high grade nature of prostatic duct carcinoma noting 5 year survival rates as low as 30% and metastases at presentation in up to 50% (9).

Small cell carcinoma

Small cell carcinoma is truly the “stone cold felon” of this group. 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 (Fig 5a-b). In approximately half of the reported cases there is a mixture of small cell and acinar carcinoma (4). Crush artifact can be prominent sometimes making the diagnosis difficult and masquerading as crushed lymphocytes. Small cell carcinoma of the prostate is almost always positive for neuroendocrine markers such as synpatophysin, CD56, sometimes chromogranin, stains with CAM5.2/low molecular weight cytokeratins, and even TTF1 (pulmonary maker).

Interestingly prostatic small cell carcinoma has been shown to have a pattern of gene expression quite similar to its pulmonary counterpart (4). PSA and PSAP immunostains are often negative and noncontributory. The main differential of small cell carcinoma of the prostate includes metastatic pulmonary small cell carcinoma (clinical 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 men (65-72 years) presenting with bladder outlet obstruction and/or disseminated disease. Some patients develop small cell carcinoma after therapy for conventional acinar carcinoma (4). As in pulmonary small cell carcinoma, small cell carcinoma of the prostate may manifest various paraneoplastic syndromes including hypercalemia, SIADH, hyperparathyroidism, thyrotoxicosis, Eaton-Lambert Syndrome and ectopic ACTH secretion. Patients do poorly as the average survival is only 9-17 months. These tumors should not be assigned a Gleason grade and are generally unresponsive to hormonal therapy. The best chance of any kind of remission seems to be with chemotherapy sometimes in combination with radiation (7).

Mucinous Adenocarcinoma

The diagnosis of mucinous adenocarcinoma of the prostate rests on strict adherence to three criteria. First, at least 25% of the tumor should show aggregates of cells floating in lakes of extracellular mucin. Second, no signet ring component or significant intercellular mucin should be present. Finally, an extraprostatic primary must be ruled out (Fig 6). When adhering to these criteria, this is an extremely rare tumor. One series from the Memorial Sloan Kettering Cancer Center showed that mucinous adenocarcinomas accounted for only 6/1,600 of their cases of prostatic carcinoma (2). Thankfully, these carcinomas do tend to retain immunoreactivity for both PSA and PSAP, making Immunohistochemistry a useful tool in ruling out extraprostatic primaries (i.e. colonic, bladder, lung, pancreatic). Generally these tend to be Gleason grade 4 carcinomas, however some Grade 3 tumors have been reported (2).


Due to the very limited number of cases reported, little is known about the behavior of these cancers, however there is evidence suggesting an aggressive disposition. Epstein and Leiberman reported six cases with a review of previously reported cases. In their analysis of 26 true mucinous carcinomas all but three showed eventual extraprostatic involvement. In addition, five patients developed extension into the bladder or seminal vesicles and 10 developed distant metastases (2). Finally, mucinous adenocarcinomas tend not to respond to exogenous hormone therapy or orchiectomy (2).

Signet Ring Cell Carcinoma

The classic signet ring cell adenocarcinoma is composed of cells which have their nucleus displaced by an intracytoplasmic mucin vacuole, hence the resemblance to a signet ring (Fig 7a-b). While Gleason grade 5 carcinomas can show single cells with signet ring morphology, signet ring cells should comprise more than 25% of neoplastic cells to render the diagnosis of signet ring cell carcinoma. It is imperative that metastatic involvement from another site (i.e. stomach, bladder, colon, etc.) be ruled out. In some cases an associated acinar adenocarcinoma is present. Immunohistochemistry can be of use in establishing prostatic origin as many of these carcinomas are PSA and PSAP positive.

Clinically these tumors are aggressive and behave like high grade acinar carcinomas. In one report, 41.2% had stage IV disease on presentation and patients had an overall 5 year survival of only 11.7% (3).

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

In the prostate sarcomatoid carcinoma is quite a rare entity with less than 60 cases reported (4). This is the prototype biphasic tumor consisting of both malignant appearing spindle cell (sarcoma like) and epithelial (carcinoma) components (Fig 8). In the vast majority of cases the carcinoma is high grade (mean Gleason score 9) acinar, with rarer cases cases of ductal, squamous/adenosquamous, and urothelial differentiation. The sarcomatous areas are most often composed of undifferentiated spindled cells similar to malignant fibrous histiocytoma. Osteosarcoma and chondrosarcoma are the most common heterologous elements (4).

The key to the diagnosis of sarcomatoid carcinoma is the identification of an epithelial component, to this end one can utilize various cytokeratin immunostains. In cases in which a conventional acinar adenocarcinoma is present the prostatic nature of the tumor is evident, however PSA and PSAP can occasionally be detected even in the spindled areas helping confirm prostatic origin in poorly differentiated cases. The main differential diagnoses would include: prostatic stromal tumors (i.e. STUMP, stromal sarcoma), Inflammatory Myofibroblastic Tumor (ALK positive), and even sometimes gastrointestinal stromal tumors (CD34, CKIT positive).

Patients tend to be around 70 years of age presenting with frequency, urgency, and nocturia with enlarged, nodular, hard prostates on DRE. Patients may also have a history of acinar carcinoma which has been treated by hormonal therapy and/or radiation. Cystoscopically, these often form polypoid masses. Prognosis is poor with most patients succumbing to their disease. Reported mean survivals range from 9.5 months-3 years.


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.

Lymphoepithelioma – like carcinoma (LELC)

LELC is an extremely rare variant of prostate carcinoma (only one case of primary LELC has been reported, 9)! It 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 showing vesicular chromatin and prominent nucleoli in a dense lymphoid background (host reaction) which can sometimes obscure the carcinomatous nature of the lesion (Fig 9). High power examination gives the impression of an undifferentiated malignant neoplasm. Additionally, if in doubt, one can use cytokeratin immunoperoxidase stains to bring out the epithelial nature of the tumor. Unlike nasopharyngeal carcinoma, these carcinomas have been consistently found to be EBV negative. The main differential diagnosis in LELC would include metastatic nasopharyngeal carcinoma, lymphoma, and involvement by an LELC of bladder origin (a more common site).


Squamous/Adenosquamous Carcinoma 

The histology of prostatic squamous carcinoma is no different than that of other sites, being composed of infiltrating nests, strands and sheets of cells with cytologic atypism. The hallmarks of squamous differentiation include: individual cell keratinization, intercellular bridges, and/or keratin pearl formation (Fig 10). In cases of poorly differentiated carcinomas in which squamous differentiation is suspected, positive cytokeratin 5 or 5/6 and p63 immunostaining can be confirmatory (for more on this please refer to June 2005 FOCUS on immunohistochemistry newsletter). PSA and PSAP are virtually always negative in these tumors. Adensquamous carcinomas most often show a transition between glandular and squamous differentiation. Other primary sites notorious for adenosquamous features include the lung, gallbladder, and pancreaticobiliary tract. In cases in which squamous features predominate, involvement from a urothelial primary should be excluded.


Clinically, these cancers often present with bladder obstruction, dysuria, and are quite large. Squamous carcinomas of the prostate are not related to HPV and squamous metaplasia, showing instead a propensity to develop after radiation or hormonal treatment for typical acinar adenocarcinoma, perhaps heralding the emergence of a divergent, resistant clone. Unfortunately these are aggressive with a mean survival of only 6-24 months (4).


Basal Cell/Adenoid Cystic Carcinoma

Basal Cell carcinoma of the prostate refers to various malignant basaloid tumors characterized histologically by large nests of cells with peripheral palisading and necrosis (much like basal cell carcinoma of the skin), florid basal cell hyperplasia like neoplasms, and a pattern morphologically very similar to adenoid cystic carcinoma of the salivary gland with basaloid cells surrounding microcysts containing basement membrane like material (Fig 11a-c). In some cases the main differential would include benign basal cell proliferations such as florid basal cell hyperplasia. Features that can be used to distinguish basal cell carcinoma from basal cell hyperplasia include infiltration between benign glands, extraprostatic extension, perineural invasion, and necrosis. Unlike acinar adenocarcinomas, basal cell carcinomas show strong and diffuse staining with high molecular weight cytokeratin.


Due to the rare nature of these tumors, there is little in the way of followup data. However, based on the relatively few patients with such data, the prognosis seems to be overall quite good with the majority showing no progression. Rare cancer specific deaths have been reported (4,9).

Urothelial Carcinoma

Urothelial carcinoma involving the prostate is generally detected in one of two settings. One at TUR in the form of either carcinoma in situ involving the prostatic urethra and sometimes the underlying prostatic ducts or as invasive carcinoma. When noted on prostate needle biopsy (often done for an elevated PSA, abnormal DRE, or ultrasound) urothelial carcinoma tends to portend a more serious and clinically aggressive situation, as this implies widespread prostatic involvement. Microscopically, urothelial carcinoma on needle biopsy can involve only the ducts/acini (intraductal spread), invade the prostatic stroma showing irregular, infiltrative nests cutting through a reactive, inflamed stroma, or a combination of both.

The main differential diagnosis in these cases is always with high grade prostatic acinar carcinoma. In contrast, urothelial carcinomas involving the prostate tend to show more stromalinflammation, apoptotic debris, greater nuclear pleomorphism, increased mitoses, and necrosis (Fig 12). If in doubt, the issue of prostate cancer vs. urothelial carcinoma can usually be resolved using a panel of immunoperoxidase stains including PSA, PSAP, high molecular weight cytokeratin, cytokeratin 7, and thrombomodulin. It is important first to think of the possibility that a high grade carcinoma involving the prostate could be urothelial to be able to make this diagnosis. As we will see, urothelial carcinoma involving the prostate carries much different treatment implications for the patient.


Several years ago, while a resident at The Johns Hopkins Hospital, my attending Jonathan Epstein and I wrote up a series of urothelial carcinomas diagnosed on needle biopsy (8). Of the 21 cases we reviewed clinical information was available in 17. Patients had a mean age of 71 years (range 51-87) and presented with complaints/findings of hematuria, urinary obstruction, abnormal DRE, increased PSA, hematospermia, urinary frequency, and chronic UTI. Interestingly, many (41%) had no prior or subsequent history of urothelial carcinoma outside the gland, suggesting that these tumors can often develop within the prostatic ductal system. Of the rest, 35% had concurrent bladder cancer, 12% had a prior history, and 12% subsequently developed bladder cancer. Unfortunately we found that these patients did not do well, even in the absence of invasion. Overall 10/17 patients died of their disease (mean survival 23.2 months) including 6/9 with intraductal involvement only (the other 3 alive with disease), and 4/8 with invasive carcinoma (2/8 alive with disease, 2/8 alive with no evidence of disease).

In conclusion, urothelial carcinoma of the prostate is an aggressive tumor when diagnosed on needle biopsy, and even when stromal invasion is not identified (may be that due to the extensive nature of the carcinoma areas of invasion were under-sampled). Urothelial carcinoma can “hide out” within the gland unaffected by the various intravesical therapeutic agents. The only chance for improved outcome may therefore be early and aggressive therapy (i.e. cystoprostatectomy).

Conclusions/Summary Table

This month I have reviewed the major variants of prostate cancer. Undoubtedly with time, our knowledge of some of these “rare birds” will increase as more are reported. As a whole these are common enough that we will end up wrestling with at least some of them in the course of our careers, whether as a pathologist or urologist. Many of these variants impart important clinical information to urologists and oncologists. Table 1 represents a quick overview of these cancers and their key points.


1. Brinker DA, Potter SR, Epstein JI. Ductal adenocarcinoma of the Prostate Diagnosed on needle biopsy: Correlation with clinical and radical prostatectomy findings and progression. The American Journal of Surgical Pathology 1999; 23(12): 14711479.

2. Epstein JI, Leiberman PH. Mucinous adenocarcinoma of the prostate gland. The American Journal of Surgical Pathology 1985; 9(4): 299-308.

3. Fujita K, Sugao H, Gotoh T, Yokomizo S, Itoh Y. Primary signet ring cell carcinoma of the prostate: Report and review of 42 cases. International Journal of Urology 2004; 11: 178-181.

4. Humphrey PA. Unusual carcinomas of the prostate. Presentation, The Pathology of Prostate Cancer: From population studies to the Molecule, XXVI International Congress of the International Academy of Pathology, Montreal, Canada, 9/18/06.

5. Levi AW, Epstein JI. Pseudohyperplastic prostatic adenocarcinoma on needle biopsy and simple prostatectomy. The American Journal of Surgical Pathology 2000; 24(8): 1039-1046.

6. Nelson RS, Epstein JI. Prostatic carcinoma with abundant xanthomatous cytoplasm: Foamy gland carcinoma. The American Journal of Surgical Pathology 1996; 20(4): 419-426.

7. Oesterling JE, Hauzer CG, Farrow GM. Small cell anaplastic carcinoma of the prostate: A clinical, pathologic, and immunohis tological study of 27 patients. The Journal of Urology 1992; 147: 804-807.

8. Oliai BR, Kahane H, Epstein JI. A clinicopathologic analysis of urothelial carcinomas diagnosed on prostate needle biopsy. The American Journal of Surgical Pathology 2001; 25(6): 794-801.

9. Randolph TL, Amin MB, Ro JY, Ayala AG. Histologic variants of adenocarcinoma and other carcinomas of the prostate: Pathologic criteria and clinical significance. Modern Pathology 1997; 10(6): 612-629.



Special thanks to Dr. Jonathan Epstein for his permission in using images from Epstein JI, Yang Ximing. Prostate Biopsy Interpretation, Philadelphia: Lippincott, Williams & Wilkins, 2002.


Date of last revision: September 2009


Variants of Urothelial Carcinoma


The Process of Diagnosing Prostate