Abstract
The differential diagnosis between high-grade prostate carcinoma and infiltrating urothelial carcinoma (UC) in transurethral resection prostate specimens as well as cystoprostatectomy specimens may often be challenging due to morphologic and clinical overlap of the 2 entities. Such distinction has critical therapeutic and staging consequences, yet it is hampered by both issues in morphology and by the low accuracy rates of single immunohistochemical markers, as reported in literature. This review aims to provide a comprehensive analysis of the available morphological and immunohistochemical parameters, which may allow to discriminate between prostate carcinoma and urothelial carcinoma in the proper clinical context and to discuss their diagnostic applications in daily practice.
Prostatic Involvement by Urothelial Carcinoma
The incidence of prostatic involvement by urothelial carcinoma (UC) of the urinary bladder in radical cystoprostatectomy specimens ranges from 12% to 51% according to several authors.1-9 UC may be detected in the prostate on needle core biopsies, transurethral resection of the prostate (TURP), simple prostatectomy, as well as radical prostatectomy and cystoprostatectomy specimens; the smaller the specimen, the more challenging to determine the site of origin of the tumor, especially in the absence of a known history of UC of the bladder. In addition, in some cases, prostatic UC may actually be arising from the prostatic urethra and not the bladder, with obvious staging implications; in such cases, appropriate tissue sampling is needed. 10
Two main mechanisms account for prostatic involvement by UC, with variable data on prognostic implications, as mentioned in several studies1-5,10-15: (1) colonization of prostatic ducts/acini by UC in situ of the prostatic urethra (with or without ductal carcinoma in situ) and (2) direct invasion of the prostatic stroma following tumor extension beyond the muscularis propria/detrusor muscle. In such scenario, the distinction between colonization of prostatic ducts/acini by urethral UC and prostatic stromal invasion by UC is crucial, since a stage of pT2 and pT4a is assigned, respectively, according to the latest 2017 American Joint Committee on Cancer/Tumor-Node-Metastasis classification. 14 This challenging differential diagnosis (especially in needle core biopsies or TURP specimens) often relies on subjective parameters related to stromal invasion, namely stromal desmoplastic response, host inflammatory response, and irregular urothelial nests; immunohistochemical stains have proven to be not helpful in some cases.10,15,16
Since prostatic involvement by UC is associated with worse survival outcomes and increased risk of harboring node metastases, 6 managing issues may arise in this setting. Some authors recommend to examine samples of prostatic urethra even in the presence of normal appearing mucosa, either before radical cystoprostatectomy 7 in order to choose the most proper type of urinary diversion, or by intraoperative frozen section of the apical margin in patients planning to undergo orthotopic neobladder reconstruction. 7
In the pursuit of predicting markers of prostatic involvement by UC, some studies have pointed out that distance between the tumor and the bladder neck had a significant value. 9
Establishing the nature of an undifferentiated carcinoma involving the prostate in cystoprostatectomy specimens (ie, whether it is of urothelial or prostatic origin) is essential in that patients with UC infiltrating the prostate carry a higher progression rate and lower survival compared with those with concurrent prostate carcinoma (PC). 17 A differentiation between UC and PC may be done in the metastatic setting as well, as in case of a collision nodal metastasis. Although rare, the event of UC and PAC metastasizing to the same lymph node represents the most frequent occurrence among collision metastases.18-21
Bladder Involvement by Prostatic Carcinoma
Because of its close proximity, PC frequently involves the bladder by direct invasion, particularly the neck, trigone and posterior wall, 22 sometimes protruding in the bladder lumen on the coronal plane during imaging exams, leading to a misdiagnosis of a primitive urothelial mass. 23
UC and PC may also coexist: in a large series of 1476 radical cystoprostatectomy specimens, PC was present in 37.9% patients, and clinically significant PC (featuring high Gleason score, extraprostatic extension, seminal vesicle and/or lymph node involvement, positive surgical margins) was found in 22% of them. 8 Such tumors are easily diagnosed as bladder primaries, especially when lacking a papillary UC or noninvasive flat component. 24 Even in cases in which a known history of PC is available, superimposed histologic changes, such as squamous metaplasia caused by prior radiation or hormonal therapy or poor differentiation, lead to difficulty in differentiating a primary UC from a recurrent PC on a needle biopsy or TURP. 25
A most unusual finding is the occurrence of PC arising from ectopic prostatic tissue in the bladder (and sometimes in the retrovesical space), with only sporadic cases reported in literature.26-29
Morphological Features
Usual morphological features which may help in discriminating UC from PC are summarized in Table 1. They are mainly cytological since architectural features cannot help in the distinction between the 2 entities, due to their high variability in each tumor.
Morphological Features for the Differential Diagnosis Between Prostate and Urothelial Carcinoma.
Abbreviations: CIS, carcinoma in situ; PC: prostate carcinoma; UC: urothelial carcinoma.
Immunohistochemistry
In cases that are challenging on hematoxylin and eosin, conventional immunohistochemical algorithms to assess the cell of origin in cancer of unknown primary recommend the use of CK7 and CK20 as first line markers, as UC is usually positive for both markers while PC lacks either. Subsequently, they suggest further classification by using markers of urothelial and prostatic origin in order to discriminate between PC and UC, the former being CK7+/GATA3+/PSA−, and the latter being CK7−/GATA3−/PSA+; other antibodies such as CK20 are less helpful since they may be expressed in both cancers.30,31 Additional immunohistochemical stains such as PSAP, NKX3.1, P501S, PSMA, and GATA3, P53, P63, HMWCK, thrombomodulin, uroplakin may be useful; the first set of markers is usually positive in PC, and the second in UC24,32-35 (see Tables 2 and 3). A recent update on the use of immunohistochemical markers in genitourinary tumors suggests a first panel including PSA, HMWCK, and P63, with PC being PSA+/HMWCK−/P63−, and UC being PSA−/HMWCK+/p63+. In case of equivocal results with these markers, additional antibodies such as P501S, NKX3.1 (for PC), and GATA3 (for UC) can be used.24,36,37 Moreover, technical issues (such as false positivity due to endogenous biotin, lack of positive controls) as well as the pattern, distribution, and staining intensity expected for each marker, should be kept in mind in order to correctly interpret immunostaining results. 38 A coordinate expression of immunohistochemical markers as an adjunct in a case of poorly differentiated PC is shown in Figure 1.
Meta-Review of the Immunoprofile of Urothelial Carcinoma Versus Prostate Carcinoma.
Abbreviations: PC, prostate carcinoma; UC, urothelial carcinoma; THROMBO, thrombomodulin; UPIII, uroplakin III.
Focal expression.
Suggested Panels of Antibodies to Discriminate Between PC and UC (See Text).
Abbreviations: PC, prostate carcinoma; UC, urothelial carcinoma.
Review and/or guideline articles; (I) first line markers; (II) second line markers.

A panel of immunohistochemical markers allowed to correctly diagnose a poorly differentiated prostate carcinoma (PC) on a transurethral resection of the prostate (TURP) biopsy. Tumor cells were immunoreactive to prostate markers PSA, PSAP, and P501S (with increasing intensity), while negative to CK5/6, CK7, CK20 and GATA3 (normal urothelial cells stand as positive control).
PC-Related Markers
AMACR/P504S
P504S/α-methylacyl coenzyme A racemase (AMACR) is the first gene identified by a combination of cDNA subtraction and high throughput microarrary screening from benign and malignant prostate tissue. 39 A mitochondrial and peroxisomal enzyme involved in the beta-oxidation of branched-chain fatty acids and bile acid biosynthesis, P504S/AMACR has been demonstrated as a useful positive tissue marker for prostate cancer regardless of tumor grade, with a sensitivity ranging from 82% to 100% and a specificity ranging from 79% to 100% in both low- and high-grade tumors.40-42 However, AMACR itself is not a specific marker for PC, because it may be expressed in high-grade prostatic intraepithelial neoplasia (PIN) and some benign mimickers of PC, 43 thus an antibody cocktail composed of AMACR, along with basal cell markers (HMWCK and/or p63) has been developed and is currently used in many institutions.44,45 Moreover, AMACR staining has been reported in up to 36% cases of UCs and bladder adenocarcinomas,46-48 and its expression is low to absent in atrophic PC. 49 Therefore, its role in the differential diagnosis between PC and UC remains limited.
Androgen Receptor
The androgen receptor (AR) is a member of the steroid hormone nuclear receptor superfamily that mediates androgen-induced male sexual development 50 , thus contributing to prostate growth and proliferation. 51 and is regarded as a diagnostic and prognostic marker in PC. 52 .
AR staining may be useful as a marker of prostate origin, although UCs may show positivity,53,54 which is apparently lost in high-grade high-stage tumors,55,56 yet conflicting data have been reported to date.57,58 Downes et al 53 recommend to record the staining intensity of AR, since in their study it was strong and diffuse overall in PC nuclei, even in poorly differentiated and castrate-resistant tumors, while it was weak or cytoplasmic, when present, in UCs, with statistical significant differences.
ERG
Fusion of the transmembrane protease serine 2 gene and ETS-related gene (TMPRSS2-ERG) is the most common molecular subtype of erythroblast transformation–specific (ETS) family gene fusions, occurring in approximately 50% of PCs and about 20% of high-grade PIN, which result in the overexpression of a truncated ERG protein. Therefore, if present, an ERG gene rearrangement may be detected by immunohistochemistry as well. 59 According to the literature, such ERG staining can be detected in about 40% of PC cases and, only rarely, in benign prostate glands.60-62 Because of its low sensitivity but high specificity, ERG is recommended as a second-line markers in the differential diagnosis setting. 37 In studies assessing the best antibody panel to discriminate between UC and PA, ERG has shown a prevalence as high as 35% in PC and a 100% specificity versus UC (Table 2). Overall, its sensitivity outperforms PSA, yet is lower than other prostatic markers (NKX3.1, p501S, PSMA). 54
NKX3.1
NKX3.1 is a prostate-specific androgen regulated homeobox gene located on chromosome 8p21. In human normal and neoplastic tissues, NKX3.1 expression is highly specific for prostate, testis, and breast tissue, with a nuclear staining pattern.63-65 Chuang et al 24 evaluated the sensitivity of prostate-specific immunomarkers, including NKX3.1 on 38 poorly differentiated PCs and 35 high-grade invasive UCs, reporting a sensitivity rate as high as 94.7%, which was similar to pPSA and lower than PSA and P501S; therefore, they recommended using PSA as the first screening marker in differentiating high-grade PC from high-grade UC, and other markers only when PSA shows negative or equivocal staining. Moreover, in a large series distant metastasis specimens stained less intensely for NKX3.1 compared with normal prostate and primary and metastatic PC, with a sensitivity of 90%, while all UCs were negative for this marker63,64; other studies reported sensitivity rates ranging from 69% to 92%.24,64,66-68 According to the available data, NKX3.1 is a highly sensitive and relatively specific second-line immunohistochemical marker 63 of prostatic origin when analyzing tumors of unknown origin, albeit it occasionally also may be expressed by lobular breast carcinoma. 63
Prostein (P501S)
P501S, or prostein, is a 553-amino-acid protein identified by cDNA subtraction combined with high-throughput microarray screening technique. 69 P501S mRNA show high specificity to both normal and malignant prostate tissue69-71; if positive, a golgian coarse perinuclear staining is seen.70-73 In the differential diagnosis between with UC, expression of P501S has been reported in less than 1% (3 out of 321) urothelial tumors, compared with almost 96% PCs; according to a study, its positivity raises (11%) in bladder adenocarcinomas. 74 A coordinate staining for P63 and P501S (P63-/P501s+ for PC and P63+/P501s- for UC) yielded sensitivity and specificity values of 90%-100% and 96%-100% for UC and PC, respectively, in one study. 75
PSA and PSAP
Prostate-specific antigen (PSA) and prostate-specific acid phosphatase (PSAP) are 2 well-known prostate tissue-related immunomarkers, and both are highly sensitive and relatively specific, featuring a cytoplasmic granular staining pattern with sometimes luminal accentuation. 38 However, their expression in nonprostatic tissues has been reported, and in PCs it is inversely associated with grade,52,76,77 with up to 27% of poorly differentiated tumors not expressing PSA.24,46,77-80 However, conflicting results regarding PSA sensitivity have been reported in different studies 66 ; such differences in both intensity of staining and number of stained cells may be striking when using monoclonal versus polyclonal antibodies, the latter being more sensitive,30,54 and when testing tumors with a history of androgen deprivation therapy.81,82 In a large series by Liu et al 61 assessing PSA expression (using both monoclonal and polyclonal antibodies) in 133 cases of PCs of all grades and 909 cases of nonprostatic tumors from several human tissues, positive staining was identified in 100% cases of PC (with some discrepancies in staining intensity between monoclonal and polyclonal antibodies), irrespective of tumor grade, with an overall PSA sensitivity of 100%, and specificity of 98.7%, in keeping with the results from a previous study. 83 Subsequently, in a series cytoplasmic expression of PSA was observed in some large, invasive UC (1.4%) 79 , with a potential prognostic implication. In one study comparing PSA (polyclonal antibody) staining in PCs and UCs, the latter featured 16 out of 138 positive cases. 66
As for PSAP, its sensitivity for the detection of PC appears to be similar, 82 or, as one study showed, higher than PSA. 77
PSMA
PSMA is a folate hydrolase membrane glycoprotein expressed by benign prostatic tissue, as well as by most primary and secondary PCs,24,81,82,84-86 with a higher expression in high-grade tumors than PSA, 80 although an inverse correlation between PSMA staining and Gleason score has been reported.82,84 Moreover, it is present in non-prostatic normal tissue and tumors as well, as well as in up to 17% UCs. 81 The staining pattern for PSMA seems to be different between PCs (apical, apical/cytoplasmic, or cytoplasmic with membranous accentuation) and other tumors, where PSMA has an exclusive cytoplasmic expression without membranous or apical accentuation. 87
UC-Related Markers
CK7/CK20
UC is one of the few tumors that frequently coexpresses CK7 and 20. CK7 is expressed in 87% to 100% and CK20 in 25% to 67% of tumors.46,88,89 Coexpression of CK7 and 20 is reported in 50% to 62% of UCs 90 ; expression of neither marker has been showed in up to 14% of high-grade UCs.89,90 The results are highly dependent on case selection, because CK20 expression is much lower in invasive and metastatic urothelial carcinomas (44%). 83
Although CK7 and CK20 are often used as frontline markers in the differential diagnosis between UC and PC, they actually can be seen in both and are not useful in this setting 54 ; yet, in case of a metastatic lesion of unknown origin a coordinate negativity for these markers may be used as supportive of prostatic origin. On the contrary, if the tumor expresses both CK7 and CK20, a urothelial origin may be assumed,76,83,89,91 although it should be noted that CK7 and CK20 immunostains are positive in 74% and 82% of high-grade PCs, respectively. 77
GATA3
GATA3 (GATA binding protein 3 to DNA sequence [A/T]GATA[A/G]) is a zinc finger family member. As a multifunctional nuclear transcription factor, it is involved in the proliferation, differentiation, development and function of ductal epithelial cells of breast, urothelium, epidermis, some skin adnexa, subsets of T cells, and a few other tissue types. 92 In their seminal study, Miettinen et al 36 analyzed 2500 epithelial and nonepithelial tumors by GATA3 (clone L50-823) antibody, reporting its expression in 100% low-grade UCs (22/22), 84% high-grade UCs (27/32), and 2% PCs (2/95), thus suggesting a potential discriminatory value in diagnosing UC versus PC. More recent studies seem to confirm such hypothesis, since GATA3 has been reported in >85% of UCs, irrespective of grade, showing typically diffuse, moderate to strong nuclear intensity, and is negative in PC,31,36,66,93 therefore it is the first-line marker that the International Society of Urologic Pathology recommends to confirm urothelial differentiation. 34 It has also been suggested that GATA3 might be usefully combined in a “urothelial cocktail” with the (nonnuclear) specific urothelial markers uroplakin II and uroplakin III. 94
HMWCK
One of the more frequently studied high-molecular-weight cytokeratin (HMWCK) clones, 34βE12, is expressed in 65% to 97% of urothelial tumors.46,78,83,95 According to some studies, CK34βE12 sensitivity is comparable to P63 and outperforms that of uroplakin III and thrombomodulin24,83 (see below); overall, it has been reported to range from 65% to 97%.24,46,78
A potential pitfall is its focal staining in a small number (6%) of metastatic and locally advanced PCs,78,96 especially in case of therapy-induced squamous, 97 although such staining is usually patchy and restricted to a very small percentage of cells. 24
Further HMWCKs have been investigated in single studies.
CK5/6 is positive in approximately up to 75% of UCs,88,98,99 and negative in most series of PC.77,99-101
In the study by Sakr et al, 102 CK10 was negative in all UC cases and positive in 52.4% cases of PC. Similar results were obtained by Fillies et al. 103
CK14 is more selectively expressed in squamous cell carcinoma versus UC 104 ; so far, it has not been studied in the differential diagnosis with PC.
P63
P63 is a homolog of the P53 tumor suppressor gene, predominantly expressed in the basal layer of the urothelium. P63 staining has been reported in normal urothelium 25 as well as in most high-grade urothelial,46,105 squamous, and basal cell carcinomas of the bladder. 34 Therefore, this marker has been traditionally regarded as an essential one in the differential diagnosis between UC and PC. 95 As a basal cell marker, immunostaining for P63 is commonly used to differentiate between benign (P63+) and malignant (P63−) prostatic lesions, especially on core biopsies. Nevertheless, up to 3% of all PCs are found to aberrantly express P63 in a non–basal cell distribution, and they coexist with usual-type PC in more than 80% of cases.46,106-111 A correct diagnosis may be suggested by the presence of a distinctive morphology, that is, glands, nests, and cords with atrophic cytoplasm, hyperchromatic nuclei, and prominent nucleoli, and of PSA and AMACR staining. According to some authors, P63 positive PCs are molecularly distinct with a mixed luminal/basal immunophenotype, and may carry a more favorable prognosis, 112 yet conflicting data exist in this regard. 113 Further rare P63-positive PCs are basal cell carcinomas; usually found in the transition zone, these tumors express HMWCK, while lacking PSA and AMACR.114,115
S100P
A marker of urothelial histogenesis, “placental” S100 (S100P) is a member of the S100 family of calcium-binding proteins.104,116 This antibody so far seems to be quite specific for urothelial and pancreaticobiliary tumors, 34 although its immunohistochemical expression has been reported in only 31/138 cases of UCs in a study, 66 and in up to 8% of PCs.24,116 Different staining results have been achieved by using monoclonal versus polyclonal antibodies in both UC and PC.116,117
Thrombomodulin
Thrombomodulin (CD141) is an endothelial cell associated cofactor for the thrombin-mediated activator of protein C. 25 Its sensitivity for urothelial tumors is low, ranging from 45% to 68%,24,66,83 and even lower staining rates have been reported in high-grade UCs,34,83 and it stained a few (5%) high-grade PCs in one series. 31
Uroplakin
Uroplakin III (UPIII) a transmembrane protein expressed by urothelial lining cells, especially the umbrella cells, is even less sensitive than thrombomodulin in high-grade and metastatic UC,34,118-120 since its expression has been reported to be inversely correlated to tumor grade and stage.83,117 A further issues is its staining pattern, which is often heterogeneous, focal and apical (plaque-like membranous pattern)34,118-121; a 100% specificity for UC has been reported in one study. 76
A novel marker of urothelial origin, uroplakin II has been recently reported as far more sensitive than UPIII, 122 yet validation studies are needed to assess its potential.
Unusual Histotypes
PC With Pseudopapillary Features
Since (pseudo)papillary pattern has been traditionally associated with UC, it has been traditionally considered a useful biomarker in the differential diagnosis with PC. Nevertheless, PC with pseudopapillary architecture has been described, with less than 10 cases reported in literature so far, mostly in high-grade PC. 72 According to Gordetsky et al, 72 pseudopapillae occur in areas where PC becomes loosely cohesive, losing its normal architecture of sheets and/or nests; the fragmentation of the tumor with more cohesive areas preserved around blood vessels mimics the fibrovascular cores of true papillae. Immunohistochemistry plays a pivotal role in these cases since they are positive for PSA, p501s, PSMA, AMACR, and NKX3.1, and do not express GATA3 and thrombomodulin. Focal staining for P63 and HMWCK has been reported in 2 cases.
Ductal PC
Ductal adenocarcinoma (DAC) of the prostate is an unusual subtype of PC, which is associated with an adverse prognosis. In comparison to typical (acinar) PC, DAC has a higher prevalence of extra-prostatic extension, seminal vesicle invasion, local and regional metastases, shortened time to biochemical progression, positive margins, and locally advanced disease.123,124
DAC is composed of tall, pseudostratified, columnar epithelium with high-grade, often elongated, nuclei. The cytoplasm is usually amphophilic, but may also be pale or clear.123,125 DAC usually has a papillary or cribriform architecture, but may also be solid or composed of single glands. While combinations of architectural patterns are often present,125,126 papillary architecture has been reported to be the most common and has also been described as the most useful diagnostic feature for this tumor type. 127
In their tissue microarray study of 60 DAC in comparison with other nonprostatic adenocarcinomas, Seipel et al 124 reported positivity of PSA and PSAP in all DAC cases, while prostein was negative in 3.4% (2/59) of cases. There have been reports of PSA-negative DAC cases,128,129 indicating that, for diagnostic purposes, a combination of prostate-specific markers should be employed; however, it should be taken into account that these latter studies belong to an earlier era, before the advent of automatization in immunohistochemistry.
Primary adenocarcinomas of the urinary bladder are rare and there is no ideal diagnostic immunomarker for this tumor type. In the study by Seipel et al, 124 the significantly discriminant markers between DAC and urothelial adenocarcinoma (9 cases) at univariate analysis (P < .001) were AR, p504S, P63, prostein, PSA, PSMA, PSAP, all of them but P63 being markers of prostate origin. Hence, the authors claim that DAC and urothelial adenocarcinomas may both be positive for these immunostains and additional markers for prostatic-specific antibodies and GATA3 should be added to determine a urothelial or prostatic origin.33,34
Bladder Adenocarcinoma
Primary adenocarcinoma of the bladder (BAC) is derived from the urothelium of the bladder but exhibits a pure glandular phenotype, featuring different growth patterns, namely enteric (colonic or intestinal), mucinous (colloid), signet ring cell, not otherwise specified (NOS), and mixed22,32; due to this morphological (and immunophenotypical accordingly) heterogeneity, the diagnosis of primary BAC should be made only after exclusion of secondary involvement by adenocarcinoma from other organs, including the prostate. 22 Secondary involvement of the bladder by carcinoma either by direct extension or metastasis is also uncommon but, interestingly, is actually more common than primary BAC (which accounts for approximately 2% of all primary bladder malignancies) when all sites of tumor origin are considered together.130,131 In most cases, PC features distinct morphology and can be easily identified by histology alone, since metastatic PCs predominantly show pattern 4 and/or 5 with cribriform and solid architecture, and well-formed glands (Gleason pattern 3) are only seen in a minority of cases and almost always coexist with higher-grade patterns in this setting. 96 Moreover, the nuclear pleomorphism seen in PCs involving the bladder is usually not as prominent as seen in primary BACs. 132
However, a differential diagnosis may not be supported by histology alone in cases of DAC (see above), featuring large cribriform glands with focal central necrosis, which mimic the enteric-type BAC, and in the setting of a secondary involvement of the bladder by previously treated (radiation or hormonal) PC. In such cases, a conventional panel should include both prostatic and urothelial markers; of these, PSA is always negative in BAC, while CK7, CDX2, and CA125 may be positive in BAC but are negative in PC, and CK20 may be positive in both (though less frequently in PC than in BAC).30,90,130 Those poorly differentiated/metastatic, or previously treated prostatic PCs which PSA and PSAP-negative may represent an issue,133,134 thus prompting the use of additional prostate-specific markers, including PSMA, prostein (P501S) and NKX3.1. 130
Primary UC of the Prostate With Glandular Differentiation
UC of the prostate is a rare hormone-independent variant of PC. 135 Primary UC of the prostate with glandular differentiation is a mixed type of tumor that presents glandular differentiation as well as urothelial differentiation, 32 and this condition has been reported only once so far. 136 Both P504S and 34βE12 may be expressed by primary UC with glandular differentiation, and serum PSA is unfrequently raised in this setting. 136 They are thought to carry an adverse prognosis, with lower survival rates than conventional PC.137,138
Small Cell Carcinoma
Small cell carcinoma (SmCC) is an aggressive neuroendocrine tumor that mainly occurs in the lung, and very rarely in the bladder and prostate, at least in its pure form.139,140 This tumor shows the same morphological features regardless of the site of origin, 141 and usually does not express immunohistochemical markers of histogenesis (ie, prostatic SmCCs lack PSA and PSAP), 142 as well as bladder ones are mostly negative for GATA3, thrombomodulin, uroplakin II and III, HMWCK, and CK20.76,142 Bezerra et al 141 reported a 100% specific GATA3 positivity with a 100% positive predictive value for the diagnosis of bladder SmCC, but with a low 68.7% negative predictive value. Other authors reported some rates of prostate-specific immunohistochemical markers in prostatic SmCC, such as NKX3.1, PSA, prostein, PSMA, and even ERG gene rearrangements. 143
Conclusion
Discriminating between poorly differentiated carcinomas of urothelial and prostatic origin carries implications for proper staging and treatment. Such distinction is not often feasible on morphological grounds alone, and panels of antibodies have been proposed that may support a correct diagnosis.24,34,35,38,46,53,54,66,75,78,80,82,102,144-146 In this review, we summarize these suggestions, highlight the potential issues in using single antibodies, and finally discuss some less frequent, yet intriguing differential diagnoses among tumors arising in the bladder and prostate.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
Not applicable, because this article does not contain any studies with human or animal subjects.
Informed Consent
Not applicable, because this article does not contain any studies with human or animal subjects.
Trial Registration
Not applicable, because this article does not contain any clinical trials.
