Principal Antibodies

Many different antibodies are used for the IHC evaluation of urologic neoplasms. The generally used epithelial, neuroendocrine, and mesenchymal antibodies are discussed in other chapters. In each section, we will summarize antibodies of particular importance for the neoplasms covered in that section.

Prostate-Specific Antigen

Prostate-specific antigen (PSA) is a serine protease member of the human glandular kallikrein family. PSA is a 34-kDa glycoprotein of 237 amino acids with high sequence homology with human glandular kallikrein 2 (HK2). It is almost exclusively synthesized in the prostate ductal and acinar epithelium and is found in normal, hyperplastic, and malignant prostate tissue. PSA liquefies the seminal fluid coagulum through proteolysis of the gel-forming proteins, thus releasing spermatozoa. It can reach the serum by diffusion from the luminal cells through the basal cell layer, glandular basement membrane, and extracellular matrix (ECM). Measuring total serum PSA levels is currently the mainstay of PCa detection, and numerous studies have shown that patients with PCa have, in general, elevated serum PSA levels. The most commonly used cutoff for PSA is 4 ng/mL. When serum PSA concentrations are 4 to 10 ng/mL, the incidence of cancer detection on prostate biopsy in men with a normal digital rectal examination (DRE) is approximately 25%. With serum PSA levels higher than 10 ng/mL, the incidence of PCa on biopsy increases to approximately 67%. However, the risk of cancer is proportional to the serum PSA level even at values less than 4 ng/mL. As large screening trials have demonstrated, clinically significant cancers occur in men with serum PSA levels of 2.5 to 4.0 ng/mL; thus, some experts have proposed lowering the PSA cutoff to 2.5 ng/mL to improve early detection of cancer in younger men.

Once PSA gains access into the circulation, most remains bound to serine protease inhibitors. The three most recognizable inhibitors are α-1-antichymotrypsin (α1-AT), α-2-macroglobulin, and α-1-protein inhibitor. PSA bound to α1-AT is the most immunoreactive and the most clinically useful in diagnosing PCa. A smaller fraction (5% to 40%) of the measurable serum PSA is free (noncomplexed) PSA. Therefore the total serum PSA measured reflects both free and complexed PSA. It has been demonstrated that the percentage of free PSA can improve the specificity of PSA testing for PCa. A free PSA value of less than 10% is worrisome for cancer. More recently, additional isoforms of free PSA have been discovered and were detailed in a review by Gretzer and Partin. PSA is first secreted in the form of a precursor termed pro-PSA . This inactive form of the enzyme constitutes the majority of free PSA in serum in men with PCa, making the relative increase of serum pro-PSA a risk marker of PCa. Benign PSA (BPSA) refers to a cleaved form of PSA from benign prostatic hyperplasia tissue. The measurement of the ratio of pro-PSA to BPSA has been proposed as a means of improving the accuracy of diagnosing cancer in men with a very low percentage of free PSA levels, who are at relatively high risk of cancer.

Serum PSA tests also may be used to monitor patients after therapy to detect early recurrence. Following radical prostatectomy, the serum PSA should drop to undetectable levels. Elevated serum PSA levels following radical prostatectomy (>0.2 ng/mL) indicate recurrent or persistent disease. Following radiotherapy for PCa, serum PSA values will decrease to a nadir, although not to the same extent as those following radical prostatectomy. Three subsequent rises in serum PSA values after radiotherapy indicate treatment failure.

Although PSA expression in extraprostatic tissues and tumors other than PCa have been rarely demonstrated ( Box 16.1 ), for all practical purposes, PSA expression at the IHC level is a specific and sensitive marker of prostatic lineage of differentiation with as much as 97.4% sensitivity found in a recent study from our group. Urethral, periurethral, and perianal glands are among normal tissues that have been rarely reported to show PSA reactivity. Extraprostatic neoplasms that occasionally express PSA include urethral and periurethral adenocarcinoma, cloacogenic carcinoma, pleomorphic adenoma of salivary gland, salivary duct carcinoma, and rare mammary carcinomas. Although a rare report indicated PSA expression in intestinal-type urachal adenocarcinoma of bladder, we failed to reveal such expression in a recent study of villous adenoma and adenocarcinoma of bladder. The latter is especially important from a differential diagnosis point of view, given the topographic proximity of the two organs. Although weaker intensity of PSA expression can be encountered in higher Gleason grade PCa, we were recently able to demonstrate a high degree of PSA immunostain sensitivity (97.4%) in high-grade prostate carcinoma, even when transcription-mediated amplification (TMA) sampling was used. Likewise, PSA expression is very valuable in defining a prostatic primary site of origin during the evaluation of a poorly differentiated metastatic carcinoma.

Prostate-Specific Membrane Antigen

Prostate-specific membrane antigen (PSMA) is a type II membrane glycoprotein expressed in prostate tissue and, to a lesser extent, in peripheral and central nervous system, small intestinal, and salivary gland tissues. PSMA expression has also been documented in endothelial cells of the neovasculature of many solid tumors, including renal cell carcinoma (RCC). In the prostate, PSMA is expressed by benign and malignant prostatic epithelial cells, with a higher extent of staining seen in the latter. It is also expressed by high-grade prostatic intraepithelial neoplasia (PIN). PSMA expression correlates with PCa stage and Gleason grade. The increase in both expression and enzymatic activity of PSMA in aggressive PCa points to a selective advantage imparted on cells that express PSMA, thereby contributing to the development and progression of PCa. Increased PSMA expression is an independent predictor of PCa recurrence, and PSMA expression is maintained in hormone-refractory PCa, thus increasing its utility in such settings. Several imaging strategies exploit PSMA specificity to PCa and are currently in use for PCa diagnostic imaging. Furthermore, PSMA is under investigation as a target of therapy in PCa and other solid tumors, given its expression by the neovasculature of extraprostatic tumors. Cytoplasmic and, to a lesser degree, membranous PSMA expression has been recently documented in 11% of analyzed urinary bladder adenocarcinomas, a fact worth noting when the differential diagnosis includes prostatic and bladder adenocarcinoma.

Prostatic Acid Phosphatase/Prostate-Specific Acid Phosphatase

Prostate-specific acid phosphatase (PSAP) is one of the earlier prostate lineage markers to be exploited for immunolabeling of PCa before the discovery of PSA. Currently, the use of PSAP as a marker of prostatic differentiation has declined, given its relative lack of specificity compared with PSA and the more variable staining of PSAP in higher grade PCa.

Prostein/P501S

P501S is a 553-amino acid protein localized to the Golgi complex. It is expressed in both benign and neoplastic prostate tissues. Typical P501S stain has a perinuclear cytoplasmic (Golgi) location and a speckled pattern. Expression is retained in poorly differentiated and metastatic PCa. P501S demonstrated up to 99% sensitivity in a recent study from our group by Sheridan and colleagues. In rare metastatic lesions, P501S positivity may be encountered in the presence of PSA-negative expression, making it an advantageous addition to a prostate lineage immunopanel. To date, P501S expression has not been shown in extraprostatic carcinomas, which makes it of great utility in differentiating high-grade PCa from other high-grade carcinomas, including colorectal and urothelial carcinoma (URCa).

α-Methylacyl–Coenzyme-A Racemase/P504S

Alpha-methylacyl-CoA racemase (AMACR) is mainly localized to peroxisomal structures and plays a critical role in peroxisomal beta oxidation of branched chain fatty acid. In their original detailed IHC analysis, Luo and associates demonstrated that both prostate carcinomas and high-grade PINs consistently revealed a significantly higher expression than that of matched normal prostate epithelium. Both untreated and hormone-refractory PCa metastases generally maintain a strong positive reactivity for AMACR. An overall PCa sensitivity and specificity of 97% and 92%, respectively, have been shown in a multiinstitutional study by Jiang and Woda.

Cytoplasmic AMACR staining combined with the absence of basal cell markers, such as the nuclear protein p63 and high-molecular-weight cytokeratins (HMWCKs), has proved to be of greatest utility in providing an added layer of assurance in establishing the diagnosis of PCa on small needle biopsy foci. However, AMACR expression has been repeatedly demonstrated in high-grade PIN and in some benign mimics of PCa, such as glandular and partial atrophy and adenosis; therefore AMACR is of limited utility as a single marker in resolving the differential diagnosis of PCa in such lesions. A panel of immunostains that includes AMACR, HMWCK, and p63 (positive AMACR immunostaining [ Table 16.1 ] along with negative basal cell markers) is recommended in the interrogation of atypical prostatic glandular foci.

High-Molecular-Weight Cytokeratins

HMWCKs are of great utility in highlighting the presence or absence of basal cells in a focus of atypical prostate glands. 34βE12 is currently the most widely used clone, both individually or as a component of a three-antibody cocktail, which includes a second basal cell marker, such as p63, and AMACR. Alternatively, CK5/6 can be used as the HMWCK marker individually or in combination with p63 and AMACR. A recent study by Abrahams and colleagues seems to suggest a superior sensitivity for CK5/6 as an HMWCK in prostate biopsies fixed in Hollande’s fixative. Following the initial examination of hematoxylin and eosin (H&E)-stained routine sections, the application of such a cocktail to previously prepared, unstained, intervening sections is recommended in biopsies where establishing the presence or absence of basal cells in a questionable focus will lead to a definitive resolution of a benign or malignant diagnosis, respectively.

p63

The p53 homolog p63 encodes for different isotypes that can either transactivate p53 reporter genes (TAp63) or act as p53-dominant negatives (ΔNp63), and p63 is expressed in the basal or myoepithelial cells of many epithelial organs; its germline inactivation in the mouse results in the agenesis of organs, such as skin appendages and the breast. In the prostate, p63 expression is limited to basal cells and is absent in secretory and neuroendocrine cells, and ΔNp63α isotype is the most abundantly represented isotype in normal prostate basal cells. Recent experimental evidence also suggests that the p63 gene is essential for normal stem cell function in the prostate. Several studies have confirmed the clinical utility of p63 immunostain as a prostate basal cell marker, and some studies suggest a slight sensitivity advantage for p63 over HMWCK alone. Additionally, the use of basal cell HMWCK and p63 cocktails may reduce the staining variability that may be encountered in basal cells, and may further decrease the false-negative and false-positive rates of basal cell labeling by either marker alone ( Table 16.2 ).

Finally, given the fact that immunostains for basal cell markers are typically used in a “negative” diagnostic mode, to show the absence of basal cells in PCa, sole reliance on such markers is not advocated, and the identification of a combination of major and minor histologic features of PCa is crucial for achieving clinical diagnostic accuracy. In this regard, consideration should be given to the fact that benign prostatic glands from the transition zone are subject to basal cell staining variability, which may result in occasional negative basal cell staining in such benign glands. Furthermore, basal cells can be retained, albeit very rarely, in individual glands in otherwise typical acinar PCa focus, and the constellation of diagnostic features are to be relied on in such rare cases. We have recently described an intriguing p63-positive, HMWCK-negative variant of PCa, in which nuclear p63 staining is seen in secretory PCa cells in a nonbasal distribution.

NKX3-1

NKX3-1 is a prostate-specific androgen-regulated homeobox gene required for tissue differentiation, whose loss of function leads to carcinogenesis. In the normal prostate, NKX3-1 controls differentiation and protects against oxidative damage by regulating gene expression in conjunction with other transcription factors. In cancer, a relative loss of NKX3-1 expression occurs as a result of loss of heterozygosity (LOH), promoter methylation, or alterations in NKX3-1 degradation. Downregulation of NKX3-1 protein leads to increased prostate epithelial cell proliferation, differentiation, and susceptibility to DNA damage, thereby furthering oncogenic insult. In addition to benign and malignant prostatic epithelium, NKX3-1 expression is found in normal testis, bronchial mucous glands, and infiltrating lobular carcinoma of the breast. Sensitivity of NKX3-1 for PCa has ranged from 68% to 94.7%. In poorly differentiated PCa, NKX3-1 appears to be superior to PSA, which may show a relatively focal weaker staining in that subset.

Immunohistochemistry in Small Focus of Prostate Carcinoma

The use of IHC markers to help establish the diagnosis of carcinoma in a morphologically atypical small focus of prostate glands is currently a common laboratory practice. As mentioned above, used individually or with two or three markers combined in a panel, HMWCK, p63, and AMACR offer great help in ensuring the absence of a basal layer combined with positive AMACR labeling in such small foci ( Fig. 16.1 ). Such a panel is also of use in distinguishing a small focus of PCa infiltrating adjacent to a high-grade PIN lesion from the glandular outpouching of high-grade PIN where an interrupted (patchy) layer of basal cells would still be identified with the aid of immunostains ( Figs. 16.2 and 16.3 ). Fully characterizing and delineating the group of atypical acini in question, based on a combination of established H&E morphologic features of malignancy before their interrogation by immunostains, is the key to a successful diagnostic approach in PCa. The key H&E morphologic features include small acinar architecture, single-layer lining, straight luminal borders, amphophilic cytoplasm, nuclear enlargement and atypia, the presence of prominent nucleoli, wispy or blue mucin content, dense eosinophilic secretions or “cancer” crystalloids, and the presence of mucinous fibroplasia. The demonstration of an increasing combination of the above morphologic features in the presence of a supportive immunostaining pattern will allow for a significant increment in diagnostic confidence when faced with increasingly smaller-sized atypical foci on a needle biopsy. If such a confidence level is unobtainable despite the application of immunostains, a diagnosis of focus of atypical glands suspicious of (but not diagnostic of) malignancy, should be rendered, with a recommendation for a close repeat follow-up biopsy to rule out malignancy ( Fig. 16.4 ).

(A) The utility of immunostains in support of diagnosis of a small focus of prostate carcinoma. (B) Lack of basal cell layer is illustrated by high-molecular-weight keratin (CK903). (C) Positive staining in the tumor glands for α-methylacyl–coenzyme-A racemase is shown.

(A) Cribriform high-grade prostatic intraepithelial neoplasia. (B) The preservation of basal layer staining with high-molecular-weight cytokeratin is shown.

(A and B) Atypical glands adjacent to high-grade prostatic intraepithelial neoplasia (PIN). Given the presence of an interrupted (patchy) layer of basal cells, the possibility of glandular outpouchings of PIN could not be excluded with the aid of immunostains (p63 in C and high-molecular-weight keratin in D).

Prostate adenocarcinoma diagnosis supported by absence of basal layer on immunostains in two groups of atypical glands (A) and infiltration among benign glands (B). Use of immunostains for high-molecular-weight cytokeratin (C) and p63 (D).

The routine initial use of immunostain cocktails as a screening tool before H&E examination, to facilitate the identification of basal cell negative foci, is not advocated for obvious reasons, which include cost, misallocation of resources, and potential detriment to diagnostic accuracy. On the other hand, when used judiciously, the role of HMWCK in decreasing diagnostic uncertainty expressed in prostate needle biopsies has been established in several large studies, including a College of American Pathology (CAP) Q-probes study of more than 15,000 biopsies. In a large study from our center, 34βE12 stains either helped to establish (14%), confirmed (58%), or changed (2%) our diagnoses when applied to questionable/atypical foci. In an additional 18% of cases, the diagnosis remained (or became) equivocal despite the use of HMWCK.

False-negative staining of basal cells with HMWCK can occur for a variety of technical reasons, including suboptimal antigen retrieval, and this should be taken into consideration. Finally, it is worth noting that a very low but existent false-positive HMWCK immunostaining of PCa cells can be encountered (0.2% to 2.8%), characteristically in a nonbasal-cell distribution pattern ( Fig. 16.5 ). Another complicating issue in interpreting basal cell immunostaining results is the p63-positive, HMWCK-negative rare variant of PCa mentioned previously ( Fig. 16.6 ).

A focus of prostate adenocarcinoma shows scattered tumor cells with positive expression of high-molecular-weight cytokeratin in a nonbasal distribution. Such findings alone should not deter the pathologist from making a diagnosis of carcinoma if all other features of malignancy are satisfied.

(A) Rare variant of prostate adenocarcinoma shows aberrant expression of p63 but not high-molecular-weight cytokeratin (HMWCK). (B) Note nuclear positivity for p63 in tumor glands (combo p63-34βE12). (C) Note infiltrating tumor glands showing tumor cells staining for p63. (D) The tumor is negative for HMWCK.

Immunohistochemistry in Benign Mimics of Prostate Adenocarcinoma

Prostatic Atrophy

Partial atrophy (PTAT) and postatrophic hyperplasia (PAH) are the most problematic morphologic variants of atrophy that may mimic PCa. In fact, PTAT is the most common mimic of PCa on needle biopsy, mainly because of the presence of disorganized acini lined by pale cytoplasm with occasional acini that retain a full height of cytoplasm and contain slightly enlarged nuclei with notable nucleoli. PTAT foci can morphologically mimic “atrophic” PCa. In difficult PTAT lesions, immunostains for basal cell markers help highlight the presence of basal cells. HMWCKs (34βE12, CK5/6) and p63 show patchy positivity in basal cells in at least some of the glands ( Fig. 16.7 ). Lack of positivity in some glands should not be misinterpreted as PCa as long as the negative and positive glands share similar cytologic features. It is also important to remember that AMACR can be expressed by some PTAT acini.

Foci of partial atrophy (A and B) show patchy basal cell layer positivity with high-molecular-weight keratin (C) and p63 (D). Partial atrophy glands may display positive α-methylacyl–coenzyme-A racemase staining (D and E).

Only rarely does the clinician need to resort to immunostains to recognize simple atrophy and PAH lesions. The latter two variants of atrophy demonstrate a continuous basal layer on immunostains, and they are usually negative for AMACR ( Fig. 16.8 ).

(A) Focus of prostatic postatrophic hyperplasia. (B) Continuous basal cell layer on high-molecular-weight cytokeratin immunostain.

Adenosis

Adenosis is a common mimic of PCa both on needle biopsy and on transurethral resection of prostate (TURP). Given its preferential occurrence in the transition zone, adenosis is more frequently seen in TURP (1.6%) compared with needle biopsy (0.8%). Adenosis is characterized by a nodular proliferation of small glands. Within such nodules, larger elongated glands with papillary infolding and branching lumina share identical nuclear and cytoplasmic features with the admixed smaller, more suspicious glands. In contrast, small PCa glands usually stand out cytologically from adjacent benign larger glands.

To avoid misinterpretation of adenosis, the constellation of histologic features in a given lesion should outweigh the significance of any one diagnostic feature given the fact that several features are shared between adenosis and PCa. Therefore, in difficult cases, IHC for HMWCK can be of great utility to demonstrate the patchy positivity of basal cells in adenosis ( Fig. 16.9 ). Lack of positivity for HMWCK in some of the glands should not be misinterpreted as evidence of PCa, as long as the negative and positive glands share similar cytologic features. Of note, AMACR can be focally or diffusely expressed in adenosis in as many as 10% of cases.

(A) Prostate gland with a focus of adenosis. (B) Patchy staining of basal cells by using high-molecular-weight cytokeratin.

Sclerosing Adenosis

Sclerosing adenosis is a rare lesion mainly encountered in TURP specimens performed for urinary obstruction. Very rarely, it also may be sampled on needle biopsy, leading to potential over-diagnosis as PCa. Sclerosing adenosis is composed of a relatively well-circumscribed proliferation of well-formed glands admixed with single epithelial cells set in a background of dense spindle cell proliferation. The glandular structures are similar to those seen in adenosis. Some glands are surrounded by a distinct eosinophilic hyaline sheath-like collarette. The lining epithelial cells usually lack atypia, and a basal cell layer can be appreciated on H&E. Establishing the diagnosis of sclerosing adenosis in examples that demonstrate atypical features, such as the presence of crystalloids, mitotic figures, and prominent nucleoli, requires the aid of immunostains. Basal cells and spindle cells are unique in their true myoepithelial differentiation as indicated by the coexpression of keratin and muscle-specific actin (MSA; Fig. 16.10 ). The latter is not expressed by basal cells in normal prostate glands.

Key Diagnostic Points: Atrophy, Adenosis, and Sclerosing Adenosis

• •
  

  Atrophy, adenosis, and sclerosing adenosis may mimic prostate cancer, and basal marker immunostains p63 and 34βE12 may aid in the diagnosis.

  
  
• •
  

  Adenosis and sclerosing adenosis are most commonly seen in transurethral resection of prostate specimens.

(A) Sclerosing adenosis of the prostate gland. (B) Positive smooth muscle actin staining in myoepithelial cells.

Xanthoma

Prostatic xanthomas are rare but could be potentially mislead­ing lesions in small and distorted needle-tissue fragments. Typical low microscopic appearance is that of a small, well-circumscribed, solid nodule; examples that are architecturally set as infiltrative cords and individual cells are more troubling. Xanthoma cells have a uniform appearance and contain abundant foamy cytoplasm and bland nuclei without prominent nucleoli ( Fig. 16.11 ). Mitotic activity is usually lacking. However, mitotic figures are also rare in PCa. Immunostains should be obtained if the possibility of xanthoma is suspected based on morphology. Expression of histiocytic markers, such as CD68, and lack of cytokeratin (CAM5.2) staining support the diagnosis.

(A) Focus of xanthoma involving prostate. (B) CD68 positivity.

Posttherapy Changes in Prostate Adenocarcinoma

Antiandrogen Therapy

Currently, luteinizing hormone-releasing hormone (LHRH) agonist (leuprolide), typically in association with the antiandrogen flutamide, is the most commonly used form of hormonal therapy in PCa. Both benign and neoplastic prostatic tissue can be significantly altered with hormonal therapy. Under hormone deprivation, neoplastic acini usually acquire an atrophic appearance and can mimic benign atrophic glands because of the relative lack of nuclear atypia and the absence of prominent nucleoli. At times, treated PCa glands develop pyknotic nuclei with abundant xanthomatous cytoplasm, and when present as scattered cells, they will closely resemble foamy histiocytes ( Fig. 16.12 ). IHC for PSA or pancytokeratin can aid in the diagnosis of carcinoma. As with their untreated counterparts, PCa cells after hormonal therapy demonstrate a lack of HMWCK staining. Following hormonal therapy, a decrease may be noted in immunoreactivity with prostate lineage markers such as PSA, P501S, PSMA, and PSAP. However, with the exception of tumors that develop focal squamous differentiation, resulting in adenosquamous carcinomas, most tumors will maintain at least focal labeling with these antibodies. In our laboratory, we find that using a panel of three of the above markers—PSA, PSMA, and P501S—will increase the sensitivity for prostatic differentiation. Finally, it is worth remembering that squamous components of recurrent or metastatic hormone-independent prostatic adenosquamous carcinoma will only rarely and very focally be positive for prostate lineage markers, such as PSA, PSMA, P501S, and PSAP, and will diffusely express HMWCK in these areas.

Key Diagnostic Points: Posthormonal Therapy Histology

• •
  

  Following hormonal therapy, prostate cancer cells demonstrate a lack of high-molecular-weight cytokeratin staining. Following hormonal therapy, a decrease in immunoreactivity may be seen with prostate lineage markers, such as prostate-specific antigen, P501S, prostate-specific membrane antigen, and prostate-specific acid phosphatase.

(A) Prostate adenocarcinoma showing changes prior hormonal therapy effect. Staining with α-methylacyl–coenzyme-A racemase (AMACR) highlights the scattered residual tumor cells that lack a basal cell layer. (B) Combination AMACR and CK 34βE12.

Radiation Therapy

Besides surgery, external beam radiation and or interstitial radiotherapy (brachytherapy) are two additional standard treatment options for localized PCa with a curative intent. Radiated nonneoplastic prostatic glands undergo glandular atrophy, squamous metaplasia, and cytologic atypia. The marked epithelial atypia, especially following brachytherapy, tends to persist for several years. The distinction between irradiated nonneoplastic prostatic glands and prostate carcinoma can be difficult, especially if the history of prior treatment is not provided and is not considered by the pathologist. On low magnification, radiated benign prostate glands maintain their normal architectural lobular configuration. On higher magnification, piling up of the nuclei occurs within irradiated benign glands with a recognizable basal cell layer ( Fig. 16.13 ). The finding of scattered, markedly atypical nuclei with a degenerative, hyperchromatic, and smudgy appearance within well-formed acini is typical of radiated benign glands. In contrast, radiated glands of PCa are lined by a single cell layer with typical pyknotic nuclei and foamy cytoplasm. PCa that is sufficiently differentiated to form glands rarely manifests the degree of atypia seen with radiation. In difficult cases, HMWCK can aid in the diagnosis of irradiated prostate by identifying basal cells within benign radiated glands to prevent a false-positive interpretation of carcinoma.

(A) Prostate adenocarcinoma showing changes of prior radiation therapy effect. (B) Lack of a basal cell layer is demonstrated on high-molecular-weight cytokeratin immunostain. (C) Positivity for prostate-specific antigen is shown.

Another scenario in which radiation treatment can introduce diagnostic difficulty when recurrent or residual PCa displays marked and extensive radiation effect in the form of glands or individual cells with abundant vacuolated cytoplasm that takes on a histiocytic appearance. The nuclei lack apparent nucleoli and are pyknotic with smudged chromatin. Pancytokeratin (AE1/AE3 and CAM5.2) and CD68 markers can be used to illustrate the epithelial nature of treated PCa. In most cases, treated PCa will retain its PSA and PSAP positivity as well as its expression of AMACR (P504S). However, as mentioned previously, recurrent or metastatic radiated PCa that displays a sarcomatoid, squamous, or adenosquamous phenotype may only focally be positive for prostate lineage markers while expressing HMWCK. When postradiation clinical response is evaluated, prostate biopsies are obtained 1 year after conclusion of treatment. Negative biopsies and the presence of residual PCa displaying severe radiation effect portend good response. Residual tumor without a demonstrable radiation effect is considered a strong predictor of clinical failure. The expression of proliferation markers (MIB-1 or PCNA) in postradiated cancer has also been shown to correlate with clinical failure.

Prostatic Duct Carcinoma

Less than 1% (0.4% to 0.8%) of PCa shows distinctive tall columnar cells in papillary or cribriform structures and is classified as prostatic duct adenocarcinoma, which can be encountered as a single pattern of tumor differentiation or, more frequently, is found admixed with “usual” acinar differentiation. Prostatic duct adenocarcinomas show a variety of architectural patterns that include a papillary exophytic architecture, seen in a periurethral location, lined by tall pseudostratified epithelial cells and a cribriform pattern, more commonly seen deeper within the tissue, formed by back-to-back large glands with slit-like lumina. It is not uncommon to find areas of papillary formation admixed with cribriform patterns. An important point is that ductal adenocarcinomas, because they arise in ducts, may show residual staining for HMWCK and p63 ( Fig. 16.14 ).

(A and B) Prostate ductal adenocarcinoma. (C) Lack of a basal cell layer on high-molecular-weight cytokeratin and p63 stains is shown.

Prostatic duct adenocarcinomas can invade as single glands lined by tall columnar cells, unlike the cuboidal cells that characterize acinar prostatic carcinoma. The single infiltrating glands of prostatic duct adenocarcinoma may resemble infiltrating colonic adenocarcinoma. The differentiation between prostatic duct adenocarcinoma and secondary involvement of the prostate by colonic adenocarcinoma is usually made by finding more typical prostatic duct adenocarcinoma elsewhere in the biopsy. Rarely in such settings, IHC demonstration of PSA or other prostate lineage markers, such as P501S, is needed to identify a prostatic duct adenocarcinoma. Adding β-catenin, CDX-2, and villin—all of which are positive in colon cancer—to the IHC panel can be of further utility in such a differential, although uncommonly, prostatic ductal adenocarcinomas are CDX2 positive. Prostatic duct adenocarcinoma on TURP specimens can also mimic papillary URCa. Nuclear features can be helpful in such a differential, because nuclei in URCa tend to be more pleomorphic and angulated. PSA and PSAP positivity and negative reactivity for GATA3 thrombomodulin and uroplakin in prostatic duct adenocarcinoma also can be useful.

Neuroendocrine Prostatic Neoplasms

It is somewhat controversial whether neuroendocrine differentiation in typical adenocarcinomas worsens prognosis. Three studies evaluated the prognostic role of neuroendocrine differentiation in organ-confined prostate adenocarcinoma. No prognostic role was found in the first study, whereas the two subsequent larger studies, including ours, found only a marginal prognostic role insufficient to be clinically useful. In the single study that analyzed neuroendocrine differentiation in PCa on needle biopsy, no relationship with prognosis could be established. According to the 1999 CAP Consensus Statement, neuroendocrine differentiation is still considered a category III prognostic factor not sufficiently studied to demonstrate its prognostic value.

As in other organs, the spectrum of neuroendocrine neoplasms in the prostate include carcinoid tumors, small cell carcinoma, and large cell neuroendocrine carcinoma (LCNEC) as defined in the lung by Travis and colleagues.

True carcinoid tumors of the prostate are extremely rare. Recently, a total of three such cases have been reported with documented negative immunoreactivity for PSA and PSAP, and otherwise typical carcinoid tumor morphology and immunoprofile. All three patients came to medical attention with normal serum PSA levels; they lacked clinical features of carcinoid syndrome. Several additional cases have been reported in which at least a focal “carcinoid-like” appearance has been present. None of these patients had carcinoid syndrome, and all such cases have been positive with antibodies for PSA and PSAP. They have clinically behaved like ordinary prostate carcinomas. A more appropriate designation for these lesions is prostatic adenocarcinomas with neuroendocrine differentiation.

The basis for a diagnosis of small cell carcinoma of the prostate is the presence of morphologic features similar to those found in small cell carcinomas of the lung as defined in the 1999 World Health Organization (WHO) classification criteria. In approximately 50% of the cases, the tumors are mixed small cell carcinoma and adenocarcinoma of the prostate. As with other unusual subtypes of PCa, we do not assign a Gleason score to small cell carcinoma, only to the conventional adenocarcinoma component. Immunohistochemically, the small cell component is positive for one or more neuroendocrine markers—neuron-specific enolase, synaptophysin, chromogranin, or CD56—and is negative for markers of prostatic differentiation such as PSA, PSMA, P501S, and PSAP. A minority of small cell carcinomas is positive for prostatic markers to varying degrees and may be negative for neuroendocrine markers. In a recent study by Yao and colleagues, strong chromogranin and synaptophysin positivity was present in 61% and 89%, respectively, of the studied prostatic small cell carcinoma. PSA and PSAP were positive in 17% and 24% of cases, respectively. In 24% and 35% of cases, positivity was noted for p63 and HMWCK, markers typically negative in prostatic carcinoma yet expressed in normal basal cells of the prostate. In our recent and largest study on small cell carcinoma of the prostate, we found most small-cell carcinomas (88%) to be positive for at least one neuroendocrine marker. We also found P501S and PSMA to be better in identifying the prostatic origin of small-cell carcinoma than PSA, although the majority (60%) of prostatic small cell carcinomas were still negative for all three markers ( Fig. 16.15 ). The latter, together with the above-cited heterogeneity of the prostatic small cell carcinoma immunophenotype, is consistent with an origin from multipotential transient amplifying cells that are closely related to stem cells.

(A and C) Small cell carcinoma of prostate in association with acinar prostate carcinoma. (D) Focal weak positivity for prostate-specific antigen. (E) Positive staining for P501S.

Ordonez originally reported that thyroid transcription factor 1 (TTF-1) was positive in 96% of small cell carcinomas of the lung and was negative in all three prostate small cell cancers. Subsequent studies, including our own, have demonstrated TTF-1 expression in the majority of small cell carcinomas of the prostate, limiting its utility in distinguishing primary small cell carcinoma of the prostate from a metastasis from the lung.

Small cell carcinoma of the prostate continues to have a dismal outcome, with average survival of less than 1 year. No difference in prognosis is observed among patients with pure small cell carcinomas or mixed glandular and small cell carcinomas. A tumor immunoprofile does not affect survival. A review by Mackey and colleagues concluded that hormonal manipulation and systemic chemotherapy had little effect on the natural history of prostate small cell carcinoma. Others suggest treatment of small cell carcinoma of the prostate with the same combination chemotherapy used to treat small cell carcinomas in other sites. It remains to be seen whether new targeted therapy strategies currently under investigation in small cell carcinoma of the lung may be applicable to small cell carcinoma of the prostate, expressing targets such as c-kit, Bcl-2, and CD56.

LCNEC of the prostate is an extremely rare occurrence. In the largest series on the topic by Evans and associates, only one out of seven cases was a de novo LCNEC; the remaining six cases represented progression from prior acinar adenocarcinoma following long-standing hormonal therapy. LCNEC is composed of sheets and ribbons of amphophilic cells with large nuclei, coarse chromatin, and prominent nucleoli. Mitotic activity is brisk, and foci of necrosis are common. A minor (<10%) component of conventional prostate adenocarcinoma showing hormonal deprivation effect was identified in all but the single de novo case. The LCNEC component was strongly positive for CD56, CD57, chromogranin A, synaptophysin, and P504S. PSA and PSAP expression was present in the conventional component but was focal or absent in the LCNEC areas. All six patients with available follow-up died, and mean survival was 7 months.

Urothelial Carcinoma That Involves Prostate and Prostatic Urethra

Prostatic involvement by URCa can result from direct invasion of an infiltrating bladder cancer into prostate stroma and through the extension of the urothelial tumor through an intraductal route with or without subsequent stromal invasion of the prostate. In the first scenario, direct bladder wall to prostate invasion, the prognosis of the URCa of the bladder is equivalent in survival to cases of bladder carcinoma with regional lymph node metastases. A common diagnostic problem in this setting is differentiating between a poorly differentiated URCa of the bladder and a poorly differentiated prostatic adenocarcinoma in a TURP specimen. Because the therapy differs significantly, the distinction between these two entities is crucial. Even in poorly differentiated prostatic carcinomas, relatively little pleomorphism or mitotic activity is typical compared with poorly differentiated URCa. A more subtle finding is that the cytoplasm of prostatic adenocarcinoma is often very foamy and pale, imparting a “soft” appearance. In contrast, URCas may demonstrate hard, glassy eosinophilic cytoplasm or more prominent squamous differentiation. The findings of infiltrating cords of cells or focal cribriform glandular differentiation are other features more typical of prostatic adenocarcinoma. Although the above distinction between URCa and prostatic adenocarcinoma on H&E-stained sections is valid for almost all cases, we have seen rare cases in which prostate adenocarcinoma has had marked pleomorphism identical to URCa. Consequently, in a poorly differentiated tumor that involves the bladder and prostate without any glandular differentiation typical of prostate adenocarcinoma, the case should be worked up immunohistochemically, given the high stakes of a misdiagnosis. Approximately 95% of poorly differentiated prostatic adenocarcinomas show PSA and PSAP staining, although it may be focal. Some authors have demonstrated a superiority of PSA over PSAP in staining prostatic carcinoma, whereas others have found poorly differentiated prostatic carcinomas that lacked PSA staining but still maintained immunoreactivity to PSAP. In our lab, PSA has, in general, been more sensitive. Monoclonal antibodies to PSAP have lower sensitivities than their polyclonal counterparts, but are more specific. We have compared PSA staining in a group of poorly differentiated prostatic adenocarcinomas with “poor” PSA staining to newer prostate-specific markers, including PSMA, P501S (prostein), and NKX3-1 ( Fig. 16.16 ). Completely negative staining was seen in 15% (PSA), 12% (PSMA), 17% (P501S), and 5% (NKX3-1) of the cases, and 5% were negative for all four markers. A similar 5% rate of false negativity is found when combining PSA and PSAP stains. Therefore the lack of immunoreactivity to prostate-specific markers in a poorly differentiated tumor within the prostate, especially in small samples, does not exclude the diagnosis of a poorly differentiated prostatic adenocarcinoma. With only a few exceptions, immunoperoxidase staining for PSA and PSAP is very specific for prostatic tissue. Situations that can cause diagnostic difficulty include PSA and PSAP within periurethral glands, and cystitis cystica and cystitis glandularis in both men and women. Other examples of cross-reactive staining include anal glands in men (PSA, PSAP) and urachal remnants (PSA). Some intestinal carcinoids and pancreatic islet cell tumors are strongly reactive with antibodies to PSAP, yet they are negative with antibodies to PSA. Periurethral gland carcinomas in women and various salivary gland tumors also may be PSA and PSAP positive. Although adenocarcinomas of the bladder, whether as a pure tumor or with mixed URCa, have also been reported to be positive for PSA or PSAP, there has yet to be a case reported positive for both. In a poorly differentiated tumor occurring in the bladder and the prostate, where the differential diagnosis is between high-grade prostatic adenocarcinoma and URCa, focal strong staining for either marker can be used reliably to make the diagnosis of prostatic adenocarcinoma, because PSAP and PSA false positivity have not been convincingly described in URCas. In general, various cytokeratins (CK7, CK20, HMWCK) show strong positivity in cases of URCa involving the prostate. Although CK7 and CK20 are more frequently seen in URCa, compared with adenocarcinoma of the prostate, they may also be positive in adenocarcinoma of the prostate, so in our experience, they are not that helpful in this differential diagnosis. We and others have found HMWCK to be positive in more than 90% of URCas. In contrast, HMWCK is only rarely (8%) expressed, and usually in a very small percentage of cells, in adenocarcinoma of the prostate. Another useful marker in differentiating high-grade URCa from prostatic adenocarcinoma is p63. Using tissue microarrays, we found p63 to have a greater specificity, albeit with lower sensitivity, for URCa compared with HMWCK (100% specificity and 83% sensitivity; Fig. 16.17 ). Other markers that also appear highly specific but are only of modest sensitivity for URCa include uroplakin and thrombomodulin (49% to 69% sensitivity).

High-grade prostate adenocarcinoma (A) showing weak focal prostate-specific antigen positivity (B) but strong diffuse staining for P501S (C) and prostate-specific membrane antigen (D). Nuclear staining for NKX3-1 is also encountered (E).

(A) High-grade urothelial carcinoma. (B) Focal high-molecular-weight cytokeratin positivity. Strong diffuse nuclear staining for p63 (C), diffuse cytoplasmic staining for thrombomodulin (D), and S100A protein (E).

More recently, GATA3 has become the favored marker with the highest sensitivity and specificity for URCa versus PCa. If intraductal URCa is identified on TURP or transurethral biopsy, patients usually will be recommended for radical cystoprostatectomy. The finding of intraductal URCa also has been demonstrated to increase the risk of urethral recurrence following cystoprostatectomy, so its identification may also result in prophylactic total urethrectomy. IHC stains for basal cells (HMWCK, p63) may in some cases only outline the prostatic basal cells, and in other cases may label the intraductal URCa.

The diagnosis of URCa on prostate needle biopsy is especially difficult. Clinically, URCa that involves the prostate can mimic prostatic adenocarcinoma in terms of findings on DRE and ultrasound, along with the potential for an elevated serum PSA level. Histologic features and IHC studies are therefore essential to establish the correct diagnosis. URCa that involves the prostate differs from adenocarcinoma of the prostate both architecturally and cytologically, and it typically forms nests of tumor, whereas poorly differentiated PCa tends to form sheets, individual cells, or cords. In our study, URCa that involved the prostate contained areas of necrosis in 43% of cases. Necrosis is an unusual finding even in high-grade adenocarcinoma of the prostate. The presence of an intraductal growth in which preexisting benign prostate glands are filled with solid nests of tumor also differs from high-grade PIN, which is composed of flat, tufting, papillary, or cribriform patterns. The presence of squamous differentiation seen in 14% of our cases would also be unusual for adenocarcinoma of the prostate. Cytologically, URCas that involve the prostate tend to show greater nuclear pleomorphism, variably prominent nucleoli, and increased mitotic activity compared with even poorly differentiated prostate adenocarcinoma. In high-grade adenocarcinomas of the prostate, nuclei tend to be more uniform from one to another with centrally located prominent eosinophilic nucleoli. Mitotic figures in high-grade PCa are typically not as frequent compared with what is seen in URCa on biopsy. Finally, the presence of stromal inflammation, seen in 76% of our cases of URCa on biopsy, differs from the typical lack of associated inflammation seen with ordinary adenocarcinoma of the prostate ( Fig. 16.18 ).

(A and B) Intraductal spread of urothelial carcinoma (URCa) involving prostate gland. (C) URCa undermines the secretory prostate epithelial layer, as shown on prostate-specific antigen stains. (D) URCa cells react positively with high-molecular-weight cytokeratin.

Secondary Involvement of Prostate by Colorectal Adenocarcinoma

Another source of secondary tumor extension into the prostate is the topographically adjacent colorectal tract. Here again, attention to some characteristic morphologic features should raise the possibility of a secondary spread. The presence of goblet/columnar cell differentiation, pseudostratified basally located nuclei, and characteristic “dirty necrosis” are more likely encountered in colorectal carcinoma (CRCa). One should be cautioned that single infiltrating glands of prostatic duct adenocarcinoma can resemble infiltrating colonic adenocarcinoma. The differentiation between prostatic duct adenocarcinoma and secondary involvement of the prostate by CRCa can be facilitated by finding more typical prostatic duct adenocarcinoma elsewhere within the biopsy. An IHC profile of positive nuclear CDX2 staining, positive nuclear staining for β-catenin (cytoplasmic staining can occur in PCa) and positive staining for CK20 in the face of negative reactivity for PSA, NKX3.1, and P501S can be used to confirm the diagnosis of CRCa spread.

Prostatic Mesenchymal Tumors

As in any other organ, immunostains can be of great utility in resolving a variety of spindle cell mesenchymal lesions that occur in the prostate, including benign and malignant smooth muscle neoplasms, peripheral nerve sheath tumors, and rhabdomyosarcoma ( Table 16.3 ).

Here we will focus our discussion on four lesions that pose a unique differential challenge when encountered in a prostatic biopsy: (1) stromal tumors of uncertain malignant potential (STUMPs) and stromal sarcomas, (2) smooth muscle neoplasms (leiomyomas/leiomyosarcomas), (3) solitary fibrous tumors (SFTs), and (4) gastrointestinal stromal tumors (GISTs).

Solitary Fibrous Tumor

Fewer than 20 cases of SFT involving the prostate have been reported. Some older reported cases of hemangiopericytoma of the prostate may also be today classified as SFT. Microscopically, prostatic SFTs appear similar to those identified in extraprostatic sites. Uniform spindled cells with bland nuclei are arranged in a “patternless” pattern in a background of variable ropy collagen and a hemangiopericytomatous appearance. None of the reported prostatic SFTs have behaved in an aggressive fashion. However, based on the behavior of SFTs in other sites and the finding in some prostatic SFTs of hypercellularity, pleomorphism, necrosis, and infiltrative margins, careful long-term clinical follow-up is warranted. IHC generally reveals diffuse reactivity for CD34, vimentin, and Bcl-2, although rare SFTs may lack some of these markers ( Fig. 16.19 ). Staining for CD99, β-catenin, p53, SMA, and MSA has also been reported. These tumors are typically negative for pancytokeratin, S100, and CD117 (c-kit).

(A and B) Solitary fibrous tumor involving prostate. (C) Strong CD34 positivity.

Gastrointestinal Stromal Tumor

Although GISTs may occur as a primary prostatic process on imaging and clinical studies, such cases are typically large masses that arise from the rectum or perirectal space that only compress the prostate. Very rarely, GIST may invade the prostate. There is not yet a fully documented example of a GIST arising within the prostate. Typically, GIST is not considered in the differential diagnosis of spindle cell lesions of the prostate, although the unique management of these tumors underscores the importance of recognizing them. Unfortunately, several patients have undergone pelvic exenteration, irradiation, and chemotherapy for a misdiagnosis of a GIST as a “pelvic sarcoma.” So-called prostatic GISTs present with urinary obstructive symptoms, rectal fullness, and abnormal DRE. Microscopically, they show identical features to lesions within the gastrointestinal tract. GIST is composed of spindle cells with a fascicular growth pattern, occasional epithelioid features, and focal dense collagenous stroma ( Fig. 16.20 ). When present, a fascicular or palisading growth pattern and perinuclear vacuoles along with a lack of collagen deposition aid in the discrimination of GISTs from SFTs and STUMPs. Tumors with malignant potential show elevated mitotic rates of more than 5 per 50 hpf, cytologically malignant features (high cellularity and overlapping nuclei), and/or necrosis. Immunohistochemically, CD117/c-kit is uniformly expressed in all cases, and CD34 is positive in almost all cases studied. S100, desmin, and SMA are negative. On prostate needle biopsy, before rendering a diagnosis of SFT, schwannoma, leiomyosarcoma, or stromal sarcoma, GIST should be considered in the differential diagnosis. Furthermore, immunostains for CD117 should be performed to verify the diagnosis. CD34 is not discriminatory, because it is positive in GISTs, SFTs, and specialized prostatic stromal tumors, and it is variably positive in schwannomas. However, it is typically negative in smooth muscle tumors. Strong positive staining for desmin can help discriminate smooth muscle tumors from other lesions. Similarly, positive immunoreactivity to S100 may aid in diagnosing neural tumors. SMA is typically expressed in smooth muscle tumors and is variably positive in STUMPs and GISTs and typically negative in SFTs and schwannomas.

(A and B) Gastrointestinal stromal tumor involving prostate biopsy. (C) Strong CD117 staining.

A subset of patients treated with the c-kit tyrosine kinase inhibitor imatinib (Gleevec) following the diagnosis of “prostatic” GIST demonstrated a subsequent reduction in tumor size.