Distinguishing prostatic from colorectal adenocarcinoma on biopsy samples: the role of morphology and immunohistochemistry.
Objective.--To evaluate the role of morphology and immunohistochemistry in the previously mentioned setting.
Design.--Surgical pathology and consultation records. Hematoxylin-eosin sections were reviewed in 16 cases (11 PCa, 5 CRCa). Immunohistochemistry for 9 markers was performed in 15 cases.
Results.--Dirty necrosis, seen in 5 (100%) of 5 CRCa and 2 (18%) of 11 PCa cases, and the presence of columnar cells with basal nuclei, seen in 5 (100%) of 5 CRCa and 1 (9%) of 11 PCa cases, appear to be the most useful morphologic parameters. Immunohistochemistry confirmed the value of prostate-specific antigen (PSA), CDX2, cytokeratin (CK) 20, and [beta]-catenin in the differential of CRCa (0% PS[A.sup.+], 60% CDX[2.sup.+], 80% CK[20.sup.+], and 100% [beta]-[catenin.sup.+]) versus PCa (80% PS[A.sup.+], 0%CDX[2.sup.+], 10% CK[20.sup.+], and 0% [beta]-[catenin.sup.+]). P501S had a similar sensitivity as PSA in detecting PCa (80%). Two (20%) of 10 PCa cases were positive for 1 of the 2 markers but not the other. P501S was negative in all 5 cases of CRCa.
Conclusions.--P501S is a useful marker in this setting when included together with PSA, CDX2, CK20, and [beta]-catenin. P501S labels a subset of PCa cases that are negative for PSA. Dirty necrosis and/or columnar cells with basal nuclei could also be of help.
(Arch Pathol Lab Med. 2007;131:599-603)
Prostate carcinoma (PCa) and colorectal carcinoma (CRCa) constitute 2 of the 3 leading malignancies in males both in term of frequency of diagnoses and mortality. (1) Given the anatomic proximity of the 2 organs, locally advanced PCa can involve adjacent colorectal tissue. (2-4) Similarly, advanced CRCa can rarely invade into prostate. (5,6) Obtaining an accurate diagnosis in such cases could prove difficult on a small biopsy sample of either organ. Such a distinction is important given the difference in treatment and prognosis between PCa and CRCa. (4) The current study aimed at identifying potential morphologic and immunohistochemical features that can help in this differential.
MATERIALS AND METHODS
The surgical pathology and consultation files at our institution were searched for colorectal and prostate biopsy specimens in which the differential of CRCa versus PCa was raised clinically and/or histologically. A total of 27 cases of poorly differentiated carcinoma were identified representing either PCa extending into colorectum, CRCa extending into prostate, or colorectal biopsies with poorly differentiated carcinoma in patients who had a history of PCa. The majority of the 27 cases were from outside institutions that were sent in consultation to one of us (J.I.E.). We were able to obtain hematoxylin-eosin sections in 16 cases. In 15 of the 16 cases, paraffin blocks were also available for immunohistochemical analysis.
All cases were reviewed by two of us (C.L.O. and G.J.N.) without knowledge of the final diagnoses. A consensus was reached on discrepant cases. The evaluated histologic parameters included the presence of dirty necrosis, columnar cell morphology with basally located nuclei, microacinar architecture, lymphovascular involvement, mucin production, and presence of prominent nucleoli (easily discernible at x20 objective magnification).
Immunohistochemical analysis was performed on formalin-fixed, paraffin-embedded tissue using standard technique with the antibodies listed in Table 1. Antigen unmasking was performed on some of the specimens using sodium citrate buffer (pH 6.0) at 100[degrees]C (Black and Decker Flavor Scenter Steamer Plus) for 20 minutes. Slides were developed with either Dako EnVision+ (K4061, DakoCytomation, Carpinteria, Calif) or I-VIEW DAB detection kit (760-091, Ventana Medical Systems, Inc, Tucson, Ariz) on Benchmark XT (Ventana), immunohistochemical staining module systems and counterstained with hematoxylin. CDX2 and [beta]-catenin were considered positive only when expressed in the nuclei, and cytoplasmic and/or membrane staining was considered positive with all the other markers.
Mean patient age was 64 years (range, 42-88) for the CRCa group and 74 years (range, 65-85) for PCa patients. A wide overlap in patient age range was present among the 2 groups.
The results of selected histologic parameters are summarized in Table 2. Of the 6 histologic parameters, the presence of dirty necrosis, seen in 5 (100%) of 5 cases of CRCa and 2 (18%) of 11 cases of PCa, and the presence of columnar cells with basal nuclei, seen in 5 (100%) of 5 cases of CRCa and 1 (9%) of 11 cases of PCa, were the most useful parameters. Microacinar formation was seen in only 6 (55%) of 11 cases of PCa and 1 (20%) of 5 cases of CRCa.
The immunohistochemical results for all 15 cases are detailed in Table 3 and summarized in Table 4. Prostate-specific antigen (PSA) was expressed in 80% of PCa and none of the CRCa cases. CDX2 (nuclear) was present in 60% of CRCa and none of the PCa cases. CK20 was expressed in 80% of CRCa and 10% of PCa (Figure 1, a through f) cases. Nuclear [beta]-catenin expression was encountered in all 5 CRCa but in none of the 10 cases of PCa (Figure 2, a through g).
[FIGURES 1-2 OMITTED]
In most cases with P501S expression, the expression pattern was cytoplasmic with a dotlike appearance. P501S was expressed in 8 (80%) of 10 PCa cases. Two of the 10 PCa cases revealed a discordant P501S and PSA expression. Case 7 was positive for P501S but negative for PSA. The opposite pattern of staining was encountered in case 12 (PSA positive/P501S negative). Only 1 (14%; case 8) of the 7 PCa cases was negative for both P501S and PSA (Figure 3). P501S was negative in all 5 cases of CRCa.
[FIGURE 3 OMITTED]
Prostate-specific acid phosphatase (PSAP) was expressed in only 40% of PCa. The majority of both PCa and CRCa cases were positive for P504S. Although both pCEA and CK7 were expressed in the majority of CRCa cases, they also labeled 30% of PCa cases.
Locally advanced cases of PCa and CRCa can present a diagnostic challenge as to the primary site on a small colorectal or prostatic biopsy containing a poorly differentiated carcinoma. As expected by the proximity of the 2 organs, locally advanced PCa can lead to secondary involvement of the rectum or perirectal lymph nodes. (2-4) The reported rates of such occurrence in autopsy studies are in the range of 1% to 12%. (3) Similarly, advanced CRCa can rarely extend to prostate. (5,6) In this setting, accurate assignment of the primary source has significant therapeutic and prognostic implications and will help avoid inappropriate surgery because advanced PCa is potentially responsive to androgen hormone withdrawal. (4) Other therapeutic modalities may be pursued in advanced CRCa including new targeted therapies such as epidermal growth factor receptor inhibitors. (7)
In general, morphologic features have played only a limited role in helping the pathologist assign accurate origin to a carcinoma of "unknown primary origin."
An expanding list of immunohistochemical and molecular diagnostic markers are being used in such cases. (8,9) In the current study, we evaluated the potential role of morphology and immunohistochemistry in resolving a primary origin of a poorly differentiated adenocarcinoma obtained in the setting of a locally advanced tumor in the colorectal-prostatic region.
Among the morphologic features evaluated, dirty necrosis and/or columnar cells with basal nuclei favored CRCa over PCa with all 5 cases of CRCa at least focally showing these features. However, dirty necrosis and columnar cell morphology with basal nuclei were also seen in rare cases of PCa (2 cases and 1 case, respectively). The latter is not totally unexpected given the fact that ductal variant of PCa can share such morphologic features with CRCa. Indeed the 1 PCa case in our group that demonstrated both of these features had an overall appearance consistent with ductal prostatic carcinoma. Therefore, one cannot totally rely on the presence of such features in accurately assigning the primary site of origin in this setting.
Our study evaluated the role of several immunohistochemical markers that have demonstrated tissue specificity to either PCa or CRCa and have been used by others in similar settings. These markers included PSA, PSAP, CK20, CK7, P501S, CDX2, and [beta]-catenin. (2,8,10-12) Among these, PSA has demonstrated great utility as a prostate tissue-specific marker in the workup of carcinoma of uncertain primary in older men. (13,14) However, it is well recognized that PSA is neither entirely sensitive nor specific for PCa. High-grade prostatic adenocarcinomas may lose PSA expression or express it focally, such that it is absent on limited biopsy samples, and nonprostatic tumors including salivary gland, breast, and pancreatic neoplasms can be PSA positive. (15-18) New prostate tissue-specific immunohistochemical markers used in combination with PSA could potentially enhance the sensitivity and specificity of PSA, especially in the setting of metastatic tumors and poorly differentiated locally advanced tumors involving urinary bladder or colorectum. In this setting, we evaluated the role of prostein, a novel potentially prostate-specific marker that has been recently identified using high throughput cDNA microarray analysis. (19) P501S, a monoclonal antibody that recognizes a linear intracellular epitope on prostein, has become commercially available. Using immunohistochemical analysis on tissue microarray analysis, Kalos et al (20) have shown P501S to be expressed by most primary and metastatic PCa. The authors also report that none of 4635 tissue specimens expressed prostein, including normal and malignant tissue from a wide variety of organ systems. In our series, prostein was equally sensitive in detecting prostatic differentiation as PSA. In 2 cases of PCa in our series, case 7 (PS[A.sup.-], PSA[P.sup.-], and P501[S.sup.+]) and case 12 (PS[A.sup.+], PSA[P.sup.-], and P501[S.sup.-]), combining both P501S and PSA in a routine panel was helpful in increasing the overall sensitivity of detecting evidence of prostatic differentiation in the setting of a high-grade tumor. To the best of our knowledge, our study is the first to evaluate the utility of prostein in this clinical setting. In our study, both PSA and prostein outperformed PSAP, in which its low sensitivity was at least in part due to the use of a monoclonal rather than a polyclonal antibody. (21) Our findings in regard to the utility of PSA, CK20, CDX2, and [beta]-catenin are in line with those of prior studies including our recent report on distinguishing distorted benign rectal mucosa from PCa. (2,8,22,23)
A single case of PCa in our series (case 8) was negative for all 3 markers of prostatic differentiation: PSA, PSAP, and prostein. All other markers including CDX2 and [beta]-catenin were also negative. As shown in Figure 3, the tumor demonstrated small cell morphology, which could have resulted in its lack of PSA and conceivably P501S expression. (24) Unfortunately, because of tissue exhaustion, we were unable to obtain neuroendocrine markers. This patient had a prior history of a biopsy proven low-grade prostate cancer. His tumor was left unresected and he later presented with a locally advanced tumor contiguous to his original PCa, the biopsy of which is included in our series.
None of the 5 cases of CRCa expressed P501S. This finding is not surprising considering the aforementioned large tissue-microarray study that found no evidence of P501S expression in any of the extraprostatic tissues examined including 364 colon specimens (355 colonic adenocarcinomas and adenomas and 9 normal colonic tissue). (20)
CDX2 and CK20 were expressed in 60% and 80% of CRCa cases, respectively, and in 0% and 10% of PCa cases, respectively, attesting to their utility in differentiating CRCa from PCa. However, we have recently encountered CDX2 positivity in rare cases of PCa (J.I.E., unpublished data, October 2006). Nuclear [beta]-catenin was observed in all 5 CRCa and none of the 10 PCa cases. Several of our cases of PCa had membranous and cytoplasmic staining with [beta]-catenin. Membranous [beta]-catenin expression has been shown by Bismar et al (10) in up to 88% of PCa, but nuclear staining was not identified in any of the cases.
In conclusion, a panel of PSA, P501S, CDX2, CK20, and [beta]-catenin may help in differentiating PCa from CRCa in biopsies of poorly differentiated carcinoma involving the colorectal-prostatic region. P501S stain a subset of PCa cases that are negative for PSA. Using these 2 markers in combination can increase the overall sensitivity of detecting evidence of prostatic differentiation in the setting of high-grade locally advanced or metastatic tumors.
(1.) Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin. 2005;55:10-30.
(2.) Murray SK, Breau RH, Guha AK, Gupta R. Spread of prostate carcinoma to the perirectal lymph node basin: analysis of 112 rectal resections over a 10-year span for primary rectal adenocarcinoma. Am J Surg Pathol. 2004;28:1154-1162.
(3.) Bowrey DJ, Otter MI, Billings PJ. Rectal infiltration by prostatic adenocarcinoma: report on six patients and review of the literature. Ann R Coll Surg Engl. 2003;85:382-385.
(4.) Bailey CM, Gilbert JM. Avoiding inappropriate surgery for secondary rectal cancer. Eur J Surg Oncol. 2002;28:220-224.
(5.) Gupta T, Laskar SG, Thakur M, et al. Isolated prostatic metastasis from primary sigmoid colon carcinoma. Indian J Gastroenterol. 2004;23:114-115.
(6.) Mai KT, Landry DC, Collins JP. Secondary colonic adenocarcinoma of the prostate histologically mimicking prostatic ductal adenocarcinoma. Tumori. 2002; 88:341-344.
(7.) Kumar R. Commentary: targeting colorectal cancer through molecular biology. Semin Oncol. 2005;32:S37-S39.
(8.) Dennis JL, Hvidsten TR, Wit EC, et al. Markers of adenocarcinoma characteristic of the site of origin: development of a diagnostic algorithm. Clin Cancer Res. 2005;11:3766-3772.
(9.) Netto GJ. Tumors of unknown primary origin: the role of immunohistochemistry and molecular diagnostics. In: Perry MC, ed. American Society of Clinical Oncology 2002 Educational Book. Baltimore, Md: Lippincott Williams & Wilkins; 2002:343-347.
(10.) Bismar TA, Humphrey PA, Grignon DJ, Wang HL. Expression of beta-catenin in prostatic adenocarcinomas: a comparison with colorectal adenocarcinomas. Am J Clin Pathol. 2004;121:557-563.
(11.) Magi-Galluzzi C, Luo J, Isaacs WB, Hicks JL, de Marzo AM, Epstein JI. Alpha-methylacyl-CoA racemase: a variably sensitive immunohistochemical marker for the diagnosis of small prostate cancer foci on needle biopsy. Am J Surg Pathol. 2003;27:1128-1133.
(12.) Moskaluk CA, Zhang H, Powell SM, Cerilli LA, Hampton GM, Frierson HF Jr. Cdx2 protein expression in normal and malignant human tissues: an immunohistochemical survey using tissue microarrays. Mod Pathol. 2003;16:913-919.
(13.) Stein BS, Vangore S, Petersen RO. Immunoperoxidase localization of prostatic antigens: comparison of primary and metastatic sites. Urology. 1984;24:146-152.
(14.) Hameed O, Humphrey PA. Immunohistochemistry in diagnostic surgical pathology of the prostate. Semin Diagn Pathol. 2005;22:88-104.
(15.) Fan CY, Wang J, Barnes EL. Expression of androgen receptor and prostatic specific markers in salivary duct carcinoma: an immunohistochemical analysis of 13 cases and review of the literature. Am J Surg Pathol. 2000;24:579-586.
(16.) Black MH, Magklara A, Obiezu C, et al. Expression of a prostate-associated protein, human glandular kallikrein (hK2), in breast tumours and in normal breast secretions. Br J Cancer. 2000;82:361-367.
(17.) Elgamal AA, Ectors NL, Sunardhi-Widyaputra S, Van Poppel HP,Van Damme BJ, Baert LV. Detection of prostate specific antigen in pancreas and salivary glands: a potential impact on prostate cancer overestimation. J Urol. 1996;156: 464-468.
(18.) Alanen KA, Kuopio T, Koskinen PJ, Nevalainen TJ. Immunohistochemical labelling for prostate specific antigen in non-prostatic tissues. Pathol Res Pract. 1996;192:233-237.
(19.) Xu J, Kalos M, Stolk JA, et al. Identification and characterization of prostein, a novel prostate-specific protein. Cancer Res. 2001;61:1563-1568.
(20.) Kalos M, Askaa J, Hylander BL, et al. Prostein expression is highly restricted to normal and malignant prostate tissues. Prostate. 2004;60:246-256.
(21.) Shevchuk MM, Romas NA, Ng PY, Tannenbaum M, Olsson CA. Acid phosphatase localization in prostate carcinoma: a comparison of monoclonal antibody to heteroantisera. Cancer. 1983;52:1642-1646.
(22.) Hameed O, Sublett J, Humphrey PA. Immunohistochemical stains for p63 and alpha-methylacyl-CoA racemase, versus a cocktail comprising both, in the diagnosis of prostatic carcinoma: a comparison of the immunohistochemical staining of 430 foci in radical prostatectomy and needle biopsy tissues. Am J Surg Pathol. 2005;29:579-587.
(23.) Schowinsky JT, Epstein JI. Distorted rectal tissue on prostate needle biopsy: a mimicker of prostate cancer. Am J Surg Pathol. 2006;30:866-870.
(24.) Helpap B, Kollermann J. Undifferentiated carcinoma of the prostate with small cell features: immunohistochemical subtyping and reflections on histogenesis. Virchows Arch. 1999;434:385-391.
Accepted for publication September 14, 2006.
Christopher L. Owens, MD; Jonathan I. Epstein, MD; George J. Netto, MD
From the Department of Pathology, The Johns Hopkins Hospital, Baltimore, Md.
The authors have no relevant financial interest in the products or companies described in this article.
Presented in part at the annual meeting of the United States and Canadian Academy of Pathology, Atlanta, Ga, February 2006.
Reprints: George Netto, MD, Johns Hopkins Medical Institutions, Pathology and Laboratory Medicine, 401 N Broadway, Room 2242, Baltimore, MD 21231 (e-mail: firstname.lastname@example.org).
Table 1. Summary of Antibodies and Manufacturers * Antigen Antibody/Clone Source Dilution Retrieval [beta]-Catenin BD Biosciences, 1:1000 CB (14) San Jose, Calif CDX2 (CDX2-88) BioGenex, San 1:100 CB Ramon, Calif CEA-p (polyclonal) Dako, Carpinteria, Predilute CB Calif PSA (polyclonal) Ventana, Tucson, Predilute CB Ariz PSAP (PASE 14LJ) Dako Predilute CB Racemace (AMACR Zeta, Sierra 1:100 CB P504S) Madre, Calif P501 S (10E3) Dako 1:100 CB Cytokeratin 7 Dako Predilute None (OV-TL 12/30) Cytokeratin 20 Ventana Predilute None (Ks20.8) * CB indicates citrate buffer method described in "Materials and Methods"; CEA, carcinoembryonic antigen; PSA, prostate-specific antigen; and PSAP, prostate-specific acid phosphatase. Table 2. Summary of Histologic Features on Hematoxylin-Eosin Sections From 16 Cases of Prostate Carcinoma (PCa) and Colorectal Carcinoma (CRCa) Columnar Dirty Microacinar Cells With Necrosis, Architecture, Basal Nuclei, Diagnosis No. (%) No. (%) No. (%) PCa (n = 11) 2 (18) 6 (55) 1 (9) CRCa (n = 5) 5 (100) 1 (20) 5 (100) Prominent Lymphovascular Nucleoli, Mucin, Invasion, Diagnosis No. (%) No. (%) No. (%) PCa (n = 11) 10 (91) 1 (9) 3 (27) CRCa (n = 5) 1 (20) 1 (20) 2 (40) Table 3. Immunohistochemical Findings in 10 Cases of Prostate Carcinoma (PCa) and 5 Cases of Colorectal Carcinoma (CRCa) * Case Prostein No. Specimen Diagnosis (P501 S) PSA PSAP 1 Colon Bx([dagger]) CRCa - - - 2 TURP CRCa - - - 3 Prostate Bx CRCa - - - 4 Prostate Bx CRCa - - - 5 Colon Bx([dagger]) CRCa - - - 6 Colon Bx PCa + + - 7 Colon Bx PCa + - - 8 Colon Bx PCa - - - 9 Colon Bx PCa + + - 10 Colon Bx PCa + + + 11 Colon Bx PCa + + - 12 Colon Bx PCa - + - 13 Colon Bx PCa + + + 14 Colon Bx PCa + + + 15 Colon Bx PCa + + + Case R-Catenin No. P504S CK7/CK20 pCEA CDX2 (Nuclear) 1 + +/- - - + 2 + -/+ + + + 3 + +/+ + + + 4 + +/+ + - + 5 - -/+ + + + 6 + +/+ + - - 7 - -/- - - - 8 - -/- - - - 9 + -/- - - - 10 + -/- + - - 11 - -/- - - - 12 - +/- - - - 13 + +/- + - - 14 + -/- - - - 15 + -/- - - - * PSA indicates prostate-specific antigen; PSAP, prostate-specific acid phosphatase; CK, cytokeratin; CEA, carcinoembryonic antigen; Bx, biopsy; -, negative; +, positive; and TURP, transurethral resection of prostate. ([dagger]) Both patients had a history of PCa. Table 4. Summary of Immunohistochemical Findings in 15 Cases of Prostate Carcinoma (PCa) and Colorectal Carcinoma (CRCa) * Prostein, PSA, PSAP, P504s, Diagnosis No. (%) No. (%) No. (%) No. (%) PCa (n = 10) 8 (80) 8 (80) 4 (40) 6 (60) CRCa (n = 5) 0 (0) 0 (0) 0 (0) 4 (80) CK7:CK20, pCEA, CDX2, beta]-Catenin, Diagnosis No. (%) No. (%) No. (%) No. (%) PCa (n = 10) 3 (30):1 (10) 3 (30) 0 (0) 0 (0) CRCa (n = 5) 3 (60):4 (80) 4 (80) 3 (60) 5 (100) * PSA indicates prostate-specific antigen; PSAP, prostate-specific acid phosphatase; CK, cytokeratin; and CEA, carcinoembryonic antigen.