Expression of CDX2 in benign tissue and adenocarcinoma of the prostate

Numerous studies have claimed that CDX2 is relatively specific and sensitive in establishing a gastrointestinal origin in metastatic tumors of unknown origin. We have recently seen 2 cases of prostatic adenocarcinoma (PCa) on needle biopsies with diffuse strong nuclear staining for CDX2 sent for consultation. One case was a prostatic duct adenocarcinoma in a man with a prostate-specific antigen (PSA) value of 327 ng/mL, and the other was a PCa with a Gleason score (GS) of 4 + 4 = 8 in a man with a PSA value of 15 ng/mL. An adenocarcinoma with GS 3 + 3 = 6 from the contralateral side did not express CDX2. Because documented examples of this phenomenon are rare, we investigated the immunoexpression of CDX2, using tissue microarrays (TMAs). Three slides of TMAs were used to stain 708 tissue samples (0.6 mm in diameter) containing either benign or malignant prostate tissue, as well as control tissues from various anatomical sites including colon. In total, 195 samples of primary PCa with GS of 6 (n = 41), 7 (n = 21), and 8 (n = 8); 195 samples of benign prostate tissue; and 185 samples of metastatic PCa were studied. Of 70 radical prostatectomy specimens examined for PCa in TMAs, 4 (5.7%) were positive for CDX2, showing Gleason score of 6 (n = 3) and Gleason score of 7 (n = 1). Focal moderate positive staining was seen in benign prostate tissue in 7 (11.7%) of 60 radical prostatectomy specimens. None of the metastatic PCa expressed CDX2. CDX2 may uncommonly be focally expressed in benign prostatic glands. Staining in PCa is less common and appears independent of GS and is usually patchy and focal and of lesser intensity than in colonic tissue. However, rarely strong and diffuse staining may be seen. Positive CDX2 staining in high-grade prostate cancer (ductal, cribriform, and solid) may be confused with secondary carcinoma of colonic origin. Routine histopathology, positive PSA immunostaining, and clinical findings can help confirm the correct diagnosis.     

Small glandular proliferations on needle biopsies: most common benign mimickers of prostatic adenocarcinoma sent in for expert second opinion

The current study aimed to determine the incidence of various benign mimickers of prostatic adenocarcinoma most commonly encountered in a busy consultation practice. All prostate needle biopsies from the consult service of one of the authors were prospectively evaluated over a 7-month period. Only cases with foci where the contributor questioned malignancy and which upon expert review the entire case was determined to be benign were included in this study. A total of 567 separate suspected atypical foci from 345 patients of a total of 4,046 patients (8.5%) received in consultation were identified. Of these, 281 foci (49.5%) had immunohistochemical (IHC) studies performed by the outside institution, which included high molecular weight cytokeratin (HMWCK) (n = 280), alpha-methylacyl-CoA racemase (AMACR) (P504s) (n = 45), and p63 (n = 34). The most common mimicker was partial atrophy (203 of 567; 35.8%). Technically adequate IHC for basal cells was performed in 117 cases of partial atrophy with patchy or patchy/negative staining seen in 102 of 117 (87%), with the remaining 13% of cases completely negative. A total of 15 of 19 (79%) cases of partial atrophy were positive with AMACR. Crowded benign glands, insufficiently crowded or numerous to warrant a diagnosis of adenosis, was the second most common mimicker (146 of 567; 25.7%). Crowded benign glands had patchy or patchy/negative IHC for basal cells in 66 of 81 (81%) cases with the remaining 19% of cases completely negative. A total of 7 of 11 (64%) cases of crowded glands were positive for AMACR. In the past, complete atrophy, adenosis, seminal vesicle, and granulomatous prostatitis were considered common mimickers of prostate cancer on prostatic needle biopsies. Our study shows that currently partial atrophy and crowded benign glands are the most common benign changes causing diagnostic difficulty and prompting consultation. Negative or patchy staining for basal cells and positive staining for AMACR may contribute to diagnostic difficulty in these entities.     

Urothelial carcinoma with rhabdoid features: report of 6 cases

Extrarenal rhabdoid tumors have been described in a variety of primary sites with only rare case reports of urothelial carcinomas with rhabdoid features in the literature. In this report, we describe the clinicopathologic characteristics, including clinical follow-up on 6 cases of urothelial carcinoma with prominent rhabdoid features. Four cases were retrieved from the consultation files of one of the authors and 2 were retrieved from the surgical pathology files at our institution. The patients were all men, with ages ranging from 53 to 86 years (mean, 66.5 years). Patients initially presented with hematuria or obstructive symptoms. The sites included bladder (n = 4) and renal pelvis (n = 2). All cases had a prominent rhabdoid component (mean, 60%), ranging from 40% to 80%. In addition to the rhabdoid component, multiple coexistent histological components were seen, including in situ urothelial carcinoma (carcinoma in situ) and high-grade papillary urothelial carcinoma (n = 2), poorly differentiated carcinoma with small-cell features (n = 1), sarcomatoid (n = 2), and a myxoid component (n = 2). All cases in this series had focal or diffuse positive staining with one or more cytokeratin markers (epithelial membrane antigen, CAM 5.2, AE1/AE3). Of the 6 patients, 4 were treated initially with surgery (radical cystoprostatectomy, n = 2; radical nephrectomy, n = 2). Of 6 patients, 2 died within 1 month, whereas a third patient died within 4 months. The remaining 3 patients were alive at 3, 3, and 9 months after diagnosis. The histological and immunohistochemical findings in this study serve to broaden the morphological spectrum of urothelial carcinomas with prominent rhabdoid features and add further evidence as to their poor prognosis.     

Actinomycotic Endocarditis of the Eustachian Valve: A Rare Case and a Review of the Literature

Eustachian valve endocarditis caused by Actinomyces species is extremely rare. A literature review revealed only one reported case—caused by Actinomyces israelii in an intravenous drug abuser.Our patient, a 30-year-old woman who at first appeared to be in good health, presented with fever, a large mobile mass on the eustachian valve, and extensive intra-abdominal and pelvic masses that looked malignant. Histopathologic examination of tissue found in association with an intrauterine contraceptive device revealed filamentous, branching microorganisms consistent with Actinomyces turicensis. This patient was treated successfully with antibiotic agents.In addition to presenting a new case of a rare condition, we discuss cardiac actinomycotic infections in general and eustachian valve endocarditis in particular: its predisposing factors, clinical course, sequelae, and our approaches to its management.     

Gastrointestinal stromal tumors (GISTs) on prostate needle biopsy: A clinicopathologic study of 8 cases

Gastrointestinal stromal tumors (GISTs) are typically not included in the differential diagnosis of spindle cell tumors seen on prostate needle biopsy. However, their recognition is critical due to their unique clinical management. We report the rare phenomenon of 8 cases of GISTs diagnosed on prostate needle biopsy. The mean patient age at diagnosis was 53.6 years (range: 42 to 65 years). Tumors variably presented with rectal fullness, urinary obstructive symptoms, and abnormal digital rectal examination. Four tumors were resected. One of these cases was shown to be primary in the rectum without prostatic involvement. The second case extensively involved the prostate but its epicenter was in the rectal muscularis propria. The third case was an encapsulated mass separated by a thin fibrous capsule from the prostate. The fourth case was a perirectal mass that underwent local excision. Four lesions have not been resected. On the basis of imaging studies, one seemed to be a prostatic mass, however, additional imaging investigations showed the mass to be separate from the prostate. Three cases have not yet been studied radiographically. Tumors measured 1.0, 1.7, 5.4, 7.0, 7.4, and 8.5 cm. The sizes of 2 recently diagnosed tumors remain undetermined. Histologically, all 8 GISTs showed spindled cells with a fascicular growth pattern. Additional histologic findings included focal epithelioid features (n = 3), necrosis (n = 3), mitotic rates of >5 per 50 high-power field (n = 2), and cytologically malignant features (n = 3). CD117/c-kit was diffusely positive in all 8 cases and CD34 in 7/8 cases. In all cases studied, stains for S100, desmin, and smooth muscle actin were negative. Two patients were treated with imatinib mesylate. One underwent radical prostatectomy after reduction in tumor size after imatinib administration. Another patient was treated with imatinib for several months with complete tumor response and no residual tumor seen in a subsequent local excision.Rectal or extraintestinal GIST can result in a clinical impression of a prostatic lesion. One should consider CD117/c-kit in the immunohistochemical panel to exclude GIST before diagnosing a solitary fibrous tumor, leiomyosarcoma, or specialized prostatic stromal tumor on prostate needle biopsy.     

Polyoma virus-associated cellular changes in the urine and bladder biopsy samples: a cytohistologic correlation

Polyoma (BK) virus-type cellular changes are occasionally reported in urine specimens, yet rarely detected in histologic sections of bladder biopsies. A total of 762 predominantly voided urine specimens with a cytologic diagnosis of polyoma virus-type change were retrieved from the cytopathology files of the Johns Hopkins Hospital over a 15-year period (1988-2003). Biopsies were available for 33 cases (29 patients) following or preceding the urinary cytology (mean interval, 210 days). The biopsies originated primarily from bladder (n=31) with one biopsy each from renal pelvis and urethra. Representative paraffin blocks were chosen from each case for immunoperoxidase staining with SV40 large T antigen. There were 22 males and 7 females with an age range of 34 to 79 years (mean, 64.7 years). The histologic diagnoses of the 33 tissue biopsies were: benign urothelium (n=9), urothelial carcinoma (n=21) and 1 case each of dysplasia, small cell carcinoma, and chronic lymphocytic leukemia involving lamina propria of the bladder. Only 3 of 33 biopsies on hematoxylin-stained sections showed morphologic changes of polyoma virus, which lacked sufficient tissue to perform immunohistochemistry for SV40 large T antigen (LTag). Immunohistochemical staining for LTag was positive in 7 cases. Only 2 cases showed strong/diffuse and moderate/focal staining for LTag with both representing invasive high-grade urothelial carcinoma (where no inclusions were seen on hematoxylin and eosin-stained sections) and both demonstrating positive immunostaining for p53. One of these 2 cases was from an organ transplant recipient and the other from a patient with no known immunosuppression. Our data lead to the following conclusions: 1) cytology appears to be more sensitive than histology in detecting cells with polyoma virus; 2) cytohistologic discordance might be due to: a) polyoma (BK) virus infected cells are shed from the tissue and collected in the urine; b) polyoma virus changes may be focal and not sampled in directed tissue biopsies; c) polyoma virus changes may originate from sites in the genitourinary tract other than the bladder; d) the lack of a "gold standard" to confirm the cytologic diagnoses of polyoma virus; and e) the discordance in time between the biopsy and cytology specimens in the current retrospective study. 3) Because some cytologically benign cases of polyoma virus were associated with malignant biopsies, careful morphologic evaluation is required to avoid false-negative urinary cytology samples. This investigation further examined the immunohistochemical staining pattern for SV40 LTag and p53 in noninvasive low-grade papillary urothelial carcinoma using a tissue microarray constructed from bladder biopsies of 79 randomly selected patients. Weak LTag staining was present in occasional neoplastic urothelial cells of 2 patients. The staining was present in only one of four samples from each tumor (0.63%; 2 of 316 tumor spots), which further illustrates the patchy and focal presence of virion containing cells. p53 staining in these two spots was also patchy and confined to rare nuclei.     

Risk of prostate cancer on first re-biopsy within 1 year following a diagnosis of high grade prostatic intraepithelial neoplasia is related to the number of cores sampled

PURPOSE: We determined the influence of the extent of needle biopsy sampling on the detection rate of cancer on first biopsy within 1 year following a diagnosis of HGPIN. MATERIALS AND METHODS: We identified 791 patients with HGPIN on the initial biopsy who had a followup biopsy within 1 year of their diagnosis. The mean interval from diagnosis of HGPIN to re-biopsy was 4.6 months. In the initial biopsy with HGPIN, 323 men had 8 or more cores (median 10, range 8 to 26) and 332 men had 6 core biopsies. RESULTS: In the 6 core initial sampling group, the risk of cancer on re-biopsy was 20.8% compared to only 13.3% following an initial 8 core or more sampling (p = 0.011). With 6 core biopsies for both the initial and re-biopsy the risk of cancer was 14.1% (group 1). With an initial 6 core biopsy and 8 core or more biopsy on followup, the risk of cancer was 31.9% (group 2). With 8 core or more biopsy sampling for both initial and repeat biopsies, the risk for cancer was 14.6% (group 3). The differences between groups 1 and 3 as compared to group 2 were statistically significant (p = 0.001 and p <0.0001, respectively). CONCLUSIONS: With relatively poor sampling (6 cores) on the initial biopsy, associated cancers are missed resulting in only HGPIN on the initial biopsy, and with relatively poor sampling on re-biopsy there is also a relatively low risk of finding cancer on re-biopsy (group 1). With poor sampling on the initial biopsy and better sampling on re-biopsy, some of these initially missed cancers are detected on re-biopsy yielding a higher detection of cancer (group 2). Sampling more extensively on the initial biopsy detects many associated cancers, such that when only HGPIN is found they often represent isolated HGPIN. Therefore, re-biopsy even with good sampling does not detect many additional cancers (group 3). Our study demonstrates that the risk of cancer on biopsy within 1 year following a diagnosis of HGPIN (13.3%) is not that predictive of cancer on re-biopsy if good sampling (8 or more cores) is initially performed. For patients diagnosed with HGPIN on extended initial core sampling, a repeat biopsy within the first year is unnecessary in the absence of other clinical indicators of cancer.