Interobserver reproducibility of Gleason grading of prostatic carcinoma: urologic pathologists.

Gleason grading is now the most widely used grading system for prostatic carcinoma in the United States. However, there are only a few studies of the interobserver reproducibility of this system, and no extensive study of interobserver reproducibility among a large number of experienced urologic pathologists exists. Forty-six needle biopsies containing prostatic carcinoma were assigned Gleason scores by 10 urologic pathologists. The overall weighted kappa coefficient kappa(w) for Gleason score for each of the urologic pathologists compared with each of the remaining urologic pathologists ranged from 0.56 to 0.70, all but one being at least 0.60 (substantial agreement). The overall kappa coefficient kappa for each pathologist compared with the others for Gleason score groups 2-4, 5-6, 7, and 8-10 ranged from 0.47 to 0.64 (moderate-substantial agreement), only one less than 0.50. At least 70% of the urologic pathologists agreed on the Gleason grade group (2-4, 5-6, 7, 8-10) in 38 ("consensus" cases) of the 46 cases. The 8 "nonconsensus" cases included low-grade tumors, tumors with small cribriform proliferations, and tumors whose histology was on the border between Gleason patterns. Interobserver reproducibility of Gleason grading among urologic pathologists is in an acceptable range.

Interobserver reproducibility of Gleason grading of prostatic carcinoma: general pathologist.

Only a few large studies of interobserver reproducibility of Gleason grading of prostatic carcinoma exist. Thirty-eight biopsies containing prostate cancer were distributed for Gleason grading to 41 general pathologists in Georgia. These cases had "consensus" Gleason grade groups (2-4, 5-6, 7, and 8-10) that were agreed on by at least 7 of 10 urologic pathologists. The overall kappa (kappa) coefficient for interobserver agreement for these 38 cases was 0.435, barely moderate agreement, with a kappa range from 0.00 to 0.88. There was consistent undergrading of Gleason scores 5-6 (47%), 7 (47%) and, to a lesser extent, 8-10 (25%). In cases with consensus primary patterns, there was consistent undergrading of patterns 2 (32%), 3 (39%), and 5 (30%). Pattern 2 was often (17%) mistaken for pattern 3. Pattern 4 was often undergraded (21%) and also mistaken for pattern 5 (17%). The most significant (P < .005) demographic factor associated with better interobserver agreement was having learned Gleason grading at a meeting or course. We believe that Gleason grading can be learned to a satisfactory level of interobserver reproducibility and have undertaken additional studies that support this belief.

A web-based tutorial improves practicing pathologists' Gleason grading of images of prostate carcinoma specimens obtained by needle biopsy: validation of a new medical education paradigm.

BACKGROUND: The Internet, although it is in widespread use in medicine, to the authors' knowledge has not been tested rigorously as an educational tool. The authors investigated, as a model to validate web-based education for physicians, the Gleason grading of images of prostate carcinoma tissue specimens that were obtained by needle biopsy, which provides critical information for patient management. METHODS: A free, web-based program (available at www. pathology.jhu.edu/prostate) was developed. It consisted of 20 pretutorial quiz images of prostate carcinoma specimens that were obtained by needle biopsy for grading, followed by 24 tutorial images with text describing the Gleason grading system. Subsequently, pathologists took a posttutorial quiz, which consisted of the same 20 images that were used in the pretutorial quiz. RESULTS: In 16 months, there were 2021 visits with 916 participants completing the entire web site; 643 participants (70.2%) were practicing pathologists and formed the basis of the current study. Only the location of practice within the United States compared with outside the United States (P < 0.0001) and < 5 years in practice (P = 0.003) were correlated independently with a higher pretutorial quiz score. Overall, the web-based tutorial significantly improved grading in 15 of 20 images. Of these, on average, there was an 11.9% increase (range, 6-25.3%) in assigning the correct Gleason score. Improvements were noted in images of tumors with the following grades: Gleason score, 2-4 (0 of 1 images); Gleason score, 5-6 (5 of 7 images); Gleason score, 7 (4 of 6 images); and Gleason score, 8-10 (6 of 6 images). Greater improvement after taking the tutorial was correlated with lower pretutorial scores (P < 0.0001). CONCLUSIONS: A web-based tutorial improves the accuracy of Gleason grading of practicing pathologists. To the authors' knowledge the current study is the first large scale, international study that has evaluated and validated the use of a web-based program to educate a population of widely dispersed physicians.

Pathology residents' use of a Web-based tutorial to improve Gleason grading of prostate carcinoma on needle biopsies.

Little is known about pathology residents' ability to Gleason grade or their ability to learn surgical pathology using Internet-based technology. A free Web-based program (available at www.pathology. jhu.edu/prostate) was developed that consisted of 20 pretutorial images for grading, 24 tutorial images, and the same 20 posttutorial images for Gleason grading. The grading images were selected from cases that had a consensus Gleason grade from 10 uropathology experts. In 2.5 months, 255 residents visited the website, and 151 (59%) completed it. Of those who completed the website, their year in training was known in 85 (56%): 1st year, 25.8%; 2nd year, 20%; 3rd year, 22.3%; 4th year, 14.1%; 5th year, 15.3%; and 6th year, 2.4%. Eighty percent learned Gleason grading in residency versus being self-taught, and 66% were male. In a multivariate analysis, higher pretutorial scores were associated with both their year in training (P = .001) and their hospital size (P = .003). Improvements in grading posttutorial were not related to the residents' year in training. Overall, the website significantly improved grading in 11 of 20 images and had no effect in 9 of 20 images. Improvements were noted in 1 of 1 Gleason score 4; 2 of 7 Gleason score 5 to 6; 2 of 6 Gleason score 7; and 6 of 6 Gleason score above 7 tumors. In summary, a Web-based tutorial improved Gleason grading accuracy by pathology residents to an equal extent regardless of their year in training. It is more difficult to teach residents to grade Gleason scores 5 to 7 tumors, and additional training should be concentrated in this area.

Small cell carcinoma of urinary bladder is differentiated from urothelial carcinoma by chromogranin expression, absence of CD44 variant 6 expression, a unique pattern of cytokeratin expression, and more intense gamma-enolase expression.

AIMS: Small cell (neuroendocrine) carcinoma of the urinary bladder is clinically more aggressive than urothelial (transitional cell) carcinoma. We have investigated the immunohistochemical markers most useful in diagnosing small cell carcinoma in bladder. METHODS AND RESULTS: We evaluated the expression of chromogranin A, CD44 variant 6 (CD44v6), cytokeratin (CAM 5.2), gamma-enolase, synaptophysin, and CD45 in 46 small cell carcinomas of the bladder. Small cell and urothelial carcinoma were mixed in 21 (46%) cases. The two immunohistochemical markers with best ability to discriminate between small cell and urothelial carcinoma were chromogranin A and CD44v6. Chromogranin A had 97% specificity for small cell carcinoma, staining 65% of cases with 2+/3+ mean intensity; only one case (5%) of urothelial carcinoma was weakly (1+/3+) positive. CD44v6 was 80% specific for urothelial carcinoma, with immunoreactivity in 60% of cases, compared with 7% of small cell carcinoma cases. In cases positive for CD44v6, the mean percentage of reactive urothelial carcinoma cells was 75% (range 10-100%), greater than the 12% of cells in three cases of small cell carcinoma (P = 0.31); further, the pattern of immunoreactivity was membranous vs. focal cytoplasmic, respectively. All small cell carcinomas stained with one of the three neuroendocrine markers tested; 76% of cases were reactive for synaptophysin and 93% for gamma-enolase, with specificities of 86% and 73% in comparison to urothelial carcinoma. gamma-enolase staining of small cell carcinoma was more intense (P = 0.01) than for urothelial carcinoma. Cytokeratin CAM 5.2 stained a mean 47% of cells in small cell carcinoma, always in a punctate perinuclear pattern, and 75% in urothelial carcinoma, in a membranous pattern. CONCLUSIONS: CD44v6, chromogranin A, and possibly gamma-enolase and cytokeratin (CAM 5.2) help differentiate small cell carcinoma from urothelial carcinoma.

Needle biopsies of the prostate: what constitutes adequate histologic sampling?

OBJECTIVE: The automated biopsy gun and increased screening for adenocarcinoma of the prostate have led to increased numbers of biopsies with only tiny foci of prostatic carcinoma. Consequently, the risk of failing to sample a small focus of carcinoma histologically has increased as well. Most pathologists routinely sample prostatic needle biopsies at more than 1 level. An expert panel has recently suggested that prostatic needle biopsies be sampled at at least 2 levels. However, there have been no studies measuring the amount of additional tissue sampled by multiple levels versus 1 level. METHODS: Forty-two prostatic needle biopsies were serially sectioned at 4-microm levels. Hematoxylin-eosin-stained slides were prepared from every fifth section. The total length of each biopsy was compared with the length sampled by 1 level (50% through the block) and 3 levels (25%, 50%, and 75% through the block). RESULTS: Sampling the tissue at 1 level missed an average of 23.4% of the total biopsy length. Sampling the tissue at 3 levels significantly improved this average to 7% (P = .0001). CONCLUSIONS: This study shows that a single histologic section of a prostatic needle biopsy often fails to sample a significant portion of available tissue. This could occasionally result in failure to sample a small focus of prostatic carcinoma. The authors recommend that prostatic needle biopsies be routinely sampled at 3 levels (approximately 25%, 50%, and 75% through the block).

p16 (CDKN2/MTS1) gene deletions are rare in prostatic carcinomas in the United States and Japan.

PURPOSE: The incidence of clinically apparent prostatic carcinoma is much higher in the United States than in Japan. Alterations in the p16 tumor suppressor gene have been identified in various tumor types, including cultured prostatic carcinoma cell lines. We studied the possible deletions of either exon 2 or 3 of this gene in primary clinical prostatic carcinomas from Japan and the United States. MATERIALS AND METHODS: Genomic DNA was extracted from 36 formalin-fixed, paraffin-embedded clinical prostatic carcinomas from Japan and 27 carcinomas from the United States. Exons 2 and 3 of the p16 gene were amplified using comparative multiplex polymerase chain reactions (PCR) and then analyzed for possible deletions of either exon. RESULTS: Two out of 36 (5.6%) carcinomas from Japan clearly demonstrated deletion of p16 exon 2, but this deletion was not detected in any of the 27 carcinomas from the United States. CONCLUSIONS: Although slightly higher in Japan than in the United States, the frequency of p16 exon deletions in clinical prostatic carcinomas is very low, and probably is not important in the development of this neoplasm.

Hair granuloma of the prostate.

We report the second case of so-called hair granuloma of the prostate in a transurethral resection specimen. We hypothesize that the hair was most likely embedded in bladder/neck/prostate tissue by an earlier perineal prostate needle biopsy.

Prevalence of androgen receptor gene mutations in latent prostatic carcinomas from Japanese men.

The incidence rate of clinically apparent prostatic carcinoma is 8-fold higher in the United States than in Japan, while the prevalence of latent prostatic carcinoma, a presumed precursor to clinical carcinoma, is similar in the two countries. The purpose of this study was to investigate the hypothesis that this profound difference in incidence rates of clinical carcinoma reflects distinct profiles of molecular genetic alterations in the latent precursor lesions that occur in the two countries. A significant fraction of latent carcinomas from Japanese men were found to contain inactivating mutations of the androgen receptor gene, while no such mutations were found in latent carcinomas from American men. No mutations were found in clinical carcinomas from either country. These data offer a potential molecular genetic explanation that may partially account for the distinct prostatic carcinoma incidence rates in these two populations.

Pilomatrical carcinosarcoma of the cheek with subsequent pulmonary metastases. A case report.

We report a case of malignant pilomatrixoma with a pronounced biphenotypic morphology. The lesion, which was excised from the cheek of a 36-year-old man, was composed of a large pilomatrixoma lying within a spindled, sarcomatoid stroma. Fourteen months later, the tumor metastasized to the right upper lobe of the lung. We describe the tumor's pathology, histology, and immunochemistry and discuss the differential diagnosis. We also speculate on its histogenesis.

Histochemistry of the prostate.

Histochemistry, including immunohistochemistry, is helpful to the practicing pathologist in the diagnosis of prostatic carcinoma. Of equal importance, histochemistry is being increasingly used to study the pathobiology of the prostate. This article reviews these histochemical techniques and their applications.

Cytology and DNA flow cytometry of peritoneal washings in gynecologic patients.

Peritoneal washing cytology, widely used in the management of gynecologic malignancy, entails several difficulties in interpretation. Quantitative DNA analysis by flow cytometry (FCM) holds promise as a more objective method fo diagnosis of malignancy. We performed traditional cytologic examination and single-parameter FCM DNA analysis on peritoneal washings from 136 gynecologic laparotomies, compared these results with the final pathologic findings, and analyzed sources of error. A total of 50 laparotomies were performed for benign disease. Another 86 were performed for cervical, endometrial, and ovarian carcinomas and various other cancers. In the benign group, cytology had one false suspicious but no false positive results, and FCM showed only diploid cells. In the cancer cases, cytology had five suspicious and 13 positive results and one false negative from laboratory error. On review, 16 washings contained confirmed cancer cells. FCM, performed in 13 of these cases, was diploid in 10 and aneuploid in only 3. In six of the diploid cases, visual cell counts showed that tumor cells were present in concentrations of 2.5% or less of total cells. In the remaining four diploid cases, a second DNA determination was obtained by FCM of nuclei retrieved from paraffin blocks of the tumors. These nuclei were diploid by FCM in three of the tumors and aneuploid in only one. Single-parameter DNA FCM was too insensitive to be helpful in our material.

Flow cytometric analysis of prostatic carcinoma with and without bone marrow metastasis.

We have analyzed 68 prostates obtained at autopsy, for DNA ploidy by means of flow cytometry from patients who had clinical prostatic carcinoma with (29 cases) and without (39 cases) bone marrow metastasis. Flow cytometric analysis revealed 42 diploid cases and 26 aneuploid cases in a total of 68 cases. Among the 26 cases of aneuploidy, 4 cases were tetraploid aneuploid and 22 cases were not tetraploid aneuploid. The highest incidence for the aneuploidy was found in stage D2 disease (19/29, 65.5%), while the highest incidence for diploid was seen in stage B disease (23/25, 92.0%). The correlation between ploidy and bone marrow metastasis was significant (P < 0.01). We thus reconfirm that flow cytometric analysis of paraffin embedded tissue is useful for prognostic evaluation of prostatic carcinoma.

Cytologic detection of urothelial cancer and other abnormalities in a cohort of workers exposed to aromatic amines.

A total of 541 members of a cohort at increased risk for occupational bladder cancer underwent a 33-month program of screening with urine cytology. Selected workers received further urologic study with cystoscopy and bladder biopsies. Eight workers had positive or suspicious cytologic findings. Only one of the eight had a prior history of bladder cancer. Biopsies showed invasive carcinoma and/or nonpapillary carcinoma in situ in five workers in this group, severe atypia in one, and no significant abnormality in two. Of 56 workers who had atypical cytologic findings, 16 had bladder biopsies, which showed atypia of flat urothelium in 11, nonpapillary carcinoma in situ in one, noninfiltrating papillary carcinoma in one, and no significant abnormality in three. The cytologic detection of urothelial abnormalities often required more than a single specimen. Since the cohort in this study may develop more bladder cancers with the passage of time, continued follow-up is indicated.

Flow cytometric analysis of granulosa cells from developing rat follicles.

Immature rats were treated with diethylstilboestrol (DES) or pregnant mares' serum gonadotropin (PMSG) and forward angle light-scatter (FALS) and 90 degrees light-scatter (90 degrees LS) signals were used to measure the size and the granularity (internal organization) of the granulosa cells, respectively. The results confirmed the presence of two major populations of granulosa cells in the ovaries of both groups of rats, with the same percentage of larger cells in both treatments (52.3% in DES, 49.5% in PMSG). Since DES treatment brings about granulosa cell growth while PMSG treatment causes growth and differentiation, it is evident that there is heterogeneity in granulosa cell sizes during different states of growth and differentiation. There was also heterogeneity in sizes of granulosa cells harvested from follicles of small (less than 210 microns), medium (210-420 microns) and large (greater than 420 microns) diameter. Quadrant analysis of granulosa cells in various fractions collected from Percoll gradients suggested an increase in granularity in the small and large granulosa cell populations. Cell cycle analysis of small and large granulosa cell populations collected from large follicles of rats treated with PMSG indicated that each population was distributed in G0/G1, S and G2/M phases. These results demonstrate that populations of small and large granulosa cells exist in rat ovarian follicles during various stages of growth and differentiation.

Recurrent renal cell carcinoma arising in Wilms' tumor.

A 19-month-old black girl had a radical nephrectomy for a Wilms' tumor that contained areas of epithelium indistinguishable from renal cell carcinoma. She was treated with chemotherapy but subsequently had pulmonary metastases develop and massive abdominal recurrence. The recurrent tumor was histologically renal cell carcinoma with no identifiable Wilms' tumor elements. The child died with recurrent and metastatic tumor 13 months after nephrectomy. Pathologic, immunoperoxidase, and flow cytometric studies of this unusual case are presented.

Intermediate structures in nuclear morphogenesis following metaphase from HeLaS3 cells can be isolated and temporally grouped.

Previously nuclear reformation following metaphase in HeLaS3 cells was conceptualized in terms of a stepwise process which was continuous throughout anaphase and telophase. This concept was based on a three-dimensional visualization by scanning electron microscopy (SEM) of individual, organically prepared chromatid structures (prenuclei) which could be sequentially arranged. Morphologic analysis revealed unique topographies and morphometric properties which suggested that it should be possible to isolate populations of prenuclei aqueously. Such an isolation using detergents and density centrifugation is presented which yields metaphase plates and two populations of prenuclei with distinctive morphology. Essentially, prenuclei are freed from late mitotic cells in suspension cultures of synchronized HeLaS3 cells by treatment with 0.1% Nonidet-P40 followed by treatment with a mixture of Tween 40-desoxycholate (0.5%). Critical for the isolation is the presence of a divalent cation (5 mM Mg(+)+) and an acid pH (approximately 5.8). After density centrifugation, 2N decondensing structures (late intermediates) are recovered from 42% Percoll, and a mixture of 2N predecondensing (early intermediates) and 4N metaphase plates are recovered from 52% Percoll. The latter intermediates can be further separated into highly enriched populations (greater than 94% pure) by fluorescence-activated sorting. Predecondensing structures are of the same overall morphology as prenuclei isolated previously by organic means, can also be ordered sequentially to demonstrate nuclear morphogenesis, and retain centromere/kinetochore loci. These chromosomal loci based on immunostaining of individual structures appear to be positioned centrally during chromatid reassociation and then appear to be dispersed prior to structural rearrangements leading to formation of a disc-like prenucleus.(ABSTRACT TRUNCATED AT 250 WORDS)

Flow cytometric DNA analysis.

Flow cytometric DNA analysis, including principles, techniques, and applications, is reviewed. Flow cytometry, a relatively recently developed technology, is being increasingly used in the diagnosis and treatment of a wide variety of benign and malignant diseases. Current major applications include phenotypic analysis of cells, cell sorting, and DNA analysis. DNA analysis by flow cytometry is rapid, reliable, and reproducible. Flow cytometric DNA analysis offers a quick, reliable method of analyzing cellular DNA content capable of identifying cell populations with abnormalities in total DNA content.

Flow cytometry shows recurrent aneuploid clone after remission in megakaryocytic leukemia.

Flow cytometry using a DNA label can quantitate aneuploid clones in malignant tissue. We illustrated the clinical value of this technique in a 71-year-old woman with acute megakaryocytic leukemia, which was diagnosed by staining of the blasts with factor VIII antigen and their morphologic resemblance to megakaryoblasts. Marrow cells were removed from needle biopsies by vortexing in RPMI medium, centrifuged in Ficoll-Hypaque, stained with a propidium-iodide/NP-40 mixture, and analyzed at 488 nm using an argon laser. During 3 weeks of low-dose cytosine arabinoside (ara-c) infusion therapy, hyperdiploid peak A dropped from 35% (day 0) to 2.3% (day 14) to 0% (day 21), with development of marrow hypoplasia. Similarly, hyperdiploid peak B, went from 7.6% to 9.1% to 3.5%. Subsequently, her marrow recovered normal morphology and lost the aneuploid peaks. Her blood counts recovered to near normal. Four months later, she relapsed and had a return of the day-21, incompletely eradicated peak B. There was no evidence of peak A. Repeated treatment with ara-c resulted in temporary suppression of the disease, but she died 3 months later with progressive hepatosplenomegaly. Analysis of cells from her enlarged liver, heavily infiltrated with blasts, showed a large hyperdiploid peak B. In this patient, ara-c therapy induced a remission with permanent eradication of one clone, but incomplete eradication of a second clone, which ultimately led to her relapse and death.