Problems in grading of prostatic carcinoma: interobserver reproducibility of five different grading systems

Subspecialty surgical pathologist's performances as triage pathologists on a telepathology-enabled quality assurance surgical pathology service: A human factors study

BACKGROUND

The case triage practice workflow model was used to manage incoming cases on a telepathology-enabled surgical pathology quality assurance (QA) service. Maximizing efficiency of workflow and the use of pathologist time requires detailed information on factors that influence telepathologists' decision-making on a surgical pathology QA service, which was gathered and analyzed in this study.

MATERIALS AND METHODS

Surgical pathology report reviews and telepathology service logs were audited, for 1862 consecutive telepathology QA cases accrued from a single Arizona rural hospital over a 51 month period. Ten university faculty telepathologists served as the case readers. Each telepathologist had an area of subspecialty surgical pathology expertise (i.e. gastrointestinal pathology, dermatopathology, etc.) but functioned largely as a general surgical pathologist while on this telepathology-enabled QA service. They handled all incoming cases during their individual 1-h telepathology sessions, regardless of the nature of the organ systems represented in the real-time incoming stream of outside surgical pathology cases.

RESULTS

The 10 participating telepathologists' postAmerican Board of pathology examination experience ranged from 3 to 36 years. This is a surrogate for age. About 91% of incoming cases were immediately signed out regardless of the subspecialty surgical pathologists' area of surgical pathology expertise. One hundred and seventy cases (9.13%) were deferred. Case concurrence rates with the provisional surgical pathology diagnosis of the referring pathologist, for incoming cases, averaged 94.3%, but ranged from 88.46% to 100% for individual telepathologists. Telepathology case deferral rates, for second opinions or immunohistochemistry, ranged from 4.79% to 21.26%. Differences in concordance rates and deferral rates among telepathologists, for incoming cases, were significant but did not correlate with years of experience as a practicing pathologist. Coincidental overlaps of the area of subspecialty surgical pathology expertise with organ-related incoming cases did not influence decisions by the telepathologists to either defer those cases or to agree or disagree with the referring pathologist's provisional diagnoses.

CONCLUSIONS

Subspecialty surgical pathologists effectively served as general surgical pathologists on a telepathology-based surgical pathology QA service. Concurrence rates with incoming surgical pathology report diagnoses, and case deferral rates, varied significantly among the 10 on-service telepathologists. We found no evidence that the higher deferral rates correlated with improving the accuracy or quality of the surgical pathology reports.

[Interobserver reproducibility of Gleason grading in prostate biopsy samples]

INTRODUCTION

Gleason grading is the most common method of prostate cancer classification.

AIM

The aim of the authors was to assess the reproducibility of Gleason grading among pathologists using the same needle biopsy samples.

METHOD

23 pathologists examined 37 prostate cancer biopsies stained with hematoxylin and eosin. Gleason scores were categorised into 4 groups (2-4, 5-6, 7 and 8-10). Kappa statistics were used to reflect interobserver agreement.

RESULTS

Considering all participating pathologists, grouping into one of the 4 categories resulted in an overall kappa value of 0.49. For the individual categories, the worst agreement (kappa = 0.15) was seen with well differentiated carcinomas, and the best (kappa = 0.65) with poorly differentiated ones.

CONCLUSIONS

These results suggest that Gleason grading in biopsy samples is moderately reproducible. The kappa values vary according to the differentiation of the cancer, and there is not much difference between the results of the present study and those published in the literature. To increase reproducibility, trainings should be organised, and this could improve the quality of grading.

Gleason grading: past, present and future

In 1966 Donald Gleason developed his grading and scoring system for prostatic adenocarcinoma. This classification was refined in 1974 and gained almost universal acceptance, being classified as a category 1 prognostic parameter by the College of American Pathologists. Modifications to the classification were recommended at a conference convened by the International Society of Urological Pathology (ISUP) in 2005. This modified classification has resulted in a significant upgrading of tumours, although some studies have shown a greater concordance between needle biopsy and radical prostatectomy scores when compared to classical Gleason (CG) grading. The ISUP consensus conference recommended that for needle biopsies higher tertiary patterns should be incorporated into the final Gleason score, and this has been correlated with biochemical failure, tumour volume and mortality. Recently the validity of including cribriform glands as a component of Gleason pattern 3 has been questioned and it has been recommended that all tumours showing cribriform architecture should be classified as Gleason pattern 4. The recommendations arising from the 2005 Consensus Conference were largely unsupported by validating data, yet this new grading system has achieved widespread usage. It is unfortunate that recent suggestions for further modification are similarly lacking in supporting evidence. In view of this it is recommended that the Modified Gleason Scoring Classification should continue to be utilized in its original (2005) format and that any future alterations should be implemented only when mandated by tumour-related outcome studies.

Reproducibility and reliability of tumor grading in urological neoplasms

Histopathologic tumor grading reflects the degree of differentiation of a given tumor and for most urological tumors grading is an important factor in predicting their biological aggressiveness. Consequently, the clinical management of tumor patients is often strongly influenced by the tumor grade, provided by pathologists. This implicates that an ideal grading system should not only be of high prognostic relevance, but also of high reproducibility among different pathologists. To this end individual histological grading systems have been developed for different tumor entities and even for a given tumor type several grading systems have been proposed. All of these grading systems possess an inherent degree of subjectivity and consequently, both intra- and interobserver variability exist. In this review, grading systems for the most frequent urological tumors (i.e. prostate cancer, renal cell carcinoma, and urothelial tumors) are mentioned and data on the reproducibility and reliability of the most commonly used grading systems are summarized.

A UK-based investigation of inter- and intra-observer reproducibility of Gleason grading of prostatic biopsies

AIMS

The frequency of prostatic core biopsies to detect cancer has been increasing with more widespread prostate specific antigen (PSA) testing. Gleason score has important implications for patient management but morphological reproducibility data for British practice are limited. Using literature-based criteria nine uropathologists took part in a reproducibility study.

METHODS

Each of the nine participants submitted slides from consecutive cases of biopsy-diagnosed cancer assigned to the Gleason score groups 2-4, 5-6, 7 and 8-10 in the original report. A random selection of slides was taken within each group and examined by all pathologists, who were blind to the original score. Over six circulations, new slides were mixed with previously read slides, resulting in a total of 47 of 81 slides being read more than once.

RESULTS

For the first readings of the 81 slides, the agreement with the consensus score was 78% and overall interobserver agreement was kappa 0.54 for Gleason score groups 2-4, 5-6, 7, 8-10. Kappa values for each category were 0.33, 0.56, 0.44 and 0.68, respectively. For the 47 slides read more than once, intra-observer agreement was 77%, kappa 0.66. The study identified problems in core biopsy interpretation of Gleason score at levels 2-4 and 7. Patterns illustrated by Gleason as 2 tended to be categorized as 3 because of the variable acinar size and unassessable lesional margin. In slides with consensus Gleason score 7, 13% of readings were scored 6 and in slides with consensus 6, 18% of readings were scored 7.

CONCLUSIONS

Recommendations include the need to increase objectivity of the Gleason criteria but limits of descriptive morphology may have to be accepted.

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.

Prostate cancer histologic trends in the metropolitan Detroit area, 1982 to 1996

OBJECTIVES

One of the concerns regarding the widespread use of serum prostate-specific antigen (PSA) as a screening tool for prostate cancer is the possibility that it may detect latent or clinically insignificant cancers. One indicator of clinical importance is thought to be histologic grade, with clinically unimportant cancers more likely to be well differentiated and clinically important tumors more likely to be moderately or poorly differentiated.

METHODS

Data from the metropolitan Detroit population-based Surveillance, Epidemiology, and End Results Program were examined to determine trends in prostate cancer histologic grading before and after the introduction of PSA screening.

RESULTS

From 1989 through 1996, the most recent year for which data are available, a dramatic increase in the incidence of prostate cancer occurred in the Detroit area, corresponding to the routine use of PSA as a screening test for prostate cancer. Local stage cancer demonstrated the largest increase in incidence. The incidence of moderately differentiated cancers also rose substantially during the same period; the incidence of poorly differentiated tumors remained about the same, and the incidence of well differentiated tumors decreased. Coincident with the increasing proportion of moderately differentiated cancers was a significant increase in the proportion of prostate biopsies performed (P = 0.001).

CONCLUSIONS

These population-based data add important evidence that prostate cancers identified with PSA are more likely to be moderately than well differentiated. Additionally, if the definition of clinical significance depends on histopathologic grade, this finding could further be interpreted as evidence that PSA is more likely to detect clinically significant prostate cancer.

Interobserver variation in prostate cancer Gleason scoring: are there implications for the design of clinical trials and treatment strategies?

A series of prostate cancer histological slides from 71 patients were used to measure the interobserver variation among three pathologists awarding a Gleason score. The study was prompted on account of the use of histological grade to stratify patients prior to randomization within two clinical trials currently recruiting at our centre, and a proposed study that would allocate treatment depending upon the score awarded. The pathologists were expected to award a score based upon their day to day experience, there being no consensus meeting before-hand to agree on the grey areas of the Gleason grading system. We used the kappa statistic to assess the level of agreement. This was calculated both for comparison of the raw scores awarded by the three observers, as well as the grouped scores corresponding to those groupings used for the purposes of stratification in the two trials. The extent of the interobserver variation (weighted kappa) for the raw scores (Gleason scores 2-10) was 0.16 to 0.29 and for the grouped scores (Gleason scores < or = 7 or > or = 8), kappa was 0.15 to 0.29. For the raw scores, the total agreement rate was 9.9% and the total disagreement 26.8%; for the grouped scores the total agreement rate was 43.7%. It is concluded that, despite this level of agreement there is no concern regarding stratification using the Gleason score, because of the subsequent randomization. However, using a reported Gleason score to determine treatment might be inappropriate. These data indicate the value of a central review process for pathology grading in clinical trials, especially where the treatment is directly affected by this information.

Diagnostic decision support for prostate lesions

The diagnostic evaluation of premalignant and malignant lesions of the prostate may benefit from the application of an inference network. Used as a diagnostic decision support system, an inference network provides standardized assessment of diagnostic clues which is supported by computer graphics and comparison imagery, uncertainty management by possibility and probabilistic schemes and the systematic combination of different pieces of diagnostic evidence. This assessment results in a numeric measure of belief in the final diagnosis.

The prognostic importance of Gleason grade in prostatic adenocarcinoma: a long-term follow-up study of 648 patients treated with radiation therapy

PURPOSE

It is common practice to histologically grade adenocarcinoma of the prostate using the Gleason system. Whereas the prognostic utility of this grading is well known, few studies have comprehensively evaluated it for patients undergoing definitive radiation therapy and generally accepted guidelines as to which Gleason grades should be "lumped" have not been established. This study reports the results of univariate and multivariate evaluation of the prognostic significance of Gleason grade in 648 patients followed for a median of 6.5 years after radiation therapy for T1 to T4, N0, or NX, MO prostate cancer.

METHODS AND MATERIALS

The correlation between Gleason grade and local recurrence, metastatic relapse, any disease relapse, and patient survival was evaluated using univariate and multivariate methods. Analysis was also stratified according to whether the grading was assigned on a needle biopsy or on a transurethral resection specimen.

RESULTS

The large number of Gleason grades required grouping of grades for meaningful analysis and we found that a four-tier system (grades 2 and 3, 155 patients; grades 4-6, 290 patients; grade 7, 92 patients; and grades 8-10, 111 patients) correlated best with outcome. In univariate analysis, this four-tier grouping correlated significantly with local recurrence, distant metastases, any relapse, and survival. The incidences of distant metastasis at 10 years were: grades 2 and 3, 13%; grades 4-6, 34%; grade 7, 52%; and, grades 8-10, 63%. The survival rates at 10 years were: grades 2 and 3, 64%; grades 4-6, 60%; grade 7, 46%; and grades 8-10, 24%. In multivariate analysis, Gleason grade was the single most important determinant of outcome for each endpoint. These results applied equally to needle biopsy and transurethral resection specimens.

CONCLUSION

Tumor grade is the single most significant determinant of outcome following radiotherapy for clinically localized prostate cancer. The Gleason system is a valid method for grading tumors to be irradiated. A four-tier grouping into grades 2 and 3, grades 4-6, grade 7, and grades 8-10 appears to be adequate and simple.

Do neuroendocrine cells in human prostate cancer express androgen receptor?

The presence of androgen receptors (AR) in neuroendocrine cells was investigated in benign tissue of 10 prostatectomy specimens, in 12 prostatic adenocarcinomas with focal neuroendocrine differentiation and in 1 case of a pure neuroendocrine small cell carcinoma of the prostate. Neuroendocrine cells were defined by their reactivity with an antibody to chromogranin A. Monoclonal antibody F39.4 directed against the amino-terminal domain of the AR molecule was used to detect AR. AR and chromogranin A were simultaneously visualized with a double immunofluorescence technique. The results indicate that chromogranin positive cells in both benign and malignant prostatic tissue lack detectable expression of AR. No effect of endocrine therapy was noted. These results are in agreement with the hypothesis that prostatic neuroendocrine tumour cells represent an androgen insensitive cell population, which incidentally may expand to replace the androgen-sensitive tumour cell population during androgen ablation therapy.

Application of the monoclonal antibody Ki-67 on prostate biopsies to assess the fraction of human prostatic carcinoma

The feasibility of using the monoclonal antibody Ki-67 as a proliferation marker in human prostatic carcinoma was studied on aspiration and core biopsy specimens obtained from 50 patients suspected of having prostate cancer. In 32 prostatic adenocarcinomas the Ki-67 index varied from 0.3 to 13.3% (mean 4.3) in cytological smears and from 0.8 to 17.8% (mean 5.1) in frozen sections from histological core biopsies. No significant correlation between the percentage of cells positive for Ki-67 and the histological tumor differentiation could be established. In 18 patients with benign prostatic hyperplasia the Ki-67 index varied from 0 to 3.0% (mean 1.2) and from 0 to 3.8% (mean 1.4) in cytological and histological material, respectively. The differences in the observed Ki-67 index between benign and malignant prostatic tissues are of statistical (p less than 0.001) and of clinical significance. Nine patients who underwent endocrine treatment or radiotherapy entered a followup protocol in which the Ki-67 staining procedure was applied to periodically obtained cytological aspiration biopsies. During month 1 after the start of therapy a statistically significant (p less than 0.05) decrease in the Ki-67 index to 58% of the initial values was found, while at 2 and 3 months the proliferative fraction showed a further decrease to 27 and 7%, respectively. As a marker, the monoclonal antibody Ki-67 was shown to provide a reliable method to estimate the proliferative cell fraction of human prostate cancer.

Variation of prostate-specific antigen expression in different tumour growth patterns present in prostatectomy specimens

A series of 55 randomly chosen radical prostatectomy specimens was analyzed for expression of prostate-specific antigen (PSA) by immunohistochemical techniques. Tissue sections were selected in such a manner that in addition to glandular benign prostatic hyperplasia (BPH), one or more different prostatic tumour growth patterns were present. Four monoclonal antibodies, directed against three different PSA epitopes, and one polyclonal anti-PSA antiserum were used. Expression of PSA was compared with that of prostate-specific acid phosphatase (PAP), recognized by two different polyclonal antisera. A critical dilution aimed at a maximum of staining intensity on BPH tissue sections was chosen for all antibodies. Anti-PSA and anti-PAP antisera stained essentially all BPH samples (over 90%). Irrespective of the nature of the antibodies used, PSA expression was found to be decreased in prostatic carcinoma. A clear cut relationship was found between immunoreactivity for PSA and the degree of differentiation of the tumour area. Under the experimental conditions used the PSA monoclonal antibodies stained only 1 out of 10 undifferentiated carcinomas, whereas 50% to 70% of the well- and moderately-differentiated carcinomas showed immunoreactivity. This correlation was less pronounced with the PAP staining pattern. If the PSA antibody titer was raised the percentage of clearly staining undifferentiated carcinomas could be considerably increased (up to 60%-100%), indicating that PSA expression is not absent, but lowered in most (if not all) undifferentiated carcinomas.

Histological grading of prostatic carcinoma in prostatectomy specimens. Comparison of prognostic accuracy of five grading systems

The prognostic accuracy of 5 histological grading systems (Broders, Anderson, Mostofi, Gleason and Mostofi-Schroeder) was compared. Grading was performed on 50 prostatectomy specimens by 5 pathologists. The results were averaged so as to reduce the impact of inter-observer variation. The Cox proportional hazards model was used to estimate the relationship between average grading scores and both time-to-recurrence and time-to-death by prostatic carcinoma. Age at surgery was considered to be a possible confounding factor and adjusted accordingly. The prognostic impact of the 5 grading systems (related to both recurrence and death caused by prostatic carcinoma) was judged by the likelihood ratio (LR) test score (chi 2 distributed with 1 df); for time-to-recurrence for the Mostofi-Schroeder score the LR was 6.54 and for the Gleason system it was 1.79. A stepwise procedure demonstrated that the best prognostic performance was reached with the Mostofi-Schroeder and Broders systems used together (with Mostofi-Schroeder weighted 1.5 times larger than Broders). For time-to-recurrence the median grading result was also used, giving results similar to the mean grading result. For time-to-death from prostatic carcinoma the LR test scores for all grading systems were relatively low. In this analysis the outcome of the Gleason system showed a minimum of prognostic ability, whereas the Broders and Mostofi-Schroeder systems had a reasonable predictive ability. Since the inter-observer variation of the Mostofi-Schroeder system was large, the Broders system is preferable. The restrictions and implications of this study are discussed and a brief review of the prognostic importance of grading of prostatic carcinoma is presented.

Monoclonal antibody Ki-67 defined growth fraction in benign prostatic hyperplasia and prostatic cancer

Growth fractions were assessed immunohistochemically in prostatic tissues with benign glandular hyperplasia (BPH) and in specimens of prostatic cancer using the monoclonal antibody Ki-67. This antibody is specific for a proliferation-associated nuclear antigen. In BPH tissues about 0.3% of nuclei of epithelial cells was reactive with Ki-67. The Ki-67 positive nuclei were distributed equally among the basal and luminal cells of the hyperplastic prostatic acini. In prostatic cancer the Ki-67 defined growth fraction ranged from 0.4% to 9.1% (mean value 2.9%). Cancers with a cribriform growth pattern and tumors composed of solid areas of undifferentiated cancer cells showed the highest growth fraction (average values 4.0%, respectively 7.6%). The investigated four tumors composed of undifferentiated solitary tumor cells with diffuse infiltration of the stroma demonstrated an unexpectedly low growth activity (average 1.2%). In cancers with a glandular growth pattern the Ki-67 defined growth fraction of tumor cells varied from 2.2% to 5%. Compared with other epithelial tumors these values are low, but they are in agreement with the earlier findings on prostatic cancer obtained with 3H-thymidine labeling and bromodeoxyuridine incorporation. The observed variation in the level of Ki-67 defined growth activity partly related to the histological tumor pattern suggests that Ki-67 labeling may serve as a prognostic factor additional to the current histopathological grading criteria of prostatic cancer.

Pathologic aspects of prostate cancer: prediction of malignant potential

Despite our expanding knowledge of prostatic histopathology, several controversies regarding the prediction of a tumor's biologic behavior still exist. Numerous histologic grading schemes specific for prostate carcinomas have been developed, but pathologists and urologists do not agree on which system is most accurate. Unfortunately, the Broders system, which is frequently used, tends to underestimate tumor grade and malignant potential. Recently, computer-assisted nuclear morphometry has added objectivity to the formerly subjective grading process, and improved prognostic accuracy for early and advanced disease. The relationship between tumor volume and grade remains controversial. Although it has been stated that volume and grade are directly related, our examination of radical prostatectomy specimens casts some doubt on this. While it can be shown that tumor volume is proportional to the surface area of capsular penetration by malignant cells, tumor penetration is more closely related to histologic grade.