Abstract
Background: Biopsy Gleason score (bGS) is an important tool for staging and decision making in patients with prostate cancer. Therefore, the data from biopsy should be both reproducible across different pathologists and predictive of the true underlying tumour. We evaluated the agreement between bGS with prostatectomy Gleason score (pGS) comparing patients who underwent prostate biopsy at our hospital with those who did it at an outside facility. Materials and Methods: We retrospectively analyzed patients who underwent robot-assisted radical prostatectomy at our Hospital in 2011 and 2012. Patients were divided depending on the site of prostate biopsy. We calculated a weighted κ statistic to evaluate the concordance from bGS and pGS in the two groups and to evaluate the Gleason score (GS) concordance comparing the proportion of positive cores at biopsy. Results: A total of 124 patients with completed data were identified (70 patients performed biopsy at our institution and 54 at an outside facility). The weighted κ score for GS agreement was 0.40 for our Institution and 0.27 for other facilities. The weighted κ score stratified by biopsy hospital for patients with at least 30% of positive cores was 0.46 for our hospital and 0.42 for other facilities. Conclusion: Internal prostate biopsy predicted better pGS than outside facility biopsy reports. When the percentage of biopsy-positive cores increases, the agreement between bGS and pGS is similar between the two groups. For certain cases in which an outside laboratory biopsy results in equivocal clinical decision, biopsy re-evaluation by internal pathologists can help reveal the true underlying tumor architecture and extension.
Gleason score (GS), consisting of primary and secondary grades, is the predominant grading system for prostate cancer (PCa). This system, based on the glandular architecture, defines five histological patterns or grades with decreasing differentiation (1).
Biopsy Gleason score (bGS) is an important tool for staging and decision-making in patients with PCa. In fact, bGS represents a powerful predictor of tumor biology and treatment outcome (2) together with presenting serum prostatic specific antigen (PSA), digital rectal examination and volume of cancer in the biopsy (3-4). GS of 7-10 is associated with worse prognoses, while tumours with GS of 5-6 are associated with lower progression rates after definitive therapy (1).
GS is especially important for patients considering a management alternative to radical prostatectomy (RP), such as active surveillance (AS), brachytherapy, cryotherapy or external beam radiotherapy (5). In these cases, the only tissue available is the one from the biopsy; therefore, any discrepancy between bGS and the actual cancer grading can result in under- or overtreatment (6).
Despite this fact, current evidence showed poor agreement between bGS and RP specimen Gleason score (pGS); particularly, an upgrade in pGS has been demonstrated in 50% of patients with low bGS, while a downgrade has been shown in up to 80% of patients with high bGS (7).
In order to allow for accurate decision-making, the data from biopsy should be both reproducible across different pathologists and predictive of the true underlying tumor.
We, thus, compared the agreement between bGS and pGS in patients who had biopsy in our institution, which is a tertiary referral centre, or in a referring institution, in order to investigate any significant difference in the rate of agreement.
Materials and Methods
We retrospectively analyzed the records of patients who underwent robot-assisted radical prostatectomy (RARP) at our hospital in 2011 and 2012. We evaluated the agreement between bGS and pGS, comparing patients who underwent prostate biopsy in our Institution (group 1) with those who had it in a referring institution (group 2). We also evaluated the correlation between agreement of GS and proportion of positive biopsy cores (cut-off value: 30%). Biopsy Gleason score were grouped as ≤6, 7 (subgrouping in 3+4 and 4+3) and ≥8. We calculated a weighted κ coefficient (that assigns different weights to subjects for whom the raters differ by categories) to evaluate the concordance between bGS and pGS in the two groups and to evaluate the GS concordance comparing the proportion of positive cores at biopsy. A value of the κ coefficient, adjusted for agreement expected by chance, of 0 indicates no agreement, <40% marginal agreement, 40-75% good agreement and >75% excellent agreement.
Results
A total of 124 patients were identified in our dataset. Seventy patients underwent the biopsy at our Institution (group 1), while 54 had it in a referring Institution (group 2). Patients' characteristics are reported in Table I. The mean age was 66 years in both cohorts; mean PSA was 7.2 in patients who underwent the biopsy at our Hospital and 10 in patients coming from other facilities; mean number of cores was 15.6 in group 1 and 14 in group 2.
The weighted κ score for GS agreement was 40% for our institution and 27% for referring institution (Table II).
The weighted κ score, stratified by biopsy hospital for patients with at least 30% of positive cores, was 46% for our hospital and 42% for referring institutions (Table III).
Discussion
Our results showed a quite good overall agreement between bGS and pGS in our institution, while only marginal agreement was reported when the biopsy was performed in referring institutions. A good agreement has been reported in both cases when at least 30% of the biopsy cores were positive for malignancy; the increment of the proportion of positive cores increased the concordance between the two groups. Differences in the two groups might be possibly explained by differential expertise of pathologists, considering that some referring institutions are less likely to have specialized uropathologists (8). It also must be noted that, in our institution, the same team of pathologists read both the prostate biopsy and the final surgical specimen, possibly contributing to increase concordance (9). A proportion of positive cores >30% increased the concordance between the two groups. One explanation might be that smaller cancer volumes may lead to sampling error and subsequent upgrading (10).
Several studies have investigated the correlation between the bGS and pGS, demonstrating discrepancy and especially upgrading from biopsy to prostatectomy specimens (7). This fact is particularly significant in clinical practice, since an upgrade to a higher GS is a strong and independent predictor of biochemical recurrence after attempted local curative therapy (10).
Many factors can explain the disagreement between bGS and pGS.
PCa is a multifocal disease, which is poorly imaged by ultrasound (11). Thus, transrectal ultrasound (TRUS) guided biopsies of 6-12 cores represent a small, random sampling of the overall prostatic tissue (approximately 0.04% of the average volume of the gland) (1). Therefore, it is very plausible to miss a focus of PCa with a higher GS (12).
Epstein et al. reported some of the more common pathology errors in grading needle biopsy specimens: overcalling Gleason pattern 4 on tangentially sectioned small glands of pattern 3 that mimic poorly formed glands; undercalling cribriform Gleason pattern 4 as pattern 3; undercalling GS 9-10. Furthermore, there are borderline grades between patterns 3-4 and 4-5 (6).
A role may also be played by the presence of a tertiary GS pattern (containing less than 5% of the tissue), which is usually uncommon and was not included in the standard GS (13). When the worst Gleason grade is the tertiary pattern, it should influence the final bGS (1).
Downgrading may be due to the processing of the prostatic tissue or the subjective nature of the GS (7). On the other hand, more biopsy cores with cancer, higher PSA levels and obesity seem to be predictors of upgrading (14). Obese men have larger prostates, which may contribute to poorer sampling (15).
In a nationwide study in Norway, Kvåle et al. showed that the concordance between bGS and pGS was significantly higher in pathology departments that examined >40 RP specimens annually than in departments assessing fewer specimens (16).
The Gleason grading system was revised in 2005 at the consensus conference of the International Society of Urological Pathology (ISUP) (17). Among the consensus statements, some are particularly significant: bGS 2-4 should rarely, if ever, be diagnosed; small amounts (≤5%) of biopsy Gleason pattern 3 should be ignored in the presence of Gleason pattern 4 or 5; in the case of patterns 3, 4 and 5, the GS should be the summation of the most common and the highest-grade pattern. These changes may reduce the difference between high- and low-volume pathology units (16).
Nevertheless, GS upgrading is clinically significant. A bGS <6 is often used as one of the inclusion criteria for AS (18) but cancers with a bGS of 6, that are upgraded to a pGS of 7, have the same pathological stage, margin status and biochemical failure rate as do cancers with bGS and pGS of 7 (16).
In men choosing radiotherapy, the presence of a GS 6 or 7 might affect the type and/or location of the radiotherapy and the possible addition of hormonal therapy. Furthermore, men with upgraded disease were more likely to have adverse pathologic features and biochemical progression (14).
Main limitations of our study were the small sample size, which makes it difficult to generalize the results, and its retrospective nature.
Patients' characteristics.
Percentage of agreement, κ score and weighted κ score stratified by biopsy hospital.
Percentage of agreement, weighted κ score stratified by group for patients with at least 30% of positive cores.
Conclusion
The pathological GS was better-predicted when prostate biopsy was performed in a referral center. When the percentage of biopsy-positive cores increases (>30%), the agreement between bGS and pGS is similar between the two groups.
In cases in which biopsy results obtained from a referring Center can lead to doubts about the clinical management, specimen re-evaluation by referral Center uropathologists can help revealing the underlying tumour architecture and extension.
- Received June 11, 2014.
- Revision received July 15, 2014.
- Accepted July 16, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.