INTEROBSERVER AGREEMENT OF GLEASON SCORE AND MODIFIED GLEASON SCORE IN NEEDLE BIOPSY AND IN SURGICAL SPECIMEN OF PROSTATE CANCER
( Download pdf )

SERGIO G. VELOSO, MARIO F. LIMA, PAULO G. SALLES, CYNTHIA K. BERENSTEIN, JOAO D. SCALON, EDUARDO A. BAMBIRRA

Section of Urology, Mario Penna Hospital, and Department of Pathology, School of Medicine, Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil

ABSTRACT

     Introduction: Gleason score, which has a high interobserver variability, is used to classify prostate cancer. The most recent consensus valued the tertiary Gleason pattern and recommended its use in the final score of needle biopsies (modified Gleason score). This pattern is considered to be of high prognostic value in surgical specimens. This study emphasized the evaluation of the modified score agreement in needle biopsies and in surgical specimen, as well as the interobserver variability of this score.
     Materials and Methods: Three pathologists evaluated the slides of needle biopsies and surgical specimens of 110 patients, reporting primary, secondary and tertiary Gleason patterns and after that, traditional and modified Gleason scores were calculated. Kappa test (K) assessed the interobserver agreement and the agreement between the traditional and modified scores of the biopsy and of the surgical specimen.
     Results: Interobserver agreement in the biopsy was K = 0.36 and K = 0.35, and in the surgical specimen it was K = 0.46 and K = 0.36, for the traditional and modified scores, respectively. The tertiary Gleason grade was found in 8%, 0% and 2% of the biopsies and in 8%, 0% and 13% of the surgical specimens, according to observers 1, 2 and 3, respectively. When evaluating the agreement of the traditional and modified Gleason scores in needle biopsy with both scores of the surgical specimen, a similar agreement was found through Kappa.
     Conclusion: Contrary to what was expected, the modified Gleason score was not superior in the agreement between the biopsy score and the specimen, or in interobserver reproducibility, in this study.

Key words: prostatic neoplasms; biopsy, needle; surgery; pathology
Int Braz J Urol. 2007; 33: 639-51

INTRODUCTION

     Prostate cancer tends to be morphologically heterogeneous (1), showing several patterns of differentiation, classified by Gleason system (2). Prostate needle biopsy provides random samples, which might not represent neoplasia in all its heterogeneity, generally downgrading the tumor (3-5). By clinical accompaniment, a worse prognosis was found in the patients who had small proportions of Gleason patterns 4 and 5 tumors, which are not mentioned in the Gleason score (6-10). From this observation, the concept of modified Gleason score was created incorporating these small most aggressive patterns in the patient’s score and being used in some prognostic nomograms (11,12), Figure-1.



     Several studies deal with interobserver agreement of Gleason score, with all sorts of different results (13). Other studies deal with the agreement as regards modified Gleason score in slides (14). Recently, Helpap reported better association between needle biopsy and surgical specimen using the modified Gleason score (11). Gleason histopathological classification shows high level of subjectivity. Despite its undeniable clinical importance, as a diagnostic method, the Gleason score, more precisely the modified score, needs to be evaluated in relation to its reliability. Taking this into account, we tried to evaluate the interobserver agreement and the association between needle biopsy and the surgical specimen adopting Gleason and modified Gleason scores.

MATERIALS AND METHODS

     A hundred and ten patients suffering from prostate cancer without any previous treatment and who would be referred to a radical prostatectomy agreed to participate in the research. They signed the consent term and sent their needle sextant biopsies, coming from different laboratories, to be reevaluated. Those biopsies had about two cores per sextant, mean total of 12 cores (range 6 to 24 cores). The surgical specimen was processed in the same laboratory, by partial sampling, producing about nine slides per surgical specimen (range 7 to 20), evaluating apex, distal third, mid third, proximal third, bladder neck, right and left seminal vesicles. Thus, the surgical specimen was not processed as a whole. All material was stained with hematoxylin-eosin. All the available slides of the needle biopsy and of the surgical specimens, with or without cancer, were evaluated by the observers.
     Three pathologists belonging to different services of Pathological Anatomy examined the slides of the needle and surgical samples of these patients. They did not know the clinical data nor did they know about the pairing between needle biopsy and surgical specimen. They filled in a protocol in which they should classify the primary, secondary and the most aggressive Gleason patterns of each examined area of both specimens. At the end of this task, Gleason score was calculated (the sum of primary and secondary patterns) of each sextant separately (12). The score of the specimen was the highest score found among the evaluated sextants, therefore, the global score was not calculated (4,5,7,15). From the most aggressive Gleason pattern, the tertiary pattern was determined, whenever it was possible. The modified Gleason score was calculated (the sum of primary and tertiary patterns) (12). Similarly, the highest modified score of the examined slides was adopted as the modified Gleason score of the specimen. The primary Gleason pattern was defined as the most frequent Gleason pattern of the sample. The secondary Gleason pattern was the second most frequent pattern, obligatory higher than 5% of the tumor area (12). When the secondary pattern was less than 5%, the primary pattern was repeated. The tertiary Gleason pattern corresponded to the third Gleason pattern, necessarily more aggressive than the secondary pattern (12). In order to avoid terminology confusion, the Gleason score was called traditional Gleason score, being clearly differentiated from the modified Gleason score.
     Data were collected in a data bank and statistically evaluated by Stata program version 9.1 (StatCorp. 4905 Lakeway Dr, College Station, USA). Kappa (K) and weighted Kappa test were used to evaluate the interobserver agreement and the agreement between the Gleason score of the two specimens. The interpretation of the agreement by Kappa value was done by the intervals: K < 0, poor; K = 0-0.2, slight; K = 0.2-0.4, fair; K = 0.4-0.6, moderate; K = 0.6-0.8, substantial; and K = 0.8-1.0, almost perfect (13). In the statistic inferences, in general, the level of significance of 5% was adopted and, consequently, a confidence level of 95% was used.

RESULTS

     The samples’ mean age was 63.5 +/- 7.7 years old (range 44 to 79 years old). The mean preoperative PSA was 10.2 +/- 8.2 ng/mL (range 1.2 to 53.4 ng/mL). The clinical tumor staging (digital rectal examination) was 46.7% of T1, 47.5% of T2 and 5.8% of T3. In the initial anatomic pathological test, extracapsular tumor extension was found in 17% (pT3a) and in seminal vesicles invasion in 11% (pT3b).
     The three pathologists are specialized in the same university even though they nowadays work in different hospitals and laboratories. The experience of working in surgical pathology and the weekly amount of prostate tests, criteria adopted by Taille (13), allow us to classify the observers 1 and 2 as experienced and the observer 3 as less experienced.
     Some slides, considered unsatisfactory, were rejected from the research.
     In the biopsies, there was a predominance of Gleason pattern 3 in the primary pattern, range from 66% to 86%, and of secondary pattern, range from 63% to 71% among the observers. Similarly, in the surgical specimens, Gleason pattern 3 was more frequent in the primary pattern, being found from 75% to 81%, and as secondary pattern from 60% to 69% of the observations. There was an absolute predominance of Gleason grade 3 in the primary and secondary grade in both specimens. Gleason grade 6 was also predominant in the needle biopsy. In the surgical specimen there was a similar proportion of Gleason score 6 and 7. Table-1 shows the distribution of traditional and modified Gleason scores.



     Interobserver agreement in needle biopsy as regards to primary Gleason grade was reasonable to moderate, according to Kappa. In the surgical specimen, the agreement was moderate to substantial. In the secondary Gleason pattern there was a divergence among the observers, the agreement was generally low, occasionally reasonable. As for the most aggressive Gleason pattern it was from reasonable to moderate (Table-2).



     Interobserver agreement of traditional Gleason score in the needle samples was reasonable, with exact agreement among 60% to 68% and agreement +/- 1 Gleason score from 91% to 98%. In the specimens the agreement was from reasonable to moderate, with exact diagnosis from 66% to 71% and accepting difference of one unit from 96% to 99%. Modified Gleason grade presented similar agreement in both specimens, being reasonable to moderate. Exact diagnosis in the biopsy was from 58% to 69% and accepting agreement +/- 1 Gleason score from 86% to 97%. In the specimen the exact diagnosis was from 60% to 64%, accepting divergence of one unit changing from 94% to 95%. By adopting weighted Kappa, values similar to Kappa (not weighted) were found (Table-3).



     Tertiary Gleason pattern was diagnosed in 8%, 0% and 2% of the biopsies and in 8%, 0% and 13% of the surgical specimen according to observers 1, 2 and 3, respectively. Thus, traditional and modified Gleason scores, according to observer 1, were the same in 92% of both specimens. Observer 2 did not consider any pattern as tertiary, having 100% precision between the two Gleason scores. Examiner 3 had 98% of the needle biopsies and 87% of the surgical specimens with the same diagnosis between the two scores.
     Traditional and modified Gleason scores were used to evaluate the association among their scores in both specimens by each observer. For observer 1, adopting the traditional score in needle biopsy and in surgical specimen K = 0.24 was found. Adopting the modified score in the biopsy and the traditional one in the specimen, we got K = 0.21. The same happened when using the modified score in the needle biopsy and in the surgical specimen. Examiner 2 did not find any difference in the association of scores between specimens (K = 0.26). When examiner 3 used the traditional score in the needle biopsy and in the surgical specimen, the value for Kappa was 0.18 and when using the modified score in the biopsy and the traditional one in the specimen, Kappa was 0.17. Adopting the traditional Gleason score in both specimens, lower downgrading in needle biopsy was found than by adopting the modified score in both samples (Table-4).

COMMENTS

     The sample used reflects a group of patients referred to radical prostatectomy, in other words, young patients, with localized illness and generally low Gleason score. The three observers, also young, had similar academic and professional background and learned the Gleason system during medical residence in the same institution. Therefore, a good agreement among them would be expected.
     Higher agreement of primary Gleason pattern was found in the surgical specimen and not in the needle biopsy. By observing smaller areas, it is expected that more attention would be devoted to a specific area and higher agreement would happen. On the other hand, once the specimen is better represented in tissue extension, the suspected areas with borderline pattern were better examined, resulting in higher agreement. This reflects the difficulties in diagnosing secondary pattern, which besides involving the identification of Gleason patterns, demands tumor volume determination. As a rule, secondary Gleason pattern is the one that is more than 5% of the tumor area and with smaller extension than the primary pattern. Determining the tumor extension is not necessary for the diagnosis of the most aggressive Gleason pattern, the recognition of the worst pattern is sufficient. Glaessgen found a weak agreement as regards the diagnosis of the most aggressive patterns and considered that the difficulty in diagnosing them was bigger than in determining their volume (14). The experience did not influence the agreement much because it was not higher between the more experienced observers, what contradicts some authors (15,16).
     Interobserver agreement of traditional Gleason score was slightly higher in the surgical specimen than in the needle biopsy. By adopting the modified Gleason score, the agreement was similar in needle biopsy and surgical specimen. In general, adopting weighted Kappa, the agreement values were a little higher, but without altering the previous relations. It is interesting to notice that the modified Gleason score did not show any superiority over traditional score, as Glaessgen reported (14). Evaluating the agreement in relation to the patterns, it is higher in the primary pattern and in the most aggressive one (this is intimately related to tertiary pattern) and too low in the secondary pattern. The modified score would be expected to obtain a higher agreement, but this did not happen. This fact might have happened due to the small number of tertiary pattern diagnosed and, as a result, the two scores were similar. However, this number is similar to the one found in Griffiths‘ study, where the diagnostic proportion of tertiary Gleason pattern was 6% for general pathologists and 9% for uropathologists, showing weak agreement in relation to tertiary pattern (17). This pattern, in general, refers to patterns 4 or 5, which can present borderline structures making the diagnosis more difficult (3). Generally, the studies regarding the use of tertiary pattern use it in the prognostic evaluation, in surgical specimens. (8-10) Mosse, when evaluating the prognosis of patients with tertiary pattern 5, found a worse prognosis in those with Gleason score 6 or 7 in the surgical specimens. (8) It is known that, statistically, those scores are the most frequent ones.
     Considering that prostate cancer is heterogeneous and multicentric (1), it is assumed that the biopsy, which samples a small portion of it, might not represent it efficiently (3-5). Traditionally it is believed that Gleason score in needle biopsy tends to downgrade the surgical specimen, because a less differentiated pattern may not have been sampled in the biopsy (4,12). That was observed by observers 1 and 2. Taking the downgrading concept as a starting-point, some authors suggest the use of modified Gleason score, which would better reflect the real tumor characteristics for it values the most aggressive small patterns (6). The International Society of Urological Pathology (ISUP) on Gleason grading recommends the inclusion of tertiary pattern (modified Gleason score) in needle biopsies. (12) In the surgical specimens, however, it is still recommended to mention the tertiary pattern, whenever it is present, without including it in the score (Gleason score). (12) Considering that the needle biopsy downgrades the score, it was expected that the modified Gleason score would have a better agreement with the traditional Gleason score in the surgical specimen. However this fact could not be demonstrated. The modified score in needle biopsy compared to traditional and modified score in the specimen, presented the same Kappa values or even slightly inferior ones when adopting the traditional score in the biopsy. The best representation in the biopsy was not proved when adopting the modified Gleason score. This fact, as previously mentioned, might have happened due to the low diagnosis of tertiary pattern. Helpap, on the contrary, evaluating slides of 368 patients, found improvement of the exact agreement between the two specimens using the modified Gleason score instead of the traditional score, ranging from 58% to 78% (8). However, he did not use the Kappa test to evaluate the real agreement, nor reported the diagnostic proportion of tertiary pattern.

CONCLUSIONS

     In this study, the modified Gleason score did not prove to be superior in reproducibility compared to the traditional Gleason score, both in the needle biopsy and in the surgical specimen. Contrary to what was expected, the use of the modified score in the biopsy was not superior to the traditional score, comparing to the Gleason scores of the specimen. Within the aim of the study, the modified Gleason score was not superior to the traditional one. These conclusions might be due to the methodology used, as well as to the observers involved. Isolated morphological analysis is based in criteria of low reproducibility. It is necessary to reevaluate the association between the two Gleason scores, using different samples with a higher amount of tertiary pattern.

ACKNOWLEDGMENT

     Dr. Sergio G. Veloso has a CNPq Grant, Ministry of Technology, Brazil.

CONFLICT OF INTEREST

     None declared.

REFERENCES

1. Aihara M, Wheeler TM, Ohori M, Scardino PT: Heterogeneity of prostate cancer in radical prostatectomy specimens. Urology. 1994; 43: 60-6; discussion 66-7.
2. Billis A, Pompeo AC: Adenocarcinoma da próstata. Int Braz J Urol. 2003; 29 (suppl 1): 27-34.
3. Montironi R, Mazzuccheli R, Scarpelli M, Lopez-Beltran A, Fellegara G, Algaba F: Gleason grading of prostate cancer in needle biopsies or radical prostatectomy specimens: contemporary approach, current clinical significance and sources of pathology discrepancies. BJU Int. 2005; 95: 1146-52.
4. Lopez-Beltran A, Mikuz G, Luque RJ, Mazzucchelli R, Montironi R: Current practice of Gleason grading of prostate carcinoma. Virchows Arch. 2006; 448: 111-8.
5. Egevad L, Allsbrook WC Jr, Epstein JI: Current practice of Gleason grading among genitourinary pathologists. Hum Pathol. 2005; 36: 5-9.
6. Pan CC, Potter SR, Partin AW, Epstein JI: The prognostic significance of tertiary Gleason patterns of higher grade in radical prostatectomy specimens: a proposal to modify the Gleason grading system. Am J Surg Pathol. 2000; 24: 563-9.
7. Rioux-Leclercq NC, Chan DY, Epstein JI: Prediction of outcome after radical prostatectomy in men with organ-confined Gleason score 8 to 10 adenocarcinoma. Urology. 2002; 60: 666-9.
8. Helpap B, Egevad L: The significance of modified Gleason grading of prostatic carcinoma in biopsy and radical prostatectomy specimens. Virchows Arch. 2006; 449: 622-7.
9. Epstein JI, Allsbrook WC Jr, Amin MB, Egevad LL; ISUP Grading Committee: The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma. Am J Surg Pathol. 2005; 29: 1228-42.
10. De la Taille A, Viellefond A, Berger N, Boucher E, De Fromont M, Fondimare A, et al.: Evaluation of the interobserver reproducibility of Gleason grading of prostatic adenocarcinoma using tissue microarrays. Hum Pathol. 2003; 34: 444-9.
11. Glaessgen A, Hamberg H, Pihl CG, Sundelin B, Nilsson B, Egevad L: Interobserver reproducibility of modified Gleason score in radical prostatectomy specimens. Virchows Arch. 2004; 445: 17-21.
12. Hollenbeck BK, Bassily N, Wei JT, Montie JE, Hayasaka S, Taylor JM, et al.: Whole mounted radical prostatectomy specimens do not increase detection of adverse pathological features. J Urol. 2000; 164: 1583-6.
13. Kunz GM Jr, Epstein JI: Should each core with prostate cancer be assigned a separate gleason score? Hum Pathol. 2003; 34: 911-4.
14. Glaessgen A, Hamberg H, Pihl CG, Sundelin B, Nilsson B, Egevad L: Interobserver reproducibility of percent Gleason grade 4/5 in total prostatectomy specimens. J Urol. 2002; 168: 2006-10.
15. Allsbrook WC Jr, Mangold KA, Johnson MH, Lane RB, Lane CG, Epstein JI: Interobserver reproducibility of Gleason grading of prostatic carcinoma: general pathologist. Hum Pathol. 2001; 32: 81-8. Erratum in: Hum Pathol 2001; 32: 1417.
16. Allsbrook WC Jr, Mangold KA, Johnson MH, Lane RB, Lane CG, Amin MB, et al.: Interobserver reproducibility of Gleason grading of prostatic carcinoma: urologic pathologists. Hum Pathol. 2001; 32: 74-80.
17. Griffiths DF, Melia J, McWilliam LJ, Ball RY, Grigor K, Harnden P, et al.: A study of Gleason score interpretation in different groups of UK pathologists; techniques for improving reproducibility. Histopathology. 2006; 48: 655-62.

____________________
Accepted after revision:
August 8, 2007

_______________________
Correspondence address:
Dr. Sergio Geraldo Veloso
Rua Henrique Benfenatti, 237
São João del-Rei, MG, 36307-042, Brazil
Fax: + 55 32 3371-8003
E-mail: velososg@ig.com.br

1. Epstein JI, Allsbrook WC Jr., Amin MB, Egevad L: The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason grading of prostatic carcinoma. Am J Surg Pathol. 2005; 29: 1228-42.
2. Helpap B, Egevad L: Correlation of modified Gleason grading with pT stage of prostatic carcinoma after radical prostectomy. Anal Quant Cytol Histol (in press).
3. Helpap B, Egevad L: Correlation of modified Gleason grading of prostate carcinoma with serum PSA, age and tumor extent in needle biopsy specimens. Anal Quant Cytol Histol (in press).
4. Helpap B, Egevad L: The significance of modified Gleason grading of prostatic carcinoma in biopsy and radical prostatectomy specimens. Virchows Arch. 2006; 449: 622-7.
5. Glaessgen A, Hamberg H, Pihl CG, Sundelin B, Nilsson B, Egevad L: Interobserver reproducibility of modified Gleason score in radical prostatectomy specimens. Virchows Arch. 2004; 445: 17-21.
6. Glaessgen A, Hamberg H, Pihl CG, Sundelin B, Nilsson B, Egevad L: Interobserver reproducibility of percent Gleason grade 4/5 in prostate biopsies. J Urol. 2004; 171: 664-7.
7. Albertsen PC, Hanley JA, Barrows GH, Penson DF, Kowalczyk PD, Sanders MM, et al.: Prostate cancer and the Will Rogers phenomenon. J Natl Cancer Inst. 2005; 97: 1248-53.

1. Epstein JI, Allsbrook WC Jr., Amin MB, Egevad LL, ISUP Grading Committee: The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma. Am J Surg Pathol. 2005; 29: 1228-42.
2. Garnett JE, Oyasu R, Grayhack JT: The accuracy of diagnostic biopsy specimens in predicting tumor grades by Gleason’s classification of radical prostatectomy specimens. J Urol. 1984; 131: 690-3.
3. Mills SE, Fowler JE Jr: Gleason histologic grading of prostatic carcinoma. Correlation between biopsy and prostatectomy specimens. Cancer. 1986; 57: 346-9.
4. Humphrey PA: Prostate Pathology. Chicago, ASCP Press. 2003; p. 138.
5. Algaba F, Chivite A, Santaularia JM, Oliver A: Evidence of the radical prostatectomy Gleason score in the biopsy Gleason score. Actas Urol. Esp. 2004; 28: 21-6.
6. Rubin MA, Dunn R, Kambham PA, Misick CP, O’Toole KM: Should a Gleason score be assigned to a minute focus of carcinoma on prostate biopsy? Am J Surg Pathol. 2000; 24: 1634-40.
7. Steinberg DM, Sauvageot J, Piantadosi S, Epstein JI: Correlation of prostate needle biopsy and radical prostatectomy Gleason grade in academic and community setting. Am J Surg Pathol. 1997; 21: 566-76.
8. Cintra ML, Billis A: Histologic grading of prostatic adenocarcinoma. Intra-observer reproducibility of the Mostofi, Gleason, and Böcking grading systems. Int Urol Nephrol. 1991; 23: 449-54.
9. Özdamar SO, Sarikaya S, Yildiz L, Atilla MK, Kandemir B, Yildiz S: Intra-observer and interobserver reproducibility of WHO and Gleason histologic grading systems in prostatic adenocarcinomas. Int Urol Nephrol. 1996; 28: 73-7.
10. Egevad L, Allsbrook WC, Epstein JI: Current practice of Gleason grading among genitourinary pathologists. Hum Pathol. 2005; 36: 5-9.