Accuracy of Grading Gleason Score 7 ProstaticAdenocarcinoma on Needle Biopsy: Influence of Percent

Pattern 4 and Other Histological Factors

Abdelrazak Meliti,1 Evita Sadimin,1 Mario Diolombi,1 Francesca Khani,1

and Jonathan I. Epstein2*1The Departments of Pathology, The Johns Hopkins Medical Institutions, Baltimore, Maryland

2The Departments of Urology and Oncology, The Johns Hopkins Medical Institutions, Baltimore, Maryland

BACKGROUND. Recognition of Gleason pattern 4 in prostatic needle biopsies is crucial forboth prognosis and therapy. Recently, it has been recommended to record percent pattern 4when Gleason score 7 cancer is the highest grade in a case.METHODS. Four hundred and five prostate needle core biopsies received for a secondopinion at our institution from February–June, 2015 were prospectively diagnosed withprostatic adenocarcinoma Gleason score 7 as the highest score on review by a consultanturological pathologist. Percentage of core involvement by cancer, percentage of Gleasonpattern 4 per core, distribution of Gleason pattern 4 (clustered, scattered), morphology ofpattern 4 (cribriform, non-cribriform), and whether the cancer was continuous or discontinu-ous were recorded.RESULTS. Better agreement was noted between the consultant and referring pathologistswhen pattern 4 was clustered as opposed to dispersed in biopsies (P¼ 0.009). The percentageof core involvement by cancer, morphology of pattern 4, and continuity of cancer did notaffect the agreement between the consultant and referring pathologists. There was a trend(P¼ 0.06) for better agreement based on the percent of pattern 4.CONCLUSIONS. When pattern 4 is scattered amongst pattern 3 as opposed to beingdiscrete foci, there is less interobserver reproducibility in grading Gleason score 7 cancer, andin this setting pathologists should consider obtaining second opinions either internally withintheir group or externally. Prostate # 2017 Wiley Periodicals, Inc.

KEY WORDS: prostate cancer; gleason grading; percent pattern 4; interobserver

INTRODUCTION

Gleason grading is widely used, and is the offi-cially recommended histopathologic grading systemfor prostatic carcinoma in the United States [1–6].Identification of Gleason pattern 4 in prostatic needlebiopsies is fundamental both prognostically andtherapeutically. The Gleason grading system hasundergone several revisions in the last decade.Changes in the grading system that could affect thediagnosis of Gleason score 7 first occurred in 2005 andmore recently in 2014 [3,4]. How the extent of pattern4 and morphological variations of pattern 4 influenceinterobserver agreement in diagnosing Gleason 7 onneedle core biopsies in the modern era has not beenstudied.

MATERIALS AND METHODS

Four hundred and five consecutive prostate needlecore biopsies with prostatic adenocarcinoma Gleasonscore 7 as the highest grade received for a second

Conflicts of interest: The authors have no conflict of interests andno funding.�Correspondence to: Jonathan I. Epstein, MD, The Departments ofUrology and Oncology, The Johns Hopkins Medical Institutions,401N Broadway St, Rm 2242, Baltimore, MD 21231.E-mail: jepstein@jhmi.eduReceived 13 December 2016; Accepted 13 January 2017DOI 10.1002/pros.23314Published online in Wiley Online Library(wileyonlinelibrary.com).

The Prostate

� 2017 Wiley Periodicals, Inc.

opinion at our institution from February–June, 2015were studied. The agreement between an experturological pathology consultant and general submit-ting pathologists was studied. The Gleason gradesprovided by the consultant pathologist was dividedinto two groups (3þ 4¼ 7, Grade Group 2) and(4þ 3¼ 7, Grade Group 3), and then each group wascompared to general pathologists grading (under-grading, same grading, and over-grading). For eachcase, a set of the morphologic factors that couldcontribute and influence the level of concordancewhen assigning Gleason score 7 between urologicalpathologists and general pathologists were recorded.These included quantitative and qualitative factorspresent in the needle biopsy of the prostate such aspercentage of core involvement by cancer, percentageof Gleason pattern 4 per core assessed by “eye-ball”estimate in 10% increments, distribution of Gleasonpattern 4 (clustered, scattered), morphology of pattern4 (cribriform, non-cribriform), and whether the cancerwas continuous or discontinuous. Non-cribriformpattern 4 included both poorly-formed and fusedglands. Statistics were performed using STATA (Stata-Corp, College Station, TX).

RESULTS

The overall percentage of tumor per core rangedfrom 5% to 100% with an average of 47% and theoverall percentage of pattern 4 per core provided bythe consultant ranged from 5% to 90% with anaverage of 34%. Two hundred and ninety biopsies(71.6%) were assigned Gleason score 3þ 4¼ 7 (GradeGroup 2) and 115 biopsies (28.4%) were assignedGleason score 4þ 3¼ 7 (Grade Group 3) on theconsultant’s review. When the consultant assignedGleason score 3þ 4¼ 7 (Grade Group 2), there wasagreement with the referring pathologists in 187biopsies (64.5%), 79 biopsies (27.2 %) were under-graded, and 24 biopsies (8.3%) were over-gradedrelative to the consultant. When the consultantassigned Gleason score 4þ 3¼ 7 (Grade Group 3),there was agreement with the referring pathologistsin 54 biopsies (47%), 50 biopsies (43.5%) were under-graded, and 11 biopsies (9.5%) were over-gradedrelative to the consultant.

Greater agreement was noted between the consul-tant and referring pathologists when pattern 4 wasclustered (P¼ 0.009) in biopsies with Gleason scores3þ 4¼ 7 (Grade Group 2) and 4þ 3¼ 7 (GradeGroup 3) (Figs. 1–4). The percentage of core involve-ment by cancer, morphology of pattern 4, and conti-nuity of cancer did not affect the agreement betweenthe consultant and referring pathologists (Table I)(Fig. 5). One-way ANOVA test was conducted to

explore the difference between the referring patholo-gists’ grading groups in terms of overall percentage ofthe tumor and percentage of pattern 4 provided bythe consultant pathologist. The analysis was per-formed separately for samples graded as 3þ 4 (GradeGroup 2) and 4þ 3 (Grade Group 3) by the consultingpathologist. There was a trend (P¼ 0.06) for betteragreement based on the percent of pattern 4. Whenthe consultant diagnosed Gleason 3þ 4 (GradeGroup 2) and the contributors graded lower, in 80% ofthe cases there was <20% pattern 4. When theconsultant diagnosed Gleason 4þ 3 (Grade Group 3)and the contributors graded lower, 70% of the cases

Fig. 1. Grade assigned by general pathologists in cases assignedGleason score 3þ 4¼ 7 (Grade Group 2) by a consultanturopathologist.

Fig. 2. Grade assigned by general pathologists in cases assignedGleason score 4þ 3¼ 7 (Grade Group 3) by a consultanturopathologist.

2 Meliti et al.

The Prostate

had 50–70% pattern 4. Under-grading was associatedwith a higher percentage of tumor, while over-grading was associated with a smaller percentage(Fig. 6).

DISCUSSION

Gleason score 7 prostatic carcinomas are tradition-ally subdivided into 3þ 4¼ 7 (Grade Group 2) and4þ 3¼ 7 (Grade Group 3), depending whether pattern4 is predominant. Published studies have shownthat the greater the amount of Gleason pattern 4,the worse the prognosis [7–11]. Multiple studieshave examined the inter-observer reproducibility ofGleason grading of prostatic adenocarcinoma bothamongst urologic pathologists as well as betweengeneral pathologists. A consistent finding in studiesthat examined the inter-observer reproducibility ofGleason grading of prostatic carcinoma amongst

urologic pathologists is the presence of substantialinter-observer agreement for Gleason score 7 [12–14].On the other hand, other studies have demonstratedconsistent under-grading of Gleason scores withGleason score 7 being under-diagnosed by generalpathologists [15–16]. A recent study examinedthe inter-observer reproducibility among urologicpathologists, addressing the impact of diagnosing“poorly formed glands” Gleason pattern 4 prostaticadenocarcinoma [17]. The study showed fair overallreproducibility among urologic pathologists for diag-nosing “poorly formed glands” of Gleason pattern 4.McKenney et al. also has demonstrated that interob-server reproducibility for the histological distinctionof tangentially sectioned Gleason pattern 3 fromGleason pattern 4 was only fair (Light’s k 0.27) [18].

In the present study, we assessed the agreementbetween a urologic pathologist and general patholo-gists when assigning Gleason score 7 on needle corebiopsy. The most significant factor that we demon-strated to improve interobserver reproducibility wasthe presence of clustered distribution of pattern 4. Itmakes sense that it is easier to assess the percent ofpattern 4 when the pattern 4 is localized as a cluster ofeither poorly formed, fused, or cribriform glands asopposed to scattered amongst well-formed glands ofpattern 3 (Fig. 4). There was a trend for lowerpercentage of pattern 4 to result in an increasedlikelihood of undergrading. It is intuitive that cases of3þ 4 (Grade Group 2) with <20% pattern 4 wouldhave an increased potential of being undergraded asGleason 3þ 3 (Grade Group 1). Similarly, Cases of4þ 3¼ 7 (Grade Group 3) with 50–70% of pattern 4would have an increased potential of being under-graded as Gleason 3þ 4¼ 7 (Grade Group 2).

Although not included in the ISUP manuscripton the 2014 consensus conference, it was recom-mended at the meeting the percent pattern 4 berecorded for Gleason score 7 in both needle biopsyand radical prostatectomy specimens. The majoradvantage for recording the percent pattern 4 is inbiopsy for men being considered for active surveil-lance (AS). For the appropriate patient, Gleasonscore 3þ 3¼ 6 (Grade Group 1) is accepted for mento undergo active surveillance. However, there maybe some men, depending on age, co-morbidity,extent of cancer, imaging findings, patient desire,etc., who could be a candidate for active surveil-lance with Gleason score 3þ 4¼ 7 (Grade Group 2)if the pattern 4 is limited. The amount of pattern 4is not only used for active surveillance but could beused for radiation therapy as well. Currently, thereare different radiation therapy protocols for Gleasonscore 3þ 4¼ 7 (Grade Group 2) versus Gleasonscore 4þ 3¼ 7 (Grade Group 3). A case that is

Fig. 3. Gleason score 3þ 4¼ 7 (Grade Group 2) with well-formed discrete glands of Gleason pattern 3 (left) and cluster ofpoorly-formed glands of Gleason pattern 4 (right).

Fig. 4. Portion of a core with a roughly equal mix of scatteredpoorly-formed and well-formed glands.

Accuracy Grading Gleason Score 7 3

The Prostate

borderline between these two grades would beapparent if the percent pattern 4 is recorded andthen other factors (clinical stage, PSA, number ofcores positive, etc.) could be used to decide therapy.By recording percent pattern 4 it would also explaincases, as seen in the current study, where there arediscrepant diagnoses between pathologists withborderline amounts of pattern 4. This would applyto several scenarios: (i) one pathologist grades atumor as 3þ 3¼ 6 (Grade Group 1) and the other as3þ 4¼ 7 (Grade Group 2) with limited percentpattern 4; and (ii) one pathologist grades a tumoras 3þ 4¼ 7 (Grade Group 2) and the other as4þ 3¼ 7 (Grade Group 3) with 50–70% pattern 4.Recording percent pattern would also help explainwhen one pathologist grades a tumor as 4þ 3¼ 7

(Grade Group 3) and the other as 4þ 4¼ 8 (GradeGroup 4) with 80–90% pattern 4, data not studiedin the current study.

In conclusion, including percent pattern 4 inreports on prostate adenocarcinoma on needle biopsycan help explain discrepant grading between patholo-gists in borderline cases, and improve patient care.Our study also demonstrates that when pattern 4 isscattered amongst pattern 3 as opposed to beingdiscrete foci, there is less inter-observer reproducibil-ity in grading Gleason score 7 cancer. In the settingof a core with scattered poorly-formed glands sur-rounded by well-formed glands, where the corerepresents the highest grade in the case, pathologistsshould consider obtaining second opinions eitherinternally within their group or externally.

TABLE I. Influence of Percent Gleason Pattern 4 and Other Histological Factors on Interobserver ReproducibilityWhen the Consultant Assigns Gleason Score 3þ 4¼ 7 and Gleason Score 4þ 3¼ 7

Gleason 3þ 4¼ 7 assigned by the consultant(n¼ 290)

Gleason 4þ 3¼ 7 assigned by the consultant(n¼ 115)

Undergradedn¼ 79 (27%)

Agree n¼ 187(65%)

Overgradedn¼ 24 (8%)

Undergradedn¼ 50 (43%)

Agree n¼ 54(47%)

Overgradedn¼ 11 (10%)

Percent Pattern 4 Mean 16% Mean 19% Mean 21% Mean 72% Mean 75% Mean 71%Percent cancer/

coreMean 56% Mean 45% Mean 44% Mean 58% Mean 48% Mean 27%

Cribriformpattern 4

9 (11%) 32 (17%) 1 (4%) 6 (12%) 7 (13%) 1 (9%)

Non-cribriformpattern 4

70 (89%) 155 (83%) 23 (96%) 44 (88%) 47 (87%) 10 (91%)

Discontinouscancer

26 (33%) 48 (26%) 5 (21%) 17 (34%) 18 (33%) 2 (18%)

Continuouscancer

53 (67%) 139 (74%) 19 (79%) 33 (66%) 36 (67%) 9 (82%)

Fig. 5. Gleason score 3þ 4¼ 7 (Grade Group 2) with well-formed discrete glands of Gleason pattern 3 (top) and cribriformgland of Gleason pattern 4 (bottom left).

Fig. 6. Agreement in grading between general pathologists andconsultant uropathologist with Gleason score 7 cancer relative toextent of cancer.

4 Meliti et al.

The Prostate

REFERENCES

1. Gleason DF, Mellinger GT, Veterans Administration Coopera-tive Urological Research Group. Prediction of prognosis forprostatic adenocarcinoma by combined histological gradingand clinical staging. J Urol 1974;167:953–958.

2. Gleason DF. Histologic grading of prostate cancer: A perspec-tive. Hum Pathol 1992;23:273–279.

3. Epstein JI, Allsbrook WCJ, Amin MB, Egevad LL. The 2005International Society of Urological Pathology (ISUP) consensusconference on Gleason grading of prostatic carcinoma. Am JSurg Pathol 2005;29:1228–1242.

4. Epstein JI, Egevad L, Amin MB, Delahunt B, Srigley JR,Humphrey PSA; Grading Committee. The 2014 InternationalSociety of Urological Pathology (ISUP) consensus conferenceon Gleason grading of prostatic carcinoma: Definition ofgrading patterns and proposal for a new grading system. Am JSurg Pathol 2016;40:244–252.

5. Humphrey PA. Gleason grading and prognostic factors incarcinoma of the prostate. Mod Pathol 2004;17:292–306.

6. Epstein JI, Allsbrook WC, Amin MB, Egevad LL. Update on theGleason grading system for prostate cancer: Results of aninternational consensus conference of urologic pathologists.Adv Anat Pathol 2006;13:57–59.

7. Stark JR, Perner S, Stampfer MJ, Sinnott JA, Finn S, EisensteinAS, Ma J, Fiorentino M, Kurth T, Loda M, Giovannucci EL,Rubin MA, Mucci LA. Gleason score and lethal prostate cancer:Does 3þ4¼4þ3? J Clin Oncol 2009;27:3459–3464.

8. Makarov DV, Sanderson H, Partin AW, Epstein JI. Gleason score7 prostate cancer on needle biopsy: Is the prognostic differencein Gleason scores 4 þ 3 and 3 þ 4 independent of the number ofinvolved cores? J Urol 2002;167:2440–2442.

9. Amin A, Partin A, Epstein JI. Gleason score 7 prostate cancer onneedle biopsy: Relation of primary pattern 3 or 4 to pathologicalstage and progression after radical prostatectomy. J Urol 2011;186:1286–1290.

10. Reese AC, Cowan JE, Brajtbord JS, Harris CR, Carroll PR,Cooperberg MR. The quantitative Gleason score improvesprostate cancer risk assessment. Cancer 2012;118:6046–6054.

11. Berg KD, Røder MA, Brasso K, Vainer B, Iversen P. PrimaryGleason pattern in biopsy Gleason score 7 is predictive of

adverse histopathological features and biochemical failurefollowing radical prostatectomy. Scand J Urol 2014;48:168–176.

12. Allsbrook WC, Jr., Mangold KA, Johnson MH, Lane RB,Amin MD, Bostwick DG, Humphrey PA, Jones EC, ReuterVE, Sakr W, Sesterhenn IA, Troncoso P, Wheeler TM, EpsteinJI. Interobserver reproducibility of Gleason grading of pros-tatic carcinoma: Urologic pathologists. Hum Pathol 2011;32:74–80.

13. Burchardt M, Engers R, Muller M, Burchardt T, Willers R,Epstein JI, Ackermann R, Gabbert HE, de la Taille A, RubinMA. Interobserver reproducibility of Gleason grading: Evalua-tion using prostate cancer tissue microarrays. J Cancer Res ClinOncol 2008;134:1071–1078.

14. Melia J, Moseley R, Ball RY, Griffiths DF, Grigor K, Harnden P,Jarmulowicz M, McWilliam LJ, Montironi R, Waller M, Moss S,Parkinson MC. A UK-based investigation of inter- and intra-observer reproducibility of Gleason grading of prostatic biop-sies. Histopathology 2006;48:644–654.

15. Allsbrook WC, Jr., Mangold KA, Johnson MH, Lane RB, LaneCG, Epstein JI. Interobserver reproducibility of Gleason gradingof prostatic carcinoma: General pathologists. Hum Pathol 2001;32:81–88.

16. Nakai Y, Tanaka N, Shimada K, Konishi N, Miyake M, Anai S,Fujimoto K. Review by urological pathologists improves theaccuracy of Gleason grading by general pathologists. BMCUrology 2015;15:70–77.

17. Zhou M, Li J, Cheng L, Egevad L, Deng FM, Kunju LP,Magi-Galluzzi C, Melamed J, Mehra R, Mendrino S, Osun-koya AO, Paner G, Shen SS, Tsuzuki T, Trpkov K, Tian W,Tang X, Shah RB. Diagnosis of ‘Poorly Formed Glands’Gleason pattern 4 prostatic adenocarcinoma on needle biopsyan interobserver reproducibility study among urologic path-ologists with recommendations. Am J Surg Pathol 2015;39:1331–1339.

18. McKenney JK, Simko J, Bonham M, True LD, Troyer D, HawleyS, Newcomb LF, Fazli L, Kunju LP, Nicolas MM, Yakar-Lopez F,Zhang X, Carroll PR, Brooks JD, Canary/Early DetectionResearch Network Prosate Active Surveillance Study Investiga-tors. The potential impact of reproducibility of Gleason gradingin men with early stage prostate cancer managed by activesurveillance: A multi-institutional study. J Urol 2011;186:465–469.

Accuracy Grading Gleason Score 7 5

The Prostate