CONCLUSIONS. The Epstein Criteria has a suboptimal accuracy for predicting for insignificant prostate cancer.

Transcription

1 ; 2011 EPSTEIN CRITERIA FOR INSIGNIFICANT PROSTATE CANCER OON ET AL. BJUI Epstein criteria for insignificant prostate cancer Sheng F. Oon*, R. William Watson*, John J. O Leary and John M. Fitzpatrick *Conway Institute, University College Dublin, Institute of Molecular Medicine, Trinity College Dublin, Mater Misericordae University Hospital, Dublin, Ireland Accepted for publication 31 August 2010 Study Type Prognosis (systematic review) Level of Evidence 2b OBJECTIVE To review the accuracy of the Epstein Criteria for insignificant prostate cancer and to explore the effect of the modified Gleason classification system on this system. METHODS We searched PubMed, EMBASE and the Cochrane Database using search terms Epstein Criteria, Prostate Cancer, Validation and Insignificant Cancer between 1994 to 2010 for validation articles. These were divided into pre-2005 and post-2005 and concordances for organconfined status, Gleason score 6 and insignificant cancer were analysed. What s known on the subject? and What does the study add? Overtreatment of prostate cancer is a major problem in contemporary urological practice. The Epstein Criteria reduces overtreatment by identifying insignificant prostate cancers that may be amenable to surveillance therapy. This systematic review of the Epstein Criteria validation studies provides a collective insight into the application and accuracy of the Epstein Criteria to predict for insignificant prostate cancer across different institutions and geographies. RESULTS A pre-2005 study showed concordance for insignificant prostate cancer, Gleason score 6 and organ-confined status at 84%, 90.3% and 91.6%, respectively. Five post-2005 validation studies were concordant for insignificant cancer, Gleason score 6 and organ-confined status at 37 76%, % and %, respectively. CONCLUSIONS The Epstein Criteria has a suboptimal accuracy for predicting for insignificant prostate cancer. The modification to Gleason scoring may be responsible for a reduced accuracy over time. However, significant heterogeneity in the validation studies means better quality validation studies are required. KEYWORDS prostate cancer, insignificant prostate cancer, epstein criteria, gleason score INTRODUCTION Prostate cancer is now the most common solid organ neoplasm in the western world and one of the leading causes of cancer death, with some countries considering screening for this disease [1,2]. Mortality rates for prostate cancer declined with PSA testing. However, PSA screening is a highly controversial topic. This controversy arises as a large number of patients with prostate cancer have clinically insignificant prostate cancer. Early autopsy studies revealed prostate cancer in 60 70% of older men dying from unrelated causes [3]. Conservative treatment of low-grade prostate cancers in this group also incurred no loss of life expectancy [4]. With an estimated 17% lifetime risk of prostate cancer diagnosis in the United States [5], it is clear that many prostate cancers pose no threat to life or health. PSA testing, however, has resulted in over diagnosis and overtreatment in prostate cancer, with the majority of prostate cancers likely to be clinically insignificant. In the Prostate cancer Intervention Versus Observation Trial, new cases of prostate cancer were estimated in 2007 with only deaths [6]. In the European Randomized study of Screening for Prostate Cancer trial, it was determined that 43% of the study population undergoing RP had minimal cancer and an estimated 1410 men needed to be screened with 48 prostatectomies performed to prevent a single death [7,8]. To reduce overtreatment and over diagnosis, watchful waiting and active surveillance measures were introduced. Watchful waiting refers to the symptomguided treatment of patients with prostate cancer where active treatment of cancer is completely deferred until the signs and symptoms of prostate cancer manifest themselves. Active surveillance (AS) refers to the therapeutic strategy of closely monitoring a patient with low-risk prostate cancer with regular follow-up of PSA, TRUS biopsy and DRE. When a progression of disease is detected by any of these monitoring measures, then the patient is referred for curative therapy. The Epstein Criteria, developed in 1994, is the most widely used tool for determining , doi: /j x x

2 EPSTEIN CRITERIA FOR INSIGNIFICANT PROSTATE CANCER TABLE 1 The Epstein Criteria as originally defined by Epstein, and in the Epstein Criteria validation studies Author PSA density (ng/ml) Gleason score Number of cores positive for tumour Volume of core positive for tumour Epstein [9] <0.1 <7 <3 50% or less Yes <7 1 <3 mm Yes Bastian [21] <0.15 <7 <3 50% or less Yes Jeldres [22] <7 1 <3 mm Yes Lee MC [26] <0.15 <7 <3 <50% Yes Lee SE [25] <7 <3 <50% Yes Hekal [24] <0.15 <7 <3 50% or less Yes Louie-Johnsun [23] <0.15 <7 <3 <50% Yes Clinically organ confined? FIG. 1. Comparison of the (a) original Gleason and (b) International Society of Urological Pathology 2005 modified Gleason grading systems for pattern 3 and pattern 4 carcinoma. Reproduced from Lotan TL, Epstein JI. Clinical implications of changing definitions within the Gleason grading system. Nat Rev Urol 2010; 7: , with permission. pattern 3 tumours to pattern 4. Pattern 3 was classically characterized by the presence of numerous small or large well-formed glands or cribriform patterns. In the modified Gleason system, most cribriform patterns were moved into pattern 4 (see Fig. 1) [14,15]. A comparison of the conventional and ISUPmodified Gleason systems is summarized in Table 2. prostate cancer significance. Epstein s original study was based on 157 men with clinical stage T1c prostate cancer compared with 64 similarly treated clinical T1a patients initially diagnosed via TURP [9]. In all, 16% of patients were deemed to harbour an insignificant tumour, and insignificant prostate cancer was defined as tumour volume <0.2 cm 3, organconfined and no Gleason patterns 4 or 5. The rationale for this volume was based on a previous study of 21 clinical T2 cases, where these tumour volumes exhibited no capsular penetration or evidence of biochemical relapse after 5 years [10,11]. Epstein also provided a minimal tumour category for tumours between cm 3 to account for disaccording opinions amongst clinicians over what tumour volumes should define insignificant tumour volume. These patients also showed no biochemical progression after RP but three out of 23 had extracapsular extension. Stamey [12] established the newly accepted tumour volume limit of 0.5 cm 3. The optimum preoperative criteria determined from Epstein s original study are summarized in Table 1. Epstein s criteria provided a positive predictive value of 95%, negative predictive value of 66% and overall predictive accuracy of 73% for insignificant cancer. The same criteria predicted 73% of minimal cancers. The Gleason scoring system was introduced in 1966 for the histological assessment of prostate cancer aggressiveness and is expressed as a formula: x + y = z. In biopsy specimens, the primary Gleason pattern (x) represents the most prevalent Gleason pattern observed in the biopsy. At the International Society of Urological Pathology (ISUP) Consensus Conference, the definition of the secondary Gleason pattern (y) was changed from the second most prevalent pattern to the highest cancer grade observable in the specimen [13]. The addition of the two Gleason patterns provides the overall Gleason score (z). The ISUP modification saw the upgrade of certain pathological features of Gleason Gleason 7 tumours are divided into two categories: (7a) and (7b). This is based on the evidence that prognostic differences exist between the two tumour subtypes (see Table 3). After the ISUP modification, the most frequently detected score was changed from Gleason 6 to 7, in particular [16 18]. The effect of the ISUP revision on the current application of the Gleason system and the need to specify in pathology reports and clinical trials which version of the Gleason system is used is already well recognized [19]. Instead of a formal re-classification, the ISUP version was intended to provide options in some areas for how to grade prostate cancers in RP specimens and biopsies. Many pathologists do not consider the ISUP recommendation newly derived, but a reflection of what the majority of urological pathologists were already doing in contemporary practice [20]. Others, who are against the consensus, claim that the ISUP version has not been properly validated and use a different style of grading. The inconsistency and variation in cancer grading methods for prostate cancer among pathologists is a limiting factor in determining cancer aggressiveness and significance, and uniformity of grading in prostate cancer specimens are required. The purpose of the present review was not to perform a meta-analysis of the data, but to

3 OON ET AL. TABLE 2 Some differences between the conventional and 2005 ISUP Gleason scoring methods (adapted from Uemura H, Hoshino K, Sasaki T, Miyoshi Y, Ishiguro H, Inayama Y, et al. Usefulness of the 2005 International Society of Urologic Pathology Gleason grading system in prostate biopsy and radical prostatectomy specimens, BJU Int 2009; ) Conventional Gleason scoring A diagnosis of Gleason score <4 is possible on biopsy Gleason pattern 3 includes a partial cribriform pattern, large cribriform Biopsy and prostatectomy specimens use the same Gleason scoring system Small quantities of high-grade tumour (<5%) on biopsy should be excluded Biopsy tumours are graded by listing the primary (most prevalent grade) and secondary (second most prevalent) patterns The Gleason score of prostatectomy specimens is assigned based on the primary and secondary patterns only In needle biopsies, the cores may either be assessed separately or assigned an overall score In prostatectomy specimens, the Gleason score of the largest volume of cancer seen is assigned to whole specimen, even if smaller volumes of higher Gleason score are seen 2005 ISUP Gleason scoring A diagnosis of Gleason Score <4 on biopsy is rarely if ever made. The majority of cribriform patterns are Gleason pattern 4. Only specimens with rare cribriform lesions are considered pattern 3 Biopsies and prostatectomy specimens use different Gleason scoring systems High-grade tumours of any quantity on biopsy should be included Biopsy tumours are graded by listing the primary (most prevalent grade) and secondary (highest grade seen) patterns The Gleason score of prostatectomy specimens are assigned based on the primary and secondary patterns with a comment on the tertiary pattern if seen Each core should be scored separately Where separate grades of tumour nodules exist in a prostatectomy specimen, the dominant nodules are assigned their own Gleason scores ISUP, International Society of Urological Pathology. investigate the application of the Epstein Criteria in contemporary practice given the changes in Gleason scoring. METHODS A literature search was conducted using PubMed, EMBASE and the Cochrane Library from 1994 to The search terms used were Prostate Cancer, Epstein Criteria, Validation and Insignificant Cancer. We restricted our searches to the English language and included only studies conducted with the specific purpose of validating the Epstein Criteria. Articles using similar criteria to the Epstein Criteria but which were conducted with the aim of developing predictive tables instead were excluded. We included all articles that included the use of PSA density, biopsy Gleason score, number of biopsy cores positive and percentage of core involvement. One reviewer conducted the literature search and appraised the retrieved articles. RESULTS Of 62 articles retrieved, six articles were relevant to the subject of Epstein Criteria validation. One article was published before 2005 and five after All five post-2005 articles were conducted spanning the time during which the ISUP version was introduced and no article indicated which version of Gleason scoring was used. We assumed that the articles published after 2005 had converted to the modified version of the Gleason system, and that those articles published before 2005 used the conventional Gleason system only. Studies limited by low population numbers were also included in the present review. No article complied completely with Epstein s original Criteria but all were included for analysis. The demographic details of these studies are summarized in Table 4. The parameters used to define the Epstein Criteria in these studies are summarized in Table 1. The concordance rates of these studies for organ-confined disease, Gleason score, insignificant prostate cancer and other pathological outcomes are summarized in Table 5. Bastian s study was the only validation study of the Epstein Criteria published before the ISUP concensus update [21]. Bastian s definition of Epstein s Criteria using PSA density at a limit of 0.15 ng/ml was already a subtle departure from the Epstein s original criterion of 0.1 ng/ml. Despite this, Bastian s PSA density range was ng/ml, which suggests that some patients who exceeded the limit for qualification into insignificant disease were included in the study. This may have been because of PSA density in the study being calculated post-operatively through the division of the absolute PSA by the prostate specimen weight minus the seminal vesicles, as opposed to the preoperative estimation of prostate volume on TRUS. Nevertheless, Bastian achieved a 91.6% concordance with organ-confined status and 90.3% concordance with a Gleason score 6 at surgery. In all, 21 patients were upgraded to Gleason 7 and two to Gleason 8. The overall accuracy of the Epstein Criteria to predict for insignificant prostate cancer was 84%. Jeldres single academic European institution study of 366 men took place between 1996 and 2006 [22]. All biopsy and RP specimens were assessed by a single uropathologist. The study obtained a 91.7% concordance with organ-confined disease and 76% concordance with Gleason patients were upgraded to Gleason 7, 82 of which were (93.2%) and 6 were (6.8%). Of all Gleason 7 patients, 34.1% had non-organ confined disease, whereas all Gleason 6 patients had organ-confined disease. The accuracy of the Epstein Criteria to predict for insignificant cancer was 76%. Louie-Johnsun et al. [23] studied the outcomes of patients in the United Kingdom undergoing RP for low-risk disease. Of 549 patients who qualified for low-risk disease

4 EPSTEIN CRITERIA FOR INSIGNIFICANT PROSTATE CANCER TABLE 3 Comparison of prognostic and pathological outcomes in Gleason 7a and 7b disease Chan [36] Lau [37] Khoddami [38] Han [39] Tollefson [40] Rasiah [41] Kang [42] Burdick [43] Study pt1 (%) pt2 (%) pt3 (%) pt2-t3 (%) ECE (%) SVI (%) PSM (%) PNI (%) LNI (%) Metastases within 5 years (%) year actuarial risk of biochemical progression (%) year actuarial risk of biochemical progression (%) year actuarial risk of biochemical progression (%) year cancer-specific survival (%) year risk of systemic recurrence (%) 8 15 ECE, extracapsular extension; SVI, seminal vesicle invasion; PSM, positive surgical margins; PNI, perineural invasion; LNI, lymph node invasion. according to the D Amico risk classification criteria, 74 were also deemed to have insignificant cancer as determined by the Epstein Criteria. Sixty-one per cent of these were upgraded to significant disease at surgery (insignificant cancer was defined as total tumour volume <0.5 ml and no Gleason pattern 4 or higher). Sixteen per cent were upgraded to Gleason 7 and 3% had extracapsular extension. The authors do not account for the remainder of upgrades to significant cancer and these may have been because of significant tumour volumes uncovered at surgery (the mean tumour volume was 2.5 ml) or higher upgrades to Gleason It was not stated how many patients comprised of and tumours but a bar graph provided by the authors comparing individual specimen Gleason scores demonstrated a clear predominance of tumours over tumours at surgery. Hekal et al. [24] examined a Middle Eastern cohort of 35 patients who underwent RP for T1c prostate cancer (2 via laparoscopic prostatectomy). Although the authors do not specifically define the criteria used in their methodology, they imply the use of Bastian s criteria in their discussion. They obtained an 80% concordance with organ-confined disease and 54.3% concordance with Gleason % were Gleason 7 (20% 3 + 4, 11.4% 4 + 3) and 14.3% were Gleason In all, 11.4% of patients had lymph node positive disease and in the 8-year retrospective study, 8.6% of patients developed metastases. Lee (SE) [25] studied 131 Korean men with Epstein-Criteria determined insignificant prostate cancer. At RP, 40 men had Gleason 7 disease [37 had disease (92.5%), three had (7.5%)], with no Gleason 8 10 tumours. All four patients with extracapsular extension in the study had Gleason 7 disease (3 had disease, 1 had 4 + 3). All Gleason 6 patients had organ-confined disease. The predictive accuracy of the Epstein Criteria for insignificant cancer was 69.5%. Lee (MC) [26] validated the Epstein Criteria in a cohort of 268 men who underwent RP. The authors compared a cohort of patients that fulfilled the Epstein Criteria (n = 136), against a cohort of patients with Gleason 6 prostate cancer on biopsy but did not fulfil the Epstein Criteria for significant disease (n = 132). Outcomes compared were insignificant cancer, organ-confined status and

5 OON ET AL. TABLE 4 Population demographics for the Epstein sriteria validation studies Study Bastian [21] Jeldres [22] Lee M [26] Lee S [25] Hekal [24] Louie-Johnsun [23] Study period No. of patients Mean age Mean preoperative PSA (ng/ml) Mean PSA density (ng/ml 2 ) NA Biopsy Number of cores taken > One core positive Two cores positive Mean percentage core Involvement Percentage core involvement: 50% 31.5 <50% 68.5 <5% % % % % % % % 8.44 Max percentage cancer in any core (range) 27 (1 100) 13.3 ( ) TABLE 5 Predictive accuracies of the Epstein Criteria in determining organ-confined disease, pathological Gleason score 6 and overall insignificant disease Study Bastian [21] Jeldres [22] Lee MC [26] Lee SE [25] Hekal [24] Louie-Johnsun [23] Preoperative Gleason score <6 (%) 22 (6%) 2 (1.5%) 28 (80%) =6 (%) 237 (100%) 344 (94%) 136 (100) 129 (98.5%) 7 (20%) Year of publication Post-operative Gleason Score <6 (%) 1 (0.4%) 55 (15%) 0 (0) 0 (0) 4 (11.4) =6 (%) 213 (89.9%) 223 (60.9%) 82 (60.3) 91 (69.5) 15 (42.9) =7 (%) 21 (8.9%) 88 (24.0%) 52 (38.2) 40 (30.5) 11 (31.4) =7a (%) N/A 82 (22.4%) N/A 37 (28.2) 7 (20.0) =7b (%) N/A 6 (1.6%) N/A 3 (2.3) 4 (11.4) =8 10 (%) 2 (0.8%) 0 (0%) 2 (1.5) 0 (0) 5 (14.3) Undergraded 23 (9.7%) 88 (24.1%) 54 (39.7%) 40 (30.5) 14 (40.0) 16 (21.6) Overgraded 1 (0.4%) 2 (0.5%) 0 (0) 2 (5.7) Organ confined (%) 217 (91.6) 336 (91.8) 126 (92.6) 129 (96.9) 28 (80.0) 71 (95.9) Non-organ confined (%) 20 (8.4) 30 (8.2) 10 (7.4) 4 (3.1) 7 (20.0) 3 (4.1) (lymph node invasion) 0 (0) 0 (0) 0 (0) 0 (0) 4 (11.4%) Overall insignificant prostate cancer (%) biochemical relapse-free survival. Lee used two definitions of insignificant disease: a classic Epstein definition (defined as organconfined, Gleason score <6 and tumour volume <0.5 cm 3 ) and a liberal Epstein definition (defined as organ confined, Gleason <6 and tumour of any volume). The authors reasoned that a liberal definition could be used because studies had shown that the majority of pathologically proven Gleason 6 tumours were organ-confined regardless of tumour volume [27,28]. A largescale multi-institutional trial also showed that only 0.03% of patients with organconfined Gleason <6 disease died from prostate cancer after 15 years of follow-up [29]. Overall, the Epstein Criteria cohort showed a 92% concordance with organconfined status (vs 75% concordance in the non-epstein Criteria cohort) and 60% concordance with Gleason 6 disease at RP (vs

6 EPSTEIN CRITERIA FOR INSIGNIFICANT PROSTATE CANCER 39% in the non-epstein Criteria cohort). In all, 38% of the Epstein Criteria population were upgraded to Gleason 7 and 2% had Gleason 8 10 disease (vs 60% and 1%, respectively, in the non-epstein Criteria group). The overall concordance for insignificant disease was 37% when the classic definition of cancer significance was used, and 58% concordance with the liberal definition. Of all 268 patients, 12% experienced biochemical failure within 6 years of the study. The actuarial estimated 1- and 5-year progression-free probability rates were 100% and 100% for the patients meeting Epstein s Criteria, and 98% and 83% for those that did not meet Epstein s Criteria. DISCUSSION Since its introduction in 1994, the accuracy of the Epstein Criteria for predicting for organconfined disease remains remarkable (91.6% before 2005 vs % after 2005). However, its concordance with Gleason 6 cancer fell from 91% to % before and after 2005, respectively. The overall predictability for insignificant prostate cancer also fell from 84% to 37 76%. These results suggest that the ISUP modification to Gleason grading is at least in part responsible for the reduced accuracy of the Epstein Criteria. A weakness of this review is the assumption that all the post-2005 validation articles used the modified Gleason system. In reality, it is unlikely that all these studies complied completely with the novel updated classifications throughout the timeframes of the study. In fact, it is more probable that the conventional system or indeed, both versions of the Gleason system may have been used at some stage. No validation study was also completely compliant with Epstein s initial criteria, and Bastian s study appeared to set the new limit for PSA density at 0.15 ng/ml for the studies that followed. There is considerable heterogeneity in the studies, with variations seen in the number of biopsies taken (Epstein s initial cohort had a mean and median of five cores of prostatic tissue), interpretation of biopsies, calculations of PSA densities and non-uniformity of surgical techniques. Many studies omitted the analysis of specimen tumour volumes, for example, the studies by Jeldres, Hekal and Lee (MC) which determined cancer significance based on organ-confined status and Gleason score alone. The inclusion of tumour volumes into their studies would probably increase the number of significant cancers detected, and further reduce the overall accuracy of the Epstein Criteria. By modifying the Gleason system, the qualifying criteria for the Epstein Criteria was also correspondingly changed, making the comparison of studies between the two eras difficult. As demonstrated in a recent review, the ISUP revision results in significant amounts of disease upgrading. When the conventional and modified classification systems were compared on needle biopsy specimens, the overall numbers of Gleason 6 tumours diagnosed fell from 48.4% to 22% whereas Gleason 7 numbers rose from 25.5% to 67.9% [20]. The present review shows that the Epstein Criteria validation studies also demonstrated a majority of tumour upgrades to If this finding is translated to Epstein s original study dataset, the re-evaluation of the biopsy and prostatectomy tissues under the ISUP system could possibly also see some tumours upgraded to 3 + 4, with a resultant different concordance for insignificant prostate cancer. However, with stricter criteria for Gleason scoring and fewer cases assigned a Gleason score 6 on needle biopsies under the ISUP modification, a biopsy score of 6 should actually predict insignificant cancer more accurately. If all the samples in these validation studies were re-graded under the modified system, insignificant cancers would probably be picked up with a greater accuracy, albeit at the cost of increased over diagnosis of significant cancers. After re-grading under the modified Gleason system, 95% of Gleason tumours now remain organ-confined whereas 79% of Gleason tumours are non-organ confined. The 10-year cancer-specific survival rates for Gleason and tumours are also similar at % and %, respectively. A further upgrade to Gleason status confers a reduced 10-year survival of % and three-fold risk of bone metastases [30 33]. With these statistics in mind, could it be possible for tumours to represent a key treatment point in active surveillance patients? The finding of disease could prompt the conversion of surveillance to curative therapy, allowing curative treatment of prostate cancer to commence at a point where overall prognosis is similar to disease while avoiding the poor outcome of without overtreatment. However, longer-term followup studies to compare Gleason against outcomes are required before a confident conclusion can be made, especially as most studies to date have compared only Gleason against (see Table 3). None of these comparisons were also performed on populations after 2005 or in cohorts with specifically Epstein Criteria-determined insignificant disease. Longer-term follow-up of the validation studies in the present review would help indicate if Gleason patients are, in fact, associated with similar outcomes to Recently, a long-term study on active surveillance patients also demonstrated that 30% of patients were upgraded to higher risk and were offered definitive therapy (the criteria for definitive intervention was PSA doubling time <3 years or biopsy Gleason score progression of or higher). Fifty per cent of these radically treated patients experienced PSA failure, even although the overall and 10-year cancer specific survival was 78.6% and 97.2%, respectively [34]. Another criticism of this review is that most of the validation articles did not include tumour volumes in their analyses. Lee (MC) and other authors challenged the role of tumour volumes in determining cancer significance, given that the majority of tumours were organ-confined regardless of tumour volume and associated with excellent 15-year survivals independent of disease stage [27 29]. An article by Isbarn, however, showed that high volume Gleason cancers (three or more biopsy cores positive) actually fared more poorly than low volume tumours (two or less biopsy cores positive) and were more closely related to low volume cancers. High volume tumours were associated with twice the rate of extraprostatic disease compared with low volume tumours (11.7% vs 23.3%, respectively), and lower 5-year biochemical recurrence-free survival (RFS) (77.8% vs 95%, respectively) [35]. Again, Isbarn s results were collected between 1991 and 2006, meaning that many patients could now be under the modified system. The results, however, provide some leverage towards the consideration of low volume tumours (2 or less positive cores) to count as insignificant cancers (the 5-year biochemical RFS rate for these patients was 81.2%, higher than that for high volume patients)

7 OON ET AL. Lee s (MC) validation study [21] used two different classifications of insignificant disease based on tumour volumes. They did not, however, provide a comparison of outcomes in these two groups, providing only a comparison of Epstein Criteria vs non- Epstein Criteria patients. By increasing the threshold for tumour volume in insignificant cancer, the authors naturally demonstrated an increased detection rate of insignificant cancer in the Epstein Criteria and non-epstein Criteria groups. The analysis of individual outcomes and pathological characteristics of the groups with different tumour volumes would have contributed better to the understanding of tumour volumes in cancer significance. In conclusion, the Epstein Criteria is the most commonly used tool for predicting insignificant prostate cancer. After the 2005 ISUP modification to the Gleason grading system, the accuracy of the Epstein Criteria to predict for insignificant prostate cancer fell from 73 84% to 39 76%. The accuracy for predicting a Gleason score 6 was also reduced from 90.3% to %, but organ-confined status predictive ability remained favourable. Most of the Epstein Criteria validation studies did not apply the criteria in a proper way, which leads to a lack of valuable validation, regardless of ISUP updates of Gleason score. As no article indicated what Gleason system was in use in their study, it is also difficult to determine if the ISUP modification is directly responsible for the decreased accuracy of the Epstein Criteria. However, the abrupt change in predictive ability of Gleason score and overall cancer significance after this time suggests that the Gleason grading system is at least in part responsible. Regardless of whether the ISUP modification has affected the Epstein Criteria or not, based on these validation studies, the present review has demonstrated a slippage in the accuracy of the Epstein Criteria and a need for an agreed system to determine for insignificant prostate cancer. As a result of the heterogeneity of these validation studies, the Epstein Criteria has not been properly validated and the conclusions are not transferable into routine daily practice. It will therefore be necessary to perform a large single or multicentre assessment with clearly defined criteria. The present review highlights a potential limitation in contemporary predictive nomograms that incorporate parameters such as Gleason scoring which are liable to amendments. A change in these scoring parameters could result in subtle yet major effects to the nomogram itself, with subsequent reduced accuracy of the entire predictive tool. PSA-screening appears to be inevitable in many countries, and efforts should be directed towards the development of better methods for identifying insignificant prostate cancer in the preoperative patient, so as to minimize over diagnosis and overtreatment in patients with insignificant prostate cancer while not compromising on the detection of early and significant prostate cancer in others. ACKNOWLEDGEMENTS This work was supported by the Royal College of Surgeons in Ireland, The Urological Foundation and the Prostate Cancer Research Consortium, funded by the Irish Cancer Society. CONFLICT OF INTEREST None declared. REFERENCES 1 Oliver SE, May MT, Gunnell D. International trends in prostate-cancer mortality in the PSA ERA. Int J Cancer 2001; 92: Helgesen F, Holmberg L, Johansson JE, Bergström R, Adami HO. Trends in prostate cancer survival in Sweden, 1960 through 1988: evidence of increasing diagnosis of nonlethal tumors. J Natl Cancer Inst 1996; 88: Albertsen PC, Fryback DG, Storer BE, Kolon TF, Fine J. Long-term survival among men with conservatively treated localized prostate cancer. JAMA 1995; 274: Rullis I, Schaeffer JA, Lilien OM. Incidence of prostatic carcinoma in the elderly. Urology 1975; 6: Ilic D, O Connor D, Green S, Wilt T. Screening for prostate cancer: a Cochrane systematic review. Cancer Causes Control 2007; 18: Wilt TJ, Brawer MK, Barry MJ et al. The Prostate Cancer Intervention Versus Observation Trial: VA/NCI/AHRQ Cooperative Studies Program #407 (PIVOT): design and baseline results of a randomized controlled trial comparing radical prostatectomy to watchful waiting for men with clinically localized prostate cancer. Contemp Clin Trials 2009; 30: Schröder FH, Hugosson J, Roobol MJ et al. ERSPC Investigators. Screening and prostate-cancer mortality in a randomized European study. N Engl J Med 2009; 360: Postma R, Schröder FH, van Leenders GJ et al. Cancer detection and cancer characteristics in the The European Randomized Screening for Prostate Cancer (ERSPC) Section Rotterdam. A comparison of two rounds of screening. Eur Urol 2007; 52: Epstein JI, Walsh PC, Carmichael M, Brendler CB. Pathologic and clinical findings to predict tumor extent of nonpalpable (stage T1c) prostate cancer. JAMA 1994; 271: Epstein JI, Pizov G, Walsh PC. Correlation of pathologic findings with progression following radical retropubic prostatectomy. Cancer 1993; 71: Epstein JI, Carmichael M, Partin AW, Walsh PC. Is tumor volume an independent predictor of progression following radical prostatectomy? A multivariate analysis of 185 clinical stage B adenocarcinomas of the prostate with five years follow-up. J Urol 1993; 149: Stamey TA, Freiha FS, McNeal JE, Redwine EA, Whittemore AS, Schmid HP. Localized prostate cancer. Relationship of tumor volume to clinical significance for treatment of prostate cancer. Cancer 1993; 71: Epstein JI, Allsbrook WC, Amin MB, Egevad L, 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: Helpap B, Egevad L. Modified Gleason Grading. An updated review. Histol Histopathol 2009; 24: Gleason DF. Classification of prostatic carcinomas. Cancer Chemother Rep 1966; 50: Helpap B, Egevad L. The significance of modified Gleason grading of prostatic carcinoma in biopsy and radical prostatectomy specimens. Virchows Arch 2006; 449: Helpap B, Egevad L. [The value of the modified Gleason grading system of

8 EPSTEIN CRITERIA FOR INSIGNIFICANT PROSTATE CANCER prostatic carcinoma in routine urological diagnostics]. Urologe A 2007; 46: Helpap B, Egevad L. [Clinical insignificance of prostate cancer, are there morphological findings]. Urologe A 2009; 48: Montironi R, Cheng L, Lopez-Beltran A et al. Original Gleason system versus 2005 ISUP modified Gleason system: the importance of indicating which system is used in the patient s pathology and clinical reports. Eur Urol 2010; 58: Epstein JI. An update of the Gleason grading system. J Urol 2010; 183: Bastian PJ, Mangold LA, Epstein JI, Partin AW. Characteristics of insignificant clinical T1c prostate tumors. A contemporary analysis. Cancer 2004; 101: Jeldres C, Suardi N, Walz J et al. Validation of the contemporary epstein criteria for insignificant prostate cancer in European men. Eur Urol 2008; 54: Louie-Johnsun M, Neill M, Treurnicht K, Jarmulowicz M, Eden C. Final outcomes of patients with low-risk prostate cancer suitable for active surveillance but treated surgically. BJU Int 2009; 104: Hekal IA, El-Tabey NA, Nabeeh MA et al. Validation of Epstein criteria of insignificant prostate cancer in Middle East patients. Int Urol Nephrol 2010; 42: Lee SE, Kim DS, Lee WK et al. Application of the Epstein criteria for prediction of clinically insignificant prostate cancer in Korean men. BJU Int 2010; 105: Lee MC, Dong F, Stephenson AJ, Jones JS, Magi-Galluzzi C, Klein EA. The Epstein Criteria predict for organconfined but not insignificant disease and a high likelihood of cure at Radical Prostatectomy. Eur Urol 2010; 58: De la Taille A, Antiphon P, Salomon L et al. Prospective evaluation of a 21- sample needle biopsy procedure to designed to improve the prostate cancer detection rate. Urology 2003; 61: Makhlouf AA, Krupski TL, Kunkle D, Theodorescu D. The effect of sampling more cores on the predictive accuracy of pathological grade and tumor distribution in the prostate biopsy. BJU Int 2004; 93: Nelson WG, DeMarzo AM, Isaacs WB. Prostate cancer. N Engl J Med 2003; 349: Helpap B, Egevad L. Correlation of modified Gleason grading with pt stage of prostatic carcinoma after radical prostatectomy. Anal Quant Cytol Histol 2008; 30: Wright JL, Salinas CA, Lin DW et al. Prostate cancer specific mortality and Gleason 7 disease differences in prostate cancer outcomes between cases with Gleason and Gleason tumors in a population based cohort. J Urol 2009; 182: Isbarn H, Wanner M, Salomon G et al. Long-term data on the survival of patients with prostate cancer treated with radical prostatectomy in the prostatespecific antigen era. BJU Int 2010; 106: Stark JR, Perner S, Stampfer MJ et al. Gleason score and lethal prostate cancer: does = 4 + 3? J Clin Oncol 2009; 27: Klotz L, Zhang L, Lam A, Nam R, Mamedov A, Loblaw A. Clinical results of long-term follow-up of a large, active surveillance cohort with localized prostate cancer. J Clin Oncol 2010; 28: Isbarn H, Karakiewicz PI, Ahyai SA et al. Differences in histopathological and biochemical outcomes in patients with low Gleason score prostate cancer. BJU Int 2010; 105: Chan TY, Partin AW, Walsh PC, Epstein JI. Prognostic significance of Gleason Score 3+4 versus Gleason Score 4+3 tumor at radical prostatectomy. Urology 2000; 56: Lau WK, Blute ML, Bostwick DG, Weaver AL, Sebo TJ, Zincke H. Prognostic factors for survival of patients with pathological Gleason score 7 prostate cancer: differences in outcome between 3 and 4. J Urol 2001; 166: Khoddami SM, Shariat SF, Lotan Y et al. Predictive value of primary Gleason pattern 4 in patients with Gleason score 7 tumours treated with radical prostatectomy. BJU Int 2004; 94: Han M, Snow PB, Epstein JI et al. A neural network predicts progression for men with gleason score 3+4 versus 4+3 tumors after radical prostatectomy. Urology 2000; 56: Tollefson MK, Leibovich BC, Slezak JM, Zincke H, Blute ML. Long-term prognostic significance of primary Gleason pattern in patients with Gleason score 7 prostate cancer: impact on prostate cancer specific survival. J Urol 2006; 175: Rasiah KK, Stricker PD, Haynes AM et al. Prognostic significance of Gleason pattern in patients with Gleason score 7 prostate carcinoma. Cancer 2003; 98: Kang DE, Fitzsimons NJ, Presti JC Jr et al. SEARCH Database Study Group. Risk stratification of men with Gleason score 7 to 10 tumors by primary and secondary Gleason Score: results from the SEARCH Database. Urology 2007; 70: Burdick MJ, Reddy CA, Ulchaker J et al. Comparison of biochemical relapse-free survival between primary Gleason score 3 and primary Gleason score 4 for biopsy Gleason score 7 prostate cancer. Int J Radiat Oncol Biol Phys 2009; 73: Correspondence: Sheng F. Oon, Conway Institute, University College Dublin, Dublin 4, Ireland. sheng-fei.oon@ucd.ie Abbreviations: AS, active surveillance; RFS, recurrence-free survival; ECE, extracapsular extension; SVI, seminal vesicle invasion; PSM, positive surgical margins; PNI, perineural invasion; LNI, lymph node invasion