Multifocal High Grade Prostatic Intraepithelial Neoplasia is a Significant Risk Factor for Prostatic Adenocarcinoma

Multifocal High Grade Prostatic Intraepithelial Neoplasia is a Significant Risk Factor for Prostatic Adenocarcinoma Jennifer L. Merrimen,* Glenn Jones, Danielle Walker, Chung S. Leung, Linda R. Kapusta and John R. Srigley From the McMaster University (JLM, GJ, DW, LRK, JRS), Hamilton and Mount Sinai Hospital (JRS) and University of Toronto, Toronto (CSL, LRK), Ontario, and Queen Elizabeth II Health Sciences Centre (JLM), Halifax, Nova Scotia, Canada

Purpose: There is debate in the literature on the role of high grade prostatic intraepithelial neoplasia as a risk factor for subsequent prostatic adenocarcinoma detection on prostatic needle biopsy. We determined whether high grade prostatic intraepithelial neoplasia on initial prostatic needle biopsy is an independent risk factor for prostatic adenocarcinoma and whether differences exist between prostatic adenocarcinoma in patients with previous high grade prostatic intraepithelial neoplasia and those with a benign diagnosis. Materials and Methods: Pathological findings in prostatic needle biopsies in 12,304 men who underwent initial prostatic needle biopsy in an 8-year period were analyzed. Patients were included in the analysis when the initial diagnosis was high grade prostatic intraepithelial neoplasia alone or a benign diagnosis and at least 1 followup prostatic needle biopsy was performed. The primary study outcome was prostatic adenocarcinoma and secondary outcome measurements were cancer characteristics, such as Gleason score and extent of tissue involvement with prostatic adenocarcinoma. Results: In the high grade prostatic intraepithelial neoplasia group of 564 patients and the benign group of 845, 27.48% and 22.01%, respectively, were diagnosed with prostatic adenocarcinoma on followup prostatic needle biopsy (p ⫽ 0.02). When age, prostate specific antigen and sampling extent were adjusted for, the adenocarcinoma risk after an initial diagnosis of high grade prostatic intraepithelial neoplasia remained significant (OR 1.38, p ⫽ 0.03). The risk was related to the extent of high grade prostatic intraepithelial neoplasia in the initial sample with a greater likelihood of adenocarcinoma when multiple prostatic sites were involved by high grade prostatic intraepithelial neoplasia. Patients in whom prostatic adenocarcinoma developed after a benign diagnosis on initial prostatic needle biopsy had greater tumor volume. However, mean followup was longer in the benign group than in the high grade prostatic intraepithelial neoplasia group (2.35 vs 1.36 years). Conclusions: Patients with an initial diagnosis of high grade prostatic intraepithelial neoplasia, especially when multifocal, are at greater risk for subsequent prostatic adenocarcinoma than those with a benign diagnosis. Results suggest that followup should be more rigorous in patients with multifocal high grade prostatic intraepithelial neoplasia.

Abbreviations and Acronyms ASAP ⫽ atypical small acinar proliferation DRE ⫽ digital rectal examination HGPIN ⫽ high grade prostatic intraepithelial neoplasia PCa ⫽ prostatic adenocarcinoma PNB ⫽ prostatic needle biopsy PSA ⫽ prostate specific antigen Submitted for publication December 15, 2008. * Correspondence: Division of Anatomical Pathology, Queen Elizabeth II Health Sciences Centre, Room 724, Mackenzie Building, 5788 University Ave., Halifax, Nova Scotia, Canada B3H 1V8 (telephone: 902-473-2931; Fax: 902-473-1049; e-mail: [email protected]).

Key Words: prostate, adenocarcinoma, prostatic intraepithelial neoplasia, risk, biopsy PROSTATE cancer represents about 28% of all new cancers diagnosed in men

and is responsible for about 11% of cancer deaths in men.1 PNB is rou-

0022-5347/09/1822-0485/0 THE JOURNAL OF UROLOGY® Copyright © 2009 by AMERICAN UROLOGICAL ASSOCIATION

Vol. 182, 485-490, August 2009 Printed in U.S.A. DOI:10.1016/j.juro.2009.04.016

www.jurology.com

485

486

PROSTATIC INTRAEPITHELIAL NEOPLASIA AS PROSTATE CANCER RISK FACTOR

tinely performed to evaluate patients with abnormal DRE and/or serum PSA. In 5% to 10% of cases biopsy does not reveal clearly benign or clearly malignant findings but rather HGPIN and/or ASAP.2 HGPIN is a neoplastic transformation of the lining epithelium of preexisting prostatic ducts and acini without stromal invasion that shares a clinical, morphological and genetic relationship with PCa.3,4 Although there is a definite association between HGPIN and PCa, there is still debate on whether HGPIN is an independent risk factor for subsequent PCa. Two recent reviews of this issue by expert pathologists yielded differing conclusions.2,5 One expert concluded that men do not need repeat PNB within year 1 after HGPIN diagnosis and proposed that more studies are needed to determine whether PNB should be performed several years after HGPIN diagnosis.2 The second expert concluded that HGPIN diagnosed in an extended PNB sampling warranted repeat PNB.5 Each review assessed a large number of published studies beginning in 19912 and 1995,5 respectively. However, many of these studies had design flaws, including lack of a valid control group and a small study population.2 We used a Canadian PNB database to address whether HGPIN alone on PNB is an independent risk factor for PCa and whether HGPIN quantity on initial PNB adjusts the risk of PCa. We also assessed whether PCa in patients with a previous HGPIN diagnosis is more aggressive than PCa in patients with a previous benign diagnosis.

Patients were assigned to a diagnostic category based on initial PNB pathology results regardless of whether HGPIN, ASAP or PCa was diagnosed on subsequent biopsy. HGPIN diagnostic criteria were those described by Bostwick et al.4 This category includes patients with HGPIN as the only significant abnormality in the sample. Patients with ASAP alone or ASAP with HGPIN were included in the ASAP diagnostic category. Patients with HGPIN or ASAP and a diagnosis of adenocarcinoma were included in the PCa diagnostic category. Patients were eligible for study when the initial diagnosis was HGPIN alone (unifocal or multifocal) or benign. Patients were included in analysis when at least 1 followup PNB was performed. When assessing PCa characteristics, such as Gleason score, percent tissue surface area involvement etc, the first PNB in which PCa was detected was used for analysis.

Statistical Analysis Data were transferred to Stata®, version 8 for statistical analysis. The Pearson chi-square test was used to compare the probability of PCa after a benign and a HGPIN diagnosis. Time to event plots were developed by the Kaplan-Meier method and assessed for significant differences by the log rank test, including log rank for trends when appropriate. Logistic regression analysis was done to assess the odds of PCa with a previous diagnosis of HGPIN, controlling for patient age, initial serum PSA, sampling extent and sign-out pathologist. Further regression analysis was done to assess the impact of HGPIN at multiple prostatic sites on PCa development. The Pearson chi-square and Student t tests, and ANOVA were used to compare patient and cancer characteristics in the HGPIN vs benign and unifocal vs multifocal HGPIN groups. Initial PSA values were transformed to log PSA for statistical analysis. Values are shown as the mean ⫾ SD.

MATERIALS AND METHODS Patients and Data Collection Biopsies were acquired from the practice of 28 community based urologists working in a large Canadian city. Clinical and pathological information was collected on 15,268 PNBs performed between May 1999 and June 2007. Biopsies were performed typically for PSA or DRE abnormalities. PNBs were interpreted by 3 experienced pathologists with expertise in urological pathology. Clinical variables were patient age and PSA. Pathological variables were sign-out pathologist, diagnosis, sampling extent (number of sites biopsied and total number of cores), number of sites with HGPIN, Gleason score, Gleason pattern, percent tissue involvement, perineural invasion and periprostatic fat invasion. Data were entered into a Microsoft® Access™ database.

Study Design PNBs were initially excluded if there was previous treatment, including chemotherapy, radiation of radical prostatectomy, ie prostatic bed biopsies, or previous PNB not available in the database. After these exclusions 12,304 patients underwent initial PNB during the study period. The diagnosis was benign in 4,938 patients (40.1%), HGPIN in 1,283 (10.4%), ASAP in 329 (2.7%) and PCa in 5,754 (46.8%).

RESULTS Overall HGPIN and Benign Study Population Characteristics During the study period 4,938 patients (40.1%) had a benign diagnosis and 1,283 (10.4%) had a HGPIN diagnosis. Of these patients 1,409 underwent at least 1 subsequent PNB, including 845 with a benign condition and 564 with HGPIN alone. The total number of PNBs performed per group was similar at 2.28 and 2.36 in the benign and HGPIN groups, respectively (table 1). However, mean followup was longer in the benign group than in the HGPIN group (2.4 vs 1.4 years) and patients with an initial diagnosis of HGPIN underwent repeat PNB sooner than those with an initial benign diagnosis (0.9 vs 1.9 years) (table 1). HGPIN as PCa Risk Factor A greater number of patients with HGPIN on initial PNB subsequently had PCa compared to those with a benign diagnosis on initial PNB (27.48% vs 22.01%, p ⫽ 0.019, table 1). Logistic regression analysis controlling for patient age, PSA, sampling ex-

PROSTATIC INTRAEPITHELIAL NEOPLASIA AS PROSTATE CANCER RISK FACTOR

Age Serum PSA (ng/ml) Log PSA No. cores No. sites No. followup PNBs Yrs to 1st followup PNB Followup (yrs) No. PCa

Mean ⫾ SD Benign

Mean ⫾ SD HGPIN

61.3 ⫾ 7.2 6.97 ⫾ 3.63 1.82 ⫾ 0.51 7.48 ⫾ 1.79 6.21 ⫾ 0.98 2.28 ⫾ 0.59 1.93 ⫾ 1.41 2.35 ⫾ 1.60 186

64.6 ⫾ 7.38 7.86 ⫾ 14.61 1.80 ⫾ 0.65 8.09 ⫾ 2.10 6.38 ⫾ 1.08 2.36 ⫾ 0.72 0.91 ⫾ 1.00 1.36 ⫾ 1.32 155

p Value ⬍0.001 0.003 ⬍0.00005 0.0015 0.13 ⬍0.00005 ⬍0.00005 0.018

tent and sign-out pathologist indicated that HGPIN remained an independent PCa risk factor (OR 1.38, p ⫽ 0.033, table 2). This analysis included 1,120 of the 1,409 patients since PSA values were unavailable in 289. Patients with unifocal HGPIN on initial PNB had no significantly different likelihood of PCa than those with a benign diagnosis on initial PNB (table 3). When assessing the impact of multifocal HGPIN on PCa, the OR generally increased with an increasing number of sites involved with HGPIN (table 3). The figure shows the plot for time to cancer in the benign, unifocal HGPIN and multifocal HGPIN groups (log rank test for trend p ⬍0.00005). Characteristics of Patients With PCa in Benign and HGPIN Groups In the benign group patients with PCa were slightly older (62.5 vs 61.0 years) and had slightly less extensive prostate sampling on initial biopsy (7.2 vs 7.6 cores) than those without PCa (data not shown). In the HGPIN group results were similar with PCa developing in older patients (66.2 vs 63.9 years) and in those with slightly less extensive prostate sampling on initial biopsy (7.7 vs 8.2 cores) than in those without PCa (data not shown). Also, in the benign group patients with PCa had slightly higher PSA at initial PNB (log PSA 1.91 vs 1.76) than those without PCa (data not shown). Patients with PCa after a HGPIN diagnosis were slightly older (66.2 vs 62.5 years) and underwent more extensive sampling at initial PNB (7.7 vs 7.2 cores and

Table 3. Multifocal HGPIN and PCa risk No. HGPIN Sites

No. Pts

OR (95% CI)

p Value

0 1 2 3 4 5

845 288 181 64 21 10

1.00 1.02 (0.73–1.40) 1.55 (1.08–2.21) 1.99 (1.16–3.40) 2.66 (1.10–6.40) 2.36 (0.66–8.46)

0.919 0.016 0.012 0.029 0.187

6.3 vs 6.1 sites) than patients with PCa after a benign diagnosis (table 4). There was no significant difference in initial PSA. However, the average change in PSA between initial PNB and the first PNB for PCa was 9.57 and 3.34 ng/ml in the benign and HGPIN groups, respectively (table 4). Patients with PCa after a diagnosis of multifocal HGPIN had higher initial PSA (8.86 vs 6.47 ng/ml) than patients with PCa after a unifocal HGPIN diagnosis (data not shown). No differences were noted in age or prostate sampling extent at initial PNB between these 2 groups (data not shown). PCa Characteristics and Initial PNB Diagnosis HGPIN or benign diagnosis. Patients with PCa after a benign diagnosis had greater tumor volume, as represented by a greater number of cores (1.9 vs 1.6) and sites (1.7 vs 1.5) involved by PCa, and a greater percent of positive cores (22.7% vs 18.9%), positive sites (25.7% vs 22.5%) and tissue involvement by PCa (4.8% vs 3.1%) (table 5). There were no significant differences in Gleason score, or primary or secondary Gleason grade. Patients with PCa after a benign diagnosis had a greater volume of high grade Gleason patterns 4 and 5 than patients with PCa after a HGPIN diagnosis (13.5% vs 7.7%). There Kaplan-Meier survival estimates 1.00 proportion free from cancer

Table 1. Patient characteristics in benign and HGPIN groups, and PCa risk after benign or HGPIN diagnosis

487

0.75

0.50

0.25

0.00

Table 2. PCa risk on multivariate logistic regression analysis controlling for age, PSA, sampling extent and sign-out pathologist Variable

OR

p Value

HGPIN diagnosis Log PSA Total No. cores Age Pathologist

1.38 1.28 .87 1.01 .99

0.033 0.540 0.001 0.053 0.960

0

2

4

6

8

20 4 3

0 0 0

years following first biopsy Number at risk Benign 845 HGPIN unifocal 287 HGPIN multifocal 277

430 73 56

135 14 16

Kaplan-Meier plot for time to cancer event by initial PNB pathology (log rank test for trend 2-tailed p ⬍0.00005). Solid line indicates benign findings. Short dashed line indicates unifocal HGPIN. Long dashed line indicates multifocal HGPIN.

488

PROSTATIC INTRAEPITHELIAL NEOPLASIA AS PROSTATE CANCER RISK FACTOR

Table 4. Characteristics of patients with PCa after initial PNB benign or HGPIN diagnosis

Age PSA (ng/ml) Log PSA Total No. cores biopsied Total No. sites biopsied PSA at biopsy 1 PSA at PCa detection PSA change

Mean ⫾ SD Benign

Mean ⫾ SD HGPIN

p Value

62.5 ⫾ 7.7 7.10 ⫾ 3.31 1.87 ⫾ 0.42 7.16 ⫾ 1.59 6.09 ⫾ 0.87 7.10 ⫾ 3.31 16.67 ⫾ 77.83 9.57

66.2 ⫾ 7.3 7.91 ⫾ 6.23 1.91 ⫾ 0.52 7.69 ⫾ 2.01 6.32 ⫾ 0.90 7.91 ⫾ 6.23 11.25 ⫾ 11.73 3.34

⬍0.01 Not appropriate 0.53 ⬍0.01 0.02 Not appropriate Not appropriate Not appropriate

was no significant difference in the percent of cases with perineural invasion or periprostatic fat invasion. Although patients with a HGPIN diagnosis on initial PNB had smaller volume PCa, there was no significant difference in the number of minimal or limited PCas, as defined by PCa with only Gleason pattern 3 on 1 core less than 1 mm in linear extent and involving less than 5% of the core tissue surface area.6 Unifocal or multifocal HGPIN diagnosis. Patients with PCa after a unifocal or multifocal HGPIN diagnosis on initial PNB had no significant differences in Gleason score, or primary or secondary Gleason grade (data not shown). Patients diagnosed with multifocal HGPIN on initial PNB had a higher volume of high grade Gleason patterns 4 and 5 (2.9 vs 11.2, p ⫽ 0.014). No significant difference was detected in PCa extent, as measured by the percent tissue surface area involved by PCa, the number of cores or sites involved by PCa, or the percent of positive cores or positive sites. There was no significant difference in perineural and periprostatic fat invasion (data not shown).

DISCUSSION HGPIN is pathogenically related to PCa but there is debate in the literature on its current role as a risk factor for subsequent PCa detection. Two large literature reviews summarizing published studies showed a wide range in the incidence of patients with PCa after a HGPIN diagnosis (2% to 100%).2,5 However, many reviewed studies were limited by study design flaws, including lack of a valid control group and a small study population.2 Our results show that patients diagnosed with HGPIN on initial PNB are at a greater risk for PCa than patients with a benign diagnosis on initial PNB. This risk is independent of age, PSA, sampling extent and sign-out pathologist. The risk of subsequent PCa varies with the number of sites involved with HGPIN, such that the risk of PCa generally increases with a greater number of sites with HGPIN and becomes significant when 2 or more sites are affected. Previous studies of this issue showed conflicting results.7–16 We studied a large group and results suggest that followup should be more rigorous in patients with multifocal HGPIN. There is bias in this study since the HGPIN and benign groups showed a high PCa rate (crude risk 27.5% and 22.0%, respectively, see figure). Of 4,938 patients with benign findings in the database only 845 (17.1%) went on to another biopsy. Patients in the benign group who underwent followup PNB were slightly younger and underwent slightly less extensive sampling on initial PNB than those who did not undergo followup PNB (data not shown). PSA at initial PNB was lower in benign cases without repeat PNB (data not shown). Possibly clinicians noted lower PSA and more extensive sampling on initial PNB, and were more confident of a benign diagnosis, choosing not to repeat PNB in such pa-

Table 5. PCa characteristics after initial PNB benign or HGPIN diagnosis

Mean ⫾ SD Gleason score Mean ⫾ SD Gleason grade: No. 1 No. 2 No. 3 % 4/5 Mean ⫾ SD sites: No. involved % Pos Mean ⫾ SD cores: No. involved % Pos Mean ⫾ SD % tissue area involvement % Perineural invasion % Periprostatic fat invasion No. min or limited PCa/total No. (%)*

Benign

HGPIN

p Value

6.48 ⫾ 0.75

6.30 ⫾ 0.64

0.10

3.15 ⫾ 0.40 3.32 ⫾ 0.54 0.09 ⫾ 0.59 13.5 ⫾ 28.0

3.06 ⫾ 0.25 3.24 ⫾ 0.52 0.02 ⫾ 0.24 7.7 ⫾ 20.5

0.06 0.19 0.47 0.04

1.71 ⫾ 1.11 25.8 ⫾ 17.0

1.48 ⫾ 0.84 22.5 ⫾ 12.9

0.04 0.05

1.9 ⫾ 1.4 22.7 ⫾ 15.9 4.8 ⫾ 7.9 17.2 0.53 60/185 (32.4)

1.62 ⫾ 1.0 18.9 ⫾ 11.9 3.1 ⫾ 4.0 10.3 0 56/155 (36.1)

0.02 0.02 0.01 0.07 0.36 0.49

* Only Gleason pattern 3 on 1 core, less than 1 mm in linear extent and involving less than 5% of core tissue surface area.

PROSTATIC INTRAEPITHELIAL NEOPLASIA AS PROSTATE CANCER RISK FACTOR

tients. However, because complete PSA profiles, ie serial values with time, were not analyzed, it is difficult to comment on PSA trends that may have prompted clinicians to repeat biopsy in some patients despite benign findings on initial PNB. Also, detailed DRE findings and family history data were not available, which are typically considered in clinical decision making. Regardless, our control group is not ideal and it showed a high rate of PCa, likely higher than would be found in a more representative sample. The PCa detection rate after an initial HGPIN diagnosis has decreased from 40% to 50% in the early 1990s to 25% to 33% in recent studies.2,5 Our results are within the contemporary range with a 27.5% PCa detection rate after an initial HGPIN diagnosis. Recent literature indicates that the change in PNB sampling from sextant to extended or double sextant protocols is largely responsible for this decrease.15 We studied a combination of sextant and extended protocols with a mean of 7.7 cores per patient. Like the study by Herawi et al,15 our series supports the importance of sampling. Patients with initially benign PNB findings who later had PCa underwent less extensive sampling on initial PNB and patients with HGPIN on initial PNB but without PCa during followup underwent more extensive sampling on initial PNB (mean 7.2 vs 8.2 cores). When assessing PCa characteristics in patients with a previous benign or HGPIN diagnosis on initial PNB, our study shows that PCa after a benign diagnosis is more extensive than PCa after a HGPIN diagnosis, as measured by greater tissue surface area, core and site involvement. An explanation may be that patients in the HGPIN group underwent biopsy more frequently than patients in the benign group (0.91 vs 1.93 years to first followup PNB) and PCa was detected at an earlier time in the disease course. The PSA change between initial PNB and the first PNB detecting PCa also showed a greater change in the benign group, correlating to a greater

489

tumor burden. Clinicians may also have a lower threshold for PSA changes in patients with a previous diagnosis of HGPIN, prompting earlier repeat PNB. When assessing PCa characteristics after a diagnosis of unifocal vs multifocal HGPIN on initial PNB, we observed no differences in Gleason score, primary or secondary Gleason grade, percent tissue surface area, number of cores or sites involved with PCa, percent of positive cores or sites, or periprostatic fat invasion. However, patients with multifocal HGPIN had PCa with a higher percent of high grade Gleason patterns 4 and 5, suggesting that these lesions may behave in a more aggressive manner. More thorough evaluation of this observation requires correlation with tumor extent and final Gleason grading in radical prostatectomy specimens, which was beyond the scope of this series. A limitation of our study is the lack of DRE data for analysis. Early analysis and review of those data revealed that DRE data were often reported using vague terms, resulting in difficulty with interpreting whether findings were truly suspicious for carcinoma. In addition, family history data were not available but they are usually considered in clinical decisions. Biopsies were interpreted by experienced urological pathologists. HGPIN is a morphological diagnosis and 1 study showed that nonexpert pathologists have significant variability in making this diagnosis (␬ 0.45).17 Our results may not be reliably transferred to all practice settings in which nonexpert pathologists examine biopsies.

CONCLUSIONS When HGPIN is multifocal on PNB, it is an independent risk factor for PCa. Results suggest that followup should be more rigorous in patients with multifocal HGPIN. We recommend that patients with multifocal HGPIN undergo repeat biopsy within 1 year.

REFERENCES 1. Canadian Cancer Statistics 2008. Toronto: Canadian Cancer Society/National Cancer Institute of Canada 2008.

prostate cancer. Overview of the prostate committee report. Scand J Urol Nephrol, suppl., 2000; 205: 1.

2. Epstein JI and Herawi M: Prostate needle biopsies containing prostatic intraepithelial neoplasia or atypical foci suspicious for carcinoma: implications for patient care. J Urol 2006; 175: 820.

5. Schlesinger C, Bostwick DG and Iczkowski KA: High-grade intraepithelial neoplasia and atypical small acinar proliferation: predictive value for cancer in current practice. Am J Surg Pathol 2005; 29: 1201.

3. Bostwick DG and Qian J: High-grade prostatic intraepithelial neoplasia. Mod Pathol 2004; 17: 360. 4. Bostwick DG, Norlén BJ and Denis L: Prostatic intraepithelial neoplasia: the preinvasive stage of

6. Rubin MA, Dunn R, Kambham N et al: Should a Gleason score be assigned to a minute focus of carcinoma on prostate biopsy? Am J Surg Pathol 2000; 24: 1634.

7. Akhavan A, Keith JD, Bastacky SI et al: The proportion of cores with high-grade prostatic intraepithelial neoplasia on extended pattern needle biopsy is significantly associated with prostate cancer on site directed repeat biopsy. BJU Int 2007; 99: 765. 8. Roscigno M, Scattoni V, Freschi M et al: Monofocal and plurifocal high-grade prostatic intraepithelial neoplasia on extended prostate biopsies: factors predicting cancer detection. Urology 2004; 63: 1105. 9. Naya Y, Ayala AG, Tamboli P et al: Can the number of cores with high grade prostatic intra-

490

PROSTATIC INTRAEPITHELIAL NEOPLASIA AS PROSTATE CANCER RISK FACTOR

epithelial neoplasia predict cancer in men who undergo repeat biopsy? Urology 2004; 63: 503. 10. Bishara T, Ramnani DM and Epstein JI: Highgrade prostatic intraepithelial neoplasia on needle biopsy: risk of cancer on repeat biopsy related to number of involved cores and morphologic pattern. Am J Surg Pathol 2004; 28: 629. 11. Netto GJ and Epstein JI: Widespread high-grade prostatic intraepithelial neoplasia on prostatic needle biopsy: a significant likelihood of subsequently diagnosed adenocarcinoma. Am J Surg Pathol 2006; 30: 1184.

12. Borboroglu PG, Sur RL, Roberts JL et al: Repeat biopsy strategy in patients with atypical small acinar proliferation or high grade prostatic intraepithelial neoplasia on initial prostate needle biopsy. J Urol 2001; 166: 866.

15. Herawi M, Kahane H, Cavallo C et al: Risk of prostate cancer on first re-biopsy within 1 year following a diagnosis of high grade prostatic intraepithelial neoplasia is related to the number of cores sampled. J Urol 2006; 175: 121.

13. Kamoi K, Troncoso P and Babaian RJ: Strategy for repeat biopsy in patients with high grade prostatic intraepithelial neoplasia. J Urol 2000; 163: 819.

16. Kronz JD, Allan CH, Shaikh AA et al: Predicting cancer following a diagnosis of high-grade prostatic intraepithelial neoplasia on needle biopsy data on men with more than one follow-up biopsy. Am J Surg Pathol 2001; 25: 1079.

14. San Francisco IF, Olumi AF, Kao J et al: Clinical management of prostatic intraepithelial neoplasia as diagnosed by extended needle biopsies. BJU Int 2003; 91: 350.

17. Allam CK, Bostwick DG, Hayes JA et al: Interobserver variability in the diagnosis of high-grade prostatic intraepithelial neoplasia and adenocarcinoma. Mod Pathol 1996; 9: 742.

EDITORIAL COMMENT This study of multifocal HGPIN in PNBs provides valuable practical information about the significance of HGPIN and suggests the most appropriate clinical response to this diagnosis. The probability of carcinoma on rebiopsy after an initial diagnosis of PCa HGPIN decreased from 40% to 50% in the 1990s to 20% to 35% currently, purportedly due to the increase from 6 to 10 to 14 cores on initial biopsy (reference 15 in article). In the current study the average number of cores is 7.7 at a total of 6.2 sites. However, the 27.5% carcinoma rate on rebiopsy is

similar to that of 10 to 14 core biopsies. Furthermore, HGPIN on more than 1 core significantly increases the probability of carcinoma on rebiopsy. This observation is supported by other recent studies.1,2. The advice to rebiopsy patients with multifocal HGPIN within 1 year is a valuable clinical guideline. Maria Shevchuk Department of Pathology Weill Cornell Medical College New York, New York

REFERENCES 1. Abdel-Khalek M, El-Baz M and Ibrahiem EH: Predictors of prostatic cancer on extended biopsy in patients with high-grade prostatic intraepithelial neoplasia: a multivariate analysis model. BJU Int 2004; 94: 528. 2. Schoenfield L, Jones JS, Zippe CD et al: subsequent risk of cancer. BJU Int 2007; 99: 770.