maximal tumor diameter and the risk of psa failure in men with specimen-confined prostate cancer
TRANSCRIPT
OtMpaPfvRr�aandCts
Pcoretpa
FDMoMn
RFl
1
ADULT UROLOGY
1
MAXIMAL TUMOR DIAMETER AND THE RISKOF PSA FAILURE IN MEN WITH SPECIMEN-CONFINED
PROSTATE CANCER
TOMAS DVORAK, MING-HUI CHEN, ANDREW A. RENSHAW, MARIAN LOFFREDO,JEROME P. RICHIE, AND ANTHONY V. D’AMICO
ABSTRACTbjectives. To evaluate whether the maximal tumor diameter (MTD) is significantly associated with the time
o postoperative prostate-specific antigen (PSA) failure.ethods. Between 1986 and 2002, 781 men with clinical Stage T1c–T2 prostate cancer underwent radical
rostatectomy. The MTD was recorded as the maximal dimension of the largest single focus of cancer fromll 3-mm step sections. The median follow-up was 5.4 years (range 0.1 to 14.9); 242 men (31%) experiencedSA failure. A Cox regression analysis was used to determine the predictors of time to postoperative PSA
ailure. Kaplan-Meier estimates of PSA failure-free survival were made, dichotomized about the median MTDalue, and compared using a two-sided log-rank test.esults. The value of the MTD was significantly associated with the time to PSA failure (adjusted hazardatio 1.04, 95% confidence interval 1.01 to 1.07, P � 0.004), controlling for preoperative PSA level (P0.0001), prostatectomy Gleason score (P �0.0001), and T stage (P �0.0001). When margin status wasdded (P � 0.0004), the MTD approached statistical significance (P � 0.07). For patients with a preoper-tive PSA level of less than 10 ng/mL, prostatectomy Gleason score of 3 � 4 � 7 or less, Stage pT2–T3a, andegative margins, the value of the MTD significantly (P � 0.05) stratified the time to PSA failure, whenichotomized about the median value (13 mm), with 7-year PSA failure estimates of 17% versus 8%.onclusions. Whether patients with traditionally low-risk but large MTD prostate cancer fare better whenreated with adjuvant radiotherapy compared with salvage radiotherapy remains to be answered in theetting of a randomized trial. UROLOGY 66: 1024–1028, 2005. © 2005 Elsevier Inc.
lc(ssptw
asapstdmp
rostate-specific antigen (PSA) is commonlyused as a marker of treatment failure,1 be-
ause local or metastatic disease in the presencef an undetectable serum PSA level is rare afteradical prostatectomy (RP).2 Although RP is anffective primary treatment for localized pros-ate cancer, up to one third of patients will ex-erience biochemical recurrence within 10 yearsfter RP.3–5
rom the Departments of Radiation Oncology and Pathology, andivision of Urology, Brigham and Women’s Hospital, Boston,assachusetts; Departments of Pathology and Radiation Oncol-
gy, and Division of Urology, Harvard Medical School, Boston,assachusetts; and Department of Statistics, University of Con-
ecticut, Storrs, ConnecticutReprint requests: Anthony V. D’Amico, M.D., Department of
adiation Oncology, Brigham and Women’s Hospital, 75rancis Street, LL-2, Boston, MA 02115. E-mail: adamico@
roc.harvard.eduSubmitted: February 8, 2005, accepted (with revisions): May
s3, 2005
© 2005 ELSEVIER INC.024 ALL RIGHTS RESERVED
Both preoperative (age, ethnicity, serum PSAevel, PSA velocity, biopsy Gleason score, and per-entage of positive biopsies) and postoperativeprostatectomy Gleason score, prostatectomytage, tumor volume, perineural invasion, margintatus, lymph node involvement, and molecularrofiles) factors have been evaluated4–8 in an efforto identify patients at high risk of disease relapseho may benefit from adjuvant therapy.McNeal9 originally suggested that the biologic
ggressiveness of prostate cancer is related to itsize. However, several other publications10–12 havergued against total tumor volume having inde-endent prognostic significance in a multivariateetting. Stamey et al.13 pioneered the use of onlyhe largest tumor volume as an independent pre-ictor, by demonstrating that secondary tumorsay not contribute to the biologic behavior of the
rimary tumor.6 Similarly, the percentage of the RP
pecimen containing carcinoma, as proposed by0090-4295/05/$30.00doi:10.1016/j.urology.2005.05.037
Hi
tmfpsvfMltd
sti
P
Ttm(b
iiusceCdcAjpT
dtpvItpvbtgmeensmtCdog
maedscts
P
wtampfdP(s7e
S
mt
C
A
P
P
P
S
M
KD
U
umphrey and Vollmer,14 has been shown to havendependent prognostic value.
A potential issue with these techniques is thathey require submitting the entire prostate speci-en, which is very labor intensive and expensive
or routine use. As a result, Renshaw et al.15 pro-osed the maximal tumor diameter (MTD) as aimple and inexpensive proxy of the total tumorolume, and provided evidence that it may predictor PSA failure in routine clinical practice.16 The
TD is defined as the maximal dimension of theargest single focus of tumor from all prostatec-omy sections and is obtained by measuring thisistance directly from the glass slide with a ruler.In this study, we investigated whether the MTD is
ignificantly associated with the time to postopera-ive PSA failure to identify patients who would bedeally suited for randomized adjuvant therapy trials.
MATERIAL AND METHODS
ATIENT SELECTION, STAGING, AND TREATMENTBetween 1988 and 2002, 781 patients with clinical Stage
1c or T2 prostate cancer underwent radical retropubic pros-atectomy at Brigham and Women’s Hospital, with an MTDeasurement made by a single genitourinary pathologist
A.A.R.). Each patient signed an approved internal reviewoard consent form before study entry.In all patients, the evaluation of stage involved history tak-
ng and physical examination, including a digital rectal exam-nation, determination of the serum PSA level, and transrectalltrasound-guided needle biopsy of the prostate; the Gleasoncore was determined by histologic examination.17 The clini-al stage was obtained from the results of the digital rectalxamination using the 2002 American Joint Committee onancer staging system.18 Preoperative serum PSA levels wererawn within 1 month before surgery. The PSA levels wereommonly measured with assays from Hybritech, Tosoh, orbbott Laboratories. No patient received neoadjuvant or ad-
uvant therapy. The preoperative clinical and postoperativeathologic characteristics of the entire cohort are shown inable I.All patients underwent limited bilateral pelvic lymph node
issection during radical retropubic prostatectomy. The pros-atectomy specimens were submitted and processed using areviously documented submission protocol19 for seminalesicle invasion, extracapsular extension, and margin status.n brief, the apical and basal margins were amputated to ahickness of 5 mm and sectioned parasagittally in a directionerpendicular to the initial transverse incision at 3-mm inter-als. The base of the seminal vesicles was amputated and theasal cross-section submitted. The prostate was then sec-ioned perpendicularly to the long axis (apical to basal) of theland along the posterior rectal surface at 5-mm intervals, withost specimens requiring four to seven cross-sections to be
ntirely sectioned. For each cross-section, a single sectionach of the right and left posterior region was submitted. Fi-ally, a single section of the mid-anterior prostate was alsoubmitted. The correlation of this method with entirely sub-itted sectioning has been previously detailed.19 The prosta-
ectomy T stage was obtained using the 2002 American Jointommission on Cancer staging system.18 Specimen-confinedisease was defined as possible extracapsular extension with-ut seminal vesicle involvement, with negative surgical mar-
ins (Stage T2–T3a, with �SM). The MTD was recorded as the sROLOGY 66 (5), 2005
aximal dimension of the largest single focus of tumor fromll sections of the tumor and was determined by marking bothnds of the tumor with a pen and measuring this distanceirectly from the glass slide with a ruler.15 Tumor foci oneparate slides that may have represented cross-sections ofontiguous tumor were treated as separate foci. The prostatec-omy Gleason score17 was determined using a primary andecondary Gleason grade.
ATIENT FOLLOW-UPFollow-up started on the day of RP. All baseline PSA values
ere obtained within 1 month of the date of surgery. Patientsypically had a serum PSA measurement and digital rectal ex-mination every 3 months after surgery for 2 years, every 6onths for an additional 3 years, and annually thereafter. No
atient received adjuvant hormonal or radiotherapy (RT) be-ore PSA failure. PSA failure was defined as two consecutiveetectable PSA values after an undetectable value. The time ofSA failure was taken as the time of the first detectable valuegreater than 0.2 ng/mL), and time 0 was defined as the day ofurgery. The median follow-up for the entire study cohort of81 patients was 5.4 years (range 0.1 to 14.9); 242 (31%)xperienced PSA failure.
TATISTICAL ANALYSISCox regression multivariable analysis20 was used to deter-ine whether the preoperative serum PSA level, MTD, pros-
atectomy Gleason score, margin status, and T stage were as-
TABLE I. Preoperative clinical andpostoperative pathologic characteristics of
781 study patients
haracteristicPatients
(n)
ge (yr)�65 511 (65.4)�65 270 (34.6)
SA (ng/mL)�4 85 (10.9)4–10 471 (60.3)10–20 160 (20.5)�20 65 (8.3)
rostatectomy Gleason score�6 383 (49.0)3 � 4 � 7 224 (28.7)4 � 3 � 7 81 (10.4)�8 93 (11.9)
athologic stagepT2 535 (68.5)pT3a 147 (18.8)pT3b 99 (12.7)
urgical margin statusNegative 570 (73.0)Positive 211 (27.0)TD (mm)Median 13Q1 9Q3 18Maximum 29
EY: MTD � maximal tumor diameter; Q1 � first quartile; Q3 � third quartile.ata in parentheses are percentages.
ociated with the time to postoperative PSA failure. The MTD
1025
aGltcawaCaas
ftsrshmpc
t0((7ofu
td
TnM
if(loafstd
ltwvufndvbnpas
C
MPP
P
P
K*†
1
nd serum PSA level were treated as continuous variables. Theleason score was treated as a categorical variable using 6 or
ess as the baseline and 3 � 4 � 7, 4 � 3 � 7, and 8 or more ashe covariates tested. Prostatectomy T stage was evaluated as aategorical variable using T2 as the baseline, and T3a and T3bs the covariates tested. Patients with positive lymph nodesere excluded. Adjusted hazard ratios for PSA failure with
ssociated 95% confidence intervals were calculated using theox model. The assumptions of the Cox model were testednd met. Logistic regression analysis was used to test for anssociation between the largest value of the MTD and theurgical margin status.
For illustrative purposes, Kaplan-Meier estimates of PSAailure-free survival, dichotomized about the median value ofhe MTD, were compared using a two-sided log-rank test inpecific clinical subsets. Because prediction of PSA failure iseasonably successful4,6,8 in patients with accepted risk factorsuch as a high preoperative PSA level (greater than 10 ng/mL),igh Gleason score (greater than 3 � 4 � 7), positive surgicalargins, or lymph node invasion, this analysis focused on
atients deemed at “low risk” of distant failure by existingriteria.
RESULTS
A statistically significant association was found be-ween the MTD and the time to PSA failure (P �.004), controlling for preoperative PSA levelP �0.0001), prostatectomy Gleason score 3 � 4 � 7P � 0.0001), prostatectomy Gleason score 4 � 3 �(P �0.0001), prostatectomy Gleason score of 8
r greater (P �0.0001), Stage T3a (P � 0.2), andor Stage T3b (P �0.0001), as shown in Table IInder the heading Multivariate 1.However, when the margin status was added to
he model (P � 0.0004), the MTD approached but
TABLE II. Adjusted and unadjusted hazardfor preoperative and postoperative cova
postoperati
ovariate
Univariate
HR 95% CI P Value AH
TD (continuous) 1.13 1.10–1.16 �0.0001 1.SA (continuous) 1.04 1.04–1.05 �0.0001 1.rostatectomy
Gleason score�6 1.00 NA NA 1.3 � 4 � 7 2.96 2.10–4.16 �0.0001 2.4 � 3 � 7 6.16 4.11–9.23 �0.0001 2.�8 8.27 5.77–11.85 �0.0001 3.
rostatectomy Tstage
pT2 1.00 NA NA 1.pT3a 2.27 1.63–3.18 �0.0001 1.pT3b 9.83 7.23–13.37 �0.0001 3.
ositive surgicalmargin 3.34 2.59–4.31 �0.0001 N
EY: PSA � prostate-specific antigen; MTD � maximal tumor diameter; NA � notSurgical margin status not included in Cox model.Surgical margin status included in Cox model.
id not reach statistical significance (P � 0.07). a
026
his can be explained by the significant associationoted on logistic regression analysis between theTD value and margin status (P �0.0001).As shown in Figure 1, the MTD value was signif-
cantly (P � 0.05) associated with the time to PSAailure when dichotomized about the median value13 mm) in men with a preoperative PSA level ofess than 10 ng/mL, prostatectomy Gleason scoref 3 � 4 � 7 or less, pathologic Stage T2 or T3a,nd negative surgical margins. The estimate of PSAailure-free survival 7 years after RP was 92% ver-us 83% for men with an MTD value at or less thanhe median compared with greater than the me-ian, respectively.
COMMENT
Although RP is an effective primary treatment forocalized prostate cancer, up to one third of pa-ients will experience biochemical recurrenceithin 10 years after RP.3–5 The role of the tumorolume as a predictor of outcome after RP remainsnder study, partly because the most clinically ef-
ective and time-effective and cost-effective man-er in which to measure it has not been clearlyefined. Total tumor volume,9–12,21 largest tumorolume,6,13 and percentage of carcinoma14 have alleen evaluated. A potential downside of these tech-iques is that they require submitting the entirerostatectomy specimen, which is labor intensivend expensive for routine use.22 As a result, Ren-haw et al.15 introduced the concept of the MTD as
ios, 95% confidence intervals, and P valueses analyzed for association with time toSA failureMultivariate 1* Multivariate 2†
95% CI P Value AHR 95% CI P Value
1.01–1.07 0.004 1.03 0.99–1.05 0.071.02–1.03 �0.0001 1.02 1.01–1.03 �0.0001
NA NA 1.00 NA NA1.40–2.88 0.0001 1.96 1.37–2.81 0.00021.85–4.48 �0.0001 2.88 1.85–4.49 �0.00012.52–5.69 �0.0001 3.83 2.55–5.75 �0.0001
NA NA 1.00 NA NA0.87–1.82 0.22 1.28 0.88–1.84 0.22.59–5.42 �0.0001 3.15 2.14–4.63 �0.0001
NA NA 1.71 1.27–2.30 0.0004
able; HR � unadjusted hazard ratio; AHR � adjusted hazard ratio.
ratriatve P
R
0403
00018879
002675
A
applic
simple and inexpensive measure that can be eas-
UROLOGY 66 (5), 2005
iwfMhm
sipeMnsM
mTpp
tp3sfiMsRtcmMswb
stMgRz[cmccanspoMcsawd
wtciptsnpaarsatnib
Fdpc
U
ly used in routine clinical practice. This parameteras shown to be associated with PSA failure by two
ollow-up studies,16,23 but one study from theayo clinic12 did not find a correlation, which may
ave been related to the availability of whole-ounted RP specimens at their institution.In this study, we found that the MTD value was
ignificantly associated with the time to PSA failuren a model that included preoperative PSA level,rostatectomy Gleason score, and T stage. How-ver, when surgical margin status was added, theTD’s significance approached (P � 0.07) but did
ot reach statistical significance, likely because of aignificant correlation (P �0.0001) between theTD value and surgical margin status.Therefore, our results suggest that MTD valueay serve as a proxy for surgical margin status.his may be important in settings in which therostate specimen is not entirely submitted for
IGURE 1. PSA failure-free survival stratified by me-ian MTD value for patients with specimen-confined androstatectomy Gleason score 3 � 4 � 7 or less prostateancer.
athologic processing, as is the case in most prac- t
ROLOGY 66 (5), 2005
ices.24 The reported data suggest that althoughartial sampling is effective, it may miss 10% to0% of positive surgical margins.25–27 Moreover, ashown in Figure 1, patients with specimen-con-ned, low-grade prostate cancer, in which theTD value exceeded the median (13 mm), had a
ignificantly greater PSA failure rate by 7 years afterP (17% versus 8%). Essentially all the failures in
his patient subgroup will be local failures, likelyaused by microscopically occult positive surgicalargins missed by sampling. Therefore, using theTD value, which is a simple, rapid, and inexpen-
ive measure, may be useful in identifying patientsho have negative surgical margins on step sectionut who are at an increased risk of PSA failure.This finding raises an important clinical question:
hould patients with specimen-confined, prostatec-omy Gleason score 3 � 4 � 7 or less, and a larger
TD value (ie, greater than 13 mm), who are at areater risk of PSA failure, be treated with adjuvantT? A single randomized trial (European Organi-ation for Research and Treatment of CancerEORTC] 22911) compared the outcomes of radi-al prostatectomy with or without RT.28 After aedian follow-up of 5 years, both biochemical and
linical progression-free survival were signifi-antly better in the RP plus RT arm, with a reason-ble side-effect profile. Follow-up, however, hadot been sufficient to assess the impact on overallurvival. Thus, the definitive answer to whetheratients withrgan-confined, low-grade disease and a largeTD (who had a 17% failure rate at 7 years and
ould, similar to EORTC 22911 patients, be con-idered at an elevated risk of PSA failure) couldlso benefit similarly from adjuvant RT comparedith salvage RT after PSA failure requires a ran-omized clinical trial.Our study had some potential limitations thatarrant discussion. First, the percentage of posi-
ive biopsies has been shown to predict for prostateancer mortality in patients with low or favorablentermediate-risk disease after RT.29 Similarly, thereoperative PSA velocity has recently been showno be significantly associated with prostate cancer-pecific mortality.7 In particular, whether the sig-ificance of the MTD value remains when both theercentage of positive biopsies and PSA velocityre incorporated in the Cox model remains to benswered. Therefore, additional studies incorpo-ating the MTD value, positive percentage of biop-ies, and pretreatment PSA velocity are needed tossess the impact of each of these parameters onhe time to PSA failure. Finally, PSA failure doesot always translate to prostate cancer death. Only
n young and healthy men would such an endpointe expected to lead to prostate cancer-specific mor-
ality, given adequate time. To better understand1027
wtoaauMac
MgbslWtmMtPs
r8
me2
hp
iy
cmJ
fs1
Pr
bsU
cs
vrsl
vi
sst
dc
eg7
tc
dirp
al
cu
rm
B
op
mC
tdl
tP
ip
taJ
ysa
ppd
pcr
pt
1
hich men with PSA failure may experience pros-ate cancer-specific mortality, a PSA doubling timef less than 3 months has recently been suggesteds a surrogate for prostate cancer-specific mortalityfter postoperative or post-RT biochemical fail-re.30 Future studies will explore whether theTD value predicts for a short PSA doubling time
nd prostate cancer-specific mortality to assess itslinical significance further.
CONCLUSIONS
The results of our study have shown that theTD value can serve as a proxy for surgical mar-
in status in patients treated with RP. This maye important in settings in which the prostatepecimen is not entirely submitted for patho-ogic processing, as is the case in most practices.
hether patients with specimen-confined, pros-atectomy Gleason score of 3 � 4 � 7 or less,argin negative, and large, as measured by theTD, prostate cancer would fare better when
reated with adjuvant RT versus salvage RT atSA recurrence remains to be answered in theetting of a randomized trial.
REFERENCES1. Scher HI, Mazumdar M, and Kelly WK: Clinical trials in
elapsed prostate cancer: defining the target. J Natl Cancer Inst8: 1623–1634, 1996.
2. Leibman BD, Dillioglugil O, Wheeler TM, et al: Distantetastasis after radical prostatectomy in patients without an
levated serum prostate specific antigen level. Cancer 76:530–2534, 1995.
3. Pound CR, Partin AW, Eisenberger MA, et al: Naturalistory of progression after PSA elevation following radicalrostatectomy. JAMA 281: 1591–1597, 1999.
4. Ward JF, Blute ML, Slezak J, et al: The long-term clinicalmpact of biochemical recurrence of prostate cancer 5 or moreears after radical prostatectomy. J Urol 170: 1872–1876, 2003.
5. Stamey TA, Yemoto CM, McNeal JE, et al: Prostate can-er is highly predictable: a prognostic equation based on allorphological variables in radical prostatectomy specimens.Urol 163: 1155–1160, 2000.
6. Noguchi M, Stamey TA, McNeal JE, et al: Prognosticactors for multifocal prostate cancer in radical prostatectomypecimens: lack of significance of secondary cancers. J Urol70: 459–463, 2003.
7. D’Amico AV, Chen MH, Roehl KA, et al: PreoperativeSA velocity and the risk of death from prostate cancer afteradical prostatectomy. N Engl J Med 351: 125–135, 2004.
8. Khan MA, Partin AW, Mangold LA, et al: Probability ofiochemical recurrence by analysis of pathologic stage, Glea-on score, and margin status for localized prostate cancer.rology 62: 866–871, 2003.
9. McNeal JE: Cancer volume and site of origin of adeno-arcinoma in the prostate: relationship to local and distantpread. Hum Pathol 23: 258–266, 1992.
10. Epstein JI, Carmichael M, Partin AW, et al: Is tumorolume an independent predictor of progression followingadical prostatectomy? A multivariate analysis of 185 clinicaltage B adenocarcinomas of the prostate with 5 years of fol-
owup. J Urol 149: 1478–1481, 1993. 1028
11. Kikuchi E, Scardino PT, Wheeler TM, et al: Is tumorolume an independent prognostic factor in clinically local-zed prostate cancer? J Urol 172: 508–511, 2004.
12. Blute ML, Bergstralh EJ, Iocca A, et al: Use of Gleasoncore, prostate specific antigen, seminal vesicle and margintatus to predict biochemical failure after radical prostatec-omy. J Urol 165: 119–125, 2001.
13. Stamey TA, McNeal JE, Yemoto CM, et al: Biologicaleterminants of cancer progression in men with prostate can-er. JAMA 281: 1395–1400, 1999.
14. Humphrey PA, and Vollmer RT: Intraglandular tumorxtent and prognosis in prostatic carcinoma: application of arid method to prostatectomy specimens. Hum Pathol 21:99–804, 1990.15. Renshaw AA, Chang H, and D’Amico AV: Estimation of
umor volume in radical prostatectomy specimens in routinelinical practice. Am J Clin Pathol 107: 704–708, 1997.
16. Renshaw AA, Richie JP, Loughlin KR, et al: Maximumiameter of prostatic carcinoma is a simple, inexpensive, andndependent predictor of prostate-specific antigen failure inadical prostatectomy specimens: validation in a cohort of 434atients. Am J Clin Pathol 111: 641–644, 1999.17. Gleason DF, and Group at VACUR: Histologic grading
nd staging of prostatic carcinoma, in Tannenbaum M (Ed): Uro-ogic Pathology. Philadelphia, Lea & Febiger, 1977, pp 171–187.
18. Greene FL, for the American Joint Committee on Can-er and American Cancer Society: AJCC Cancer Staging Man-al. New York, Springer, 2002, pp 309–316.19. Renshaw AA: An abbreviated protocol for processing
adical prostatectomy specimens: analysis of grade, stage,argin status, and volume. J Urol Pathol 5: 183–192, 1996.20. Cox DR: Regression models and life tables. J R Stat Soc34: 187–199, 1972.21. Bostwick DG, Graham SD Jr, Napalkov P, et al: Staging
f early prostate cancer: a proposed tumor volume-basedrognostic index. Urology 41: 403–411, 1993.22. Noguchi M, Stamey TA, McNeal JE, et al: Assessment oforphometric measurements of prostate carcinoma volume.ancer 89: 1056–1064, 2000.23. Rashid M, Wojno KJ, Marcovich R, et al: Maximum
umor dimension provides a clinically useful and indepen-ently significant measure for predicting PSA-free survival fol-owing radical prostatectomy. J Urol 161(4S): 241, 1999.
24. True LD: Surgical pathology examination of the pros-ate gland: practice survey by American Society of Clinicalathologists. Am J Clin Pathol 102: 572–579, 1994.25. Hall GS, Kramer CE, and Epstein JI: Evaluation of rad-
cal prostatectomy specimens: a comparative analysis of sam-ling methods. Am J Surg Pathol 16: 315–324, 1992.26. Haggman M, Norberg M, de la Torre M, et al: Charac-
erization of localized prostatic cancer: distribution, gradingnd pT-staging in radical prostatectomy specimens. ScandUrol Nephrol 27: 7–13, 1993.27. Sehdev AE, Pan CC, and Epstein JI: Comparative anal-
sis of sampling methods for grossing radical prostatectomypecimens performed for nonpalpable (stage T1c) prostaticdenocarcinoma. Hum Pathol 32: 494–499, 2001.
28. Petrovich Z, Lieskovsky G, Langholz B, et al: Radicalrostatectomy and postoperative irradiation in patients withathological stage C (T3) carcinoma of the prostate. Int J Ra-iat Oncol Biol Phys 40: 139–147, 1998.29. D’Amico AV, Renshaw AA, Cote K, et al: Impact of the
ercentage of positive prostate cores on prostate cancer-spe-ific mortality for patients with low or favorable intermediate-isk disease. J Clin Oncol 22: 3726–3732, 2004.
30. D’Amico AV, Moul JW, Carroll PR, et al: Surrogate endoint for prostate cancer-specific mortality after radical pros-atectomy or radiation therapy. J Natl Cancer Inst 95: 1376–
383, 2003.UROLOGY 66 (5), 2005