Neoadjuvant Hormonal Therapy Prior to Radical Prostatectomy: Promises and Pitfalls

Home , Neoadjuvant therapy

Prostate Cancer and Prostatic Diseases (2000) 7, 136±144 ß 2000 Macmillan Publishers Ltd All rights reserved 1365±7852/00 $15.00 www.nature.com/pcan Review Neoadjuvant hormonal therapy prior to radical prostatectomy: promises and pitfalls

ME Gleave1*, S La Bianca1 & SL Goldenberg1 1Division of Urology, University of British Columbia, Vancouver General Hospital, D-9, 2733 Heather Street, Vancouver, BC, Canada

Prostate Cancer and Prostatic Diseases (2000) 3, 136±144.

Keywords: prostate cancer; radical prostectomy; neoadjuvant; hormonal therapy

Introduction the incidence of positive surgical margins and disease recurrence. The goal of radical prostatectomy is complete removal of The term adjuvant therapy refers to the use of systemic all cancer cells. Unfortunately, even in carefully selected therapy after management of localized disease when risk patients, up to two-thirds of clinically con®ned tumors of recurrence is known to be high. Chemotherapy regi- are understaged, and positive margin rates of 30 ± 60% are mens can cure some patients when used adjuvantly even reported following radical prostatectomy.1±3 Although though the same regimen fails to alter the natural history higher tumor stage, grade and pre-operative serum pros- of established metastatic disease. Experimental animal tate speci®c antigen (PSA) levels correlate with higher model and early clinical trial data suggest that adminis- pathologic stage and risk of disease recurrence, prognos- tration of systemic therapy is more likely to affect cure tic variables are limited in their ability to independently when tumor burden is low.10 Neoadjuvant therapy and individually predict pathologic stage and risk of extends this logic further by applying systemic therapy recurrence.4±6 In order to optimize outcomes from radical earlier in the course of the disease prior to de®nitive prostatectomy, only men with low-risk tumors (T1 or T2a, locoregional therapy. The aim of neoadjuvant hormonal Gleason score < 6, serum PSA < 10 mg/l) have tradition- therapy (NHT) prior to radical prostatectomy is reduction ally been considered ideal for cure by surgical interven- of positive margin rates and ultimately decreased disease tion. Thus, ideal candidates for radical prostatectomy are recurrence. The role of NHT prior to radical prostatec- men with the highest probability of being cured by the tomy is controversial as some clinicians argue that down- operation. Indeed, cure rates following radical prostatec- sizing occurs without down staging and any apparent tomy alone exceed 80% in men with low-risk tumors.7,8 down staging results from dif®culty in pathologic evalua- However, radical therapy may represent overtreatment in tion of the neoadjuvantly treated prostatectomy speci- some of these men because of the indolent natural history men.11,12 Others further argue that short-term follow-up of low-risk disease.9 In contrast, most men with high-risk from randomized studies comparing 3 months of NHT to localized tumors (> T2b, Gleason score 7, PSA > 10 mg/l) surgery alone show no difference in biochemical recur- and life expectancies > 10 y will die of their disease if they rence rates.13 However, NHT has many attractive theore- are not treated with curative intent.9 High-risk tumors tical features and research in this are is justi®ed and exhibit a more aggressive natural history and have important. The purpose of this article is to review the higher positive margin and recurrence rates after radical rationale behind NHT, critique the randomized studies of prostatectomy or radiotherapy alone, where unimodality 3 months of therapy, and examine the optimal duration therapy probably represents undertreatment.4±9 Para- of therapy. doxically, the therapeutic ratio (patients who actually realize a survival bene®t from a therapeutic intervention) may therefore be greater in high-risk disease if its natural history can be altered by multimodality therapy. It is important, therefore, to investigate therapies that opti- Rationale behind neoadjuvant therapy mize complete extirpation of all cancer cells and reduce prior to radical prostatectomy High positive margin rates following radical *Correspondence: M Gleave, D-9, 2733 Heather Street, Vancouver prostatectomy General Hospital, Vancouver, BC, Canada V5Z 3J5. Although higher tumor stage, grade and pre-operative Received 6 March 2000; accepted 31 May 2000 serum PSA levels correlate with higher pathologic stage, Neoadjuvant hormonal therapy ME Gleave et al no prognostic variable exists that can be applied inde- suicide' process that requires activation of a series of 137 pendently and individually to predict pathologic stage.4 genes and is characterized morphologically by shrunken Up to two-thirds of clinically con®ned tumors are under- cells with condensed and fragmented nuclei (apoptotic staged, and positive margin rates of 30 ± 60% are reported bodies). Initiation of apoptosis may, in part, be cell-cycle following radical prostatectomy.1±3 Incomplete excision speci®c or regulated by epigenetic factors and may there- places the patient at higher risk of disease recurrence.5±8 fore begin at different times in different populations of Selection of patients who are most likely to bene®t from cells. Maximal apoptotic regression of ventral rat prostate NHT remains critical; NHT is less likely to alter outcome glands and Shionogi tumors is complete within 1 ± 2 in most low-risk tumors (PSA < 10 mg/l, Gleason score weeks following castration;19 however, timing of apopto- 6, stage T1C, since most of these patients do well with sis has not been well characterized in benign or malignant surgery alone. In contrast, men with high-risk localized human prostate tissue and probably takes place over a tumors (PSA > 10 mg/l, Gleason grade > 7) are at signi®- longer period of time. cant risk of positive margins and PSA recurrence. Patients with positive margins are often offered post-operative radiation therapy to consolidate local therapy, which is Availability of PSA as a marker of tumor regression associated with increased risk of urinary incontinence, importance and anastomotic strictures. It is therefore PSA gene expression is androgen-regulated and serum important to investigate combinations of therapies that PSA levels are dependent on both androgen levels and optimize complete extirpation of all cancer cells and tumor volume.20,21 After institution of androgen ablation reduce the incidence of positive surgical margins. therapy, serum PSA levels decrease rapidly and dramati- cally due to cessation of androgen-regulated PSA gene expression and apoptosis. Although signi®cant downsiz- Reversible androgen suppression therapy ing occurs after 3 months of therapy, serum PSA does not reach undetectable levels after 3 months in most patients, The development of potent, well-tolerated and reversible which suggests that optimal duration of NHT may be agents for androgen ablation therapy provides a safe longer than 3 months. Maximal tumor regression prob- method for inducing prostate cancer cell death and ably occurs when PSA reaches its nadir level. Changes in tumor regression prior to radical prostatectomy. Several serum PSA following NHT provide objective biochemical classes of drugs induce castrate levels of testosterone information that re¯ects tumor response and may identify through suppression of LH release from the pituitary patients not responding favorably. gland, including diethylstilbestrol (DES)14 luteinizing- hormone-releasing hormone (LHRH) agonists include 15,16 goserelin, leuprolide and buserelin. The ¯are reaction Animal model studies can be prevented by cyproterone acetate or DES 1 week prior to the LHRH analog, or blocked by nonsteroidal Androgen withdrawal in the Shionogi tumor model pre- anti-androgens like ¯utamide or bicalutamide. cipitates apoptosis and tumor regression in a highly Prostate tumor cell death, not just tumor shrinkage, reproducible manner with up to a 2-log cell kill, but occurs following medical castration. Apoptosis, or pro- despite undergoing complete regression, androgen-inde- grammed cell death, occurs in normal, benign hyper- pendent tumors recur after one month.19 Conceptually, plastic and malignant prostatic epithelial cells by any the ideal time to implement curative therapy is after procedure that results in castrate levels of testoster- maximal castration-induced tumor involution but before one.17 ± 19 Apoptotic cell death appears to be a `genetic outgrowth of androgen-resistant clones. Experiments

Figure 1 Neoadjuvant castration reduces local recurrence rates in the Shionogi tumor model. Tumors were allowed to grow to 1 ± 2 g before mice were randomized to either tumor excision with wide margins and castration upon tumor recurrence (group 1), or neoadjuvant castration for 10 days followed by wide excision of the regressed tumor (group 2). The study endpoint was androgen-independent tumor recurrence which occurred in 80% of Group 1 after a median of 36 days and 44% of mice in Group 2 after a median of 42 days.

Prostate Cancer and Prostatic Diseases Neoadjuvant hormonal therapy ME Gleave et al

138 designed to determine whether local recurrence rates after historical controls, but no postoperative follow-up data tumor excision were lower with neoadjuvant compared to has been reported. Investigators at Memorial Sloan Ket- adjuvant androgen ablation demonstrated that tumor-free tering Cancer Center have conducted a series of studies survival was signi®cantly greater in the NHT group (56% beginning with a report by Aprikian et al 24 which sug- 22 vs 20%, P < 0.05; Figure 1). Although this model can not gested that most patients with locally advanced cT3 address the issue of the affect of NHT on subclinical tumors did not bene®t from between 2 and 8 months of metastases, extrapolation of these results to the clinical neoadjuvant DES. A subsequent phase II study of 141 disease suggests that NHT may help to reduce not only patients with cT1 or cT2 tumors reported signi®cant the positive margin rates but also the subsequent risk of decreases in serum PSA (> 98%) and a 38% decrease in local recurrence following radical prostatectomy. prostatic volume after 3 months of LHRH agonist with ¯utamide.25 Pathologic staging revealed increased organ- con®ned rates (74% vs 49%) and decreased positive margin rates (10% vs 33%) when compared to a group Clinical studies with short-term of 72 concurrent controls who did not receive NHT. However at 48 months follow-up, no differences in PSA neoadjuvant therapy recurrence rates were apparent.26 Several smaller nonrandomized studies by Sassine and Schulman27,28 and by Most studies assessing the effects of NHT prior to surgery HaÈggman et al 29 also reported 40 ± 50% decreases in chose, somewhat empirically, a 3 month duration. Surro- positive margin rates when groups of patients pretreated gate endpoints for assessing the ef®cacy of NHT prior to with 3 months of NHT were compared to historical radical prostatectomy include serum PSA nadir levels controls. In contrast, reports on small cohorts of patients prior to surgery, pathologic stage and positive margin with cT or cT tumors by Mcfarlane et al 11 and Oesterling rates, and most importantly, serum PSA recurrence rates 2 3 et al 12 suggested no improvement in pathologic outcome following surgery. following short-term NHT. These nonrandomized studies (reviewed in Table 1) illustrate that outcome (pathologic Nonrandomized clinical studies stage, positive margin rates, and risk of PSA recurrence) is dependent on many factors, including proportion of T3 Most nonrandomized studies report decreases in serum tumors and other high-risk features, length of postopera- PSA, prostatic volume and positive margin rates follow- tive follow-up, and time period during which the study ing 3 months of NHT (Table 1). These studies do not accrued (stage migration towards earlier diagnosis began report on time to PSA nadir or absolute PSA nadir levels. in early 1990s). Conclusions from such studies must be Soloman et al 23 reported on the largest phase II series of drawn with caution, but taken together the results sug- 156 patients with T2/T3 tumors treated with between gest that 3 months on NHT result in encouraging patho- 3 and 6 months of an LHRH agonist plus ¯utamide. logic effects and lead to initiation of Phase III randomized Positive margin rates were 11.5%, compared to 35% in studies.

Table 1 Nonrandomized studies of short-term neoadjuvant hormone therapy

Investigator Sample size Clinical stage Serum PSAa Duration of NHT Positive margin rate Follow-up

23 Soloman, 1993 156 T2/T3 NA 3 ± 6 months 11.5% NA 24 Aprikian, 1994 55 T3/T2b 20.4 (median) 3 (median) 33% 62%/26 months 29 HaÈggman, 1993 40 T2/T3 NA 3 months 31% 15%/3 months 28 Schulman, 1993 40 T2/T3 NA 3 (mean) 32% NA 25,26 Fair, 1993, 1997 74 T2 8.7 (median) 3 months 10% 18%/48 months 11 Mcfarlane, 1993 22 T3/T2b 14.8 3 months 32% NA 12 Oesterling, 1993 22 T3/T2b 30 1.5 months 68% NA

aPretreatment serum PSA (mg/l).

Table 2 Randomized studies of short-term neoadjuvant hormone therapy

Investigator Sample size Clinical stage Type of NHT Change in serum PSA Change in TRUS volume Positive margin rate

30 a b Labrie, 1995 161 T2/T3 3 months L FNANA8%vs 34% 31 3 b Soloway, 1995 303 T2b 3 months L F 14.3 ± < 0.5 in 70 44 ± 35 cm 18% vs 48% 32 3 c Van Poppel, 1995 130 T2b/T3 6 weeks E.P. 14 ± 1.0 mg/l 43 ± 29 cm 20% vs 46% 33 3 b Goldenberg, 1996 213 T1/T2 3 months CPA 13 ± 1.1 mg/l 43 ± 33 cm 28% vs 65% 34 b Fair, 1997 148 T1/T2 3 months G F NA NA 18% vs 37% 35 3 b Witjes, 1997 354 T2/T3 3 months G F 20 ± 0.8mg/l 38 ± 27 cm 27% vs 46% 36 b Hugosson, 1996 111 T1 ± T3a 3 months T C NA NA 23% vs 41%

a30 patients refused randomization and were included in the NHT arm. bStatistically signi®cant difference in primary endpoint. c Statistically signi®cant difference only in T2 patients. L F leuprolide acetate plus ¯utamide; G F goserelin plus ¯utamide; CPA 300 mg cyproterone acetate daily; E.P. estramustine phosphate; T triptorelin plus cyproterone acetate 50 mg bid.

Prostate Cancer and Prostatic Diseases Neoadjuvant hormonal therapy ME Gleave et al 139 Randomized clinical studies power necessary to detect signi®cant differences in biochemical recurrence rates. Second, follow-up is short and To date, seven Phase III randomized studies comparing 0 any bene®t from NHT may only become apparent after vs 3 months of NHT have been published (Table 2).30 ± 36 longer follow-up. Patterns of failure following radical The primary endpoint in all these trials was incidence of prostatectomy suggest that early failures (ie < 2 y) are positive margin disease. All seven of these studies more often distant, while most local failures (ie those demonstrate signi®cant decreases in positive surgical most likely to be affected by short-term NHT), do not margins after 3 months of NHT, reporting decreases become manifest until 3 or more years after surgery. ranging from 34 ± 65% to 8 ± 28%. These studies do not Third, approximately half the patients in these studies report on time to PSA nadir or absolute PSA nadir levels. had low-risk tumors (Gleason 6, PSA 10 mg/l) with a Soloway et al 31 published the results of a multicenter very low risk of PSA recurrence after radical prostatec- randomized study comparing 0 vs 3 months of leuprolide tomy (10 ± 20%), which could dilute any potential bene®t and ¯utamide prior to radical prostatectomy. Positive from NHT. Finally, there may be a true lack of bene®t margin rates decreased by over 50% from 42% in the from 3 months of NHT, because of an error either in control group to 17% in the NHT group. Organ-con®ned application (ie the duration of NHT was too short), or in disease (pT2) was reported twice as frequently in the NHT concept (ie no matter how NHT is applied, it will not alter group (53%) compared to the surgery alone group (22%). recurrence rates). Similarly, a randomized multicenter trial by the Canadian Androgen ablation therapy is rarely curative in Uro-Oncology Group (CUOG) using 3 months of cypro- patients with osseous metastases; however 10% of terone acetate also reported a 50% reduction in positive patients with D2 disease do survive for longer than 10 y margin rates (65% vs 28%) and doubling of the organ- after castration. Clinical observations further suggest that con®ned rate (20% vs 42%).33 Fair and Kava34 reported soft tissue disease responds more favorably to androgen results of a recent randomized trial comparing 0 vs 3 ablation with more frequent and durable complete months of NHT (LHRH plus ¯utamide) in 148 patients responses to compared to osseous metastases.39,40 with clinically con®ned prostate cancer. Overall, the inci- Patients with stage D1 lymph node-only disease have a dence of positive margins decreased from 37% to 18% median survival more than twice that of patients with D2 (P < 0.05). The decrease in positive margins was even osseous disease.40 Adjuvant androgen ablation therapy greater in high-risk patients with pre-treatment PSA following radical prostatectomy in patients with levels > 10 mg/l (63% vs 18%), suggesting that these regional lymph node metastases results in long-term men may be the most likely to bene®t from NHT. In the (15 y) survival in 30%, which suggests that a signi®cant Belgian study, no difference in positive margin rates was proportion of patients with low volume soft tissue metas- observed using only 6 weeks of estramustine phosphate tases may be cured with surgery and immediate andro- when all patients were considered (40.5 vs 46.5%); how- gen ablation.41 One would anticipate that a similar ever, when only T2 patients were considered, there was a proportion of patients with clinically localized tumors statistically signi®cant decrease in positive margin and unrecognized subclinical metastases who would disease.32 ultimately recur after radical prostatectomy alone may The randomized studies reviewed using 3 months of also be cured by NHT as long as the course of therapy is NHT in Table 2 consistently report 90% decreases in long enough. serum PSA levels, 30% decreases in prostatic volume, 50% decreases in positive margin rates in cT1 or cT2 (but not cT3) disease, higher organ-con®ned rates, and no differences in incidence of seminal vesicle involvement Clinical studies with long-term or positive lymph nodes. Caution is required, however, when interpreting stu- neoadjuvant therapy dies using pathologic stage as an endpoint despite the The timing (in terms of initiation at the molecular/cellular apparently encouraging results reviewed above. The abi- level and completion at the tumor population level) of lity of NHT to reduce biochemical and local recurrence castration-induced apoptosis of prostate carcinoma cells remains the ultimate goal and this has not yet been remains poorly de®ned, but it is likely that castration- determined. Soloway et al 37 recently reported 24 month induced apoptosis is an asynchronous event initiated at follow-up data on 256 evaluable patients in the US variable times in a heterogenous tumor population. Intergroup Study and found that 21% of the NHT Timing of apoptosis is dif®cult to characterize in human group compared to 21.6% of the surgery alone group prostatic tumors because of the absense of appropriate of had PSA levels > 0.4 mg/l (P 0.473). Similarly, the animal model systems, and because serial or comparative CUOG study has not demonstrated any differences in immunostaining of biopsies for apoptotic bodies incon- PSA recurrence rates at 36 months follow-up, with recur- sistently detects uncommon events that exist for relatively rence rates of 35% and 40% in the control and NHT arms, short periods of time. Although downsizing occurs after 3 respectively.38 These two studies suggest that 3 months of months of NHT, serum PSA has not reached undetectable NHT does not signi®cantly reduce the risk of treatment levels in most patients after this period of time. Ongoing failure. There are, however, several possible reasons for decreases in serum PSA after the second month of ther- lack of apparent difference in PSA recurrence rates apy are not due to further decreases in PSA gene expres- despite favorable changes in positive margin rates. First, sion because changes in gene expression occur rapidly the sample size of these studies was based on detecting and reach basal levels shortly after testosterone reaches differences in positive margin rates, and therefore the castrate levels.42 Therefore, decreases in serum PSA after studies were not large enough to have the statistical the second month possibly re¯ect an ongoing imbalance

Prostate Cancer and Prostatic Diseases Neoadjuvant hormonal therapy ME Gleave et al 140

Figure 2 Changes in serum PSA levels in men with clinically localized prostate cancer after initiation of 8 months of androgen withdrawal therapy.

between apoptotic prostate epithelial cell death and with negative margins (pT3M-) was reported in 13% of decreased tumor cell proliferation, which leads to further the patients. Pathologic stage and risk of positive margins reduction in tumor volume. These observations suggest increased with higher clinical T-stage, grade and pre- that optimal duration of NHT may be longer than 3 treatment serum PSA levels (Table 3). None of 65 patients months. with T1c or T2a tumors had positive margins compared to A prospective trial was initiated to study the biochem- 7.5% of 79 with T2b and 25% of 12 patients with T3a ical and pathologic effects of 8 months of NHT prior to disease. Positive margins were identi®ed in 2% of patients surgery in 156 men.43,44 Different rates of PSA decline with Gleason scores 4, 5% with scores of 5 or 6, and 11% after institution of therapy produces two distinct slopes a with Gleason scores of > 7. None of 99 patients with PSA precipitous drop during the ®rst month and a more levels < 10 mg/l had positive margins, compared to 15% gradual subsequent decrease which continued into the of 57 with PSA levels 10mg/l. Although patients with eighth month of therapy (Figure 2). The initial rapid 80 ± organ-con®ned tumors tended to have lower mean PSA 90% decrease in PSA during the ®rst month results from cessation of androgen-regulated PSA synthesis and apoptosis, while the ongoing slower 50% decline may parallel decreasing number of PSA-producing prostatic cells. Serum PSA decreases to undetectable levels or reaches nadir levels in 22% of patients after 3 months, 42% after 5 months and 84% after 8 months. After 8 months of NHT, serum PSA levels were 0.5 mg/l in 18% of patients 0.4 mg/l in 82% of patients, < 0.3 mg/l in 76% of patients, and 0.2mg/l in 66%. Most patients in this study had organ-con®ned, stage pT2 tumors (79%), and half of these had microfoci or small volume tumors occupying less than 5% of the prostate (Table 3). The positive margin rate was 5.8%, lower than that reported after 3 months of NHT in similar 34,36,37 cohorts of patients (Figure 3). No cancer was iden- Figure 3 Pathologic stage in patients treated with no NHT or 3 months of ti®ed in the radical prostatectomy specimen in 13% of NHT in the US Intergroup31 and CUOG33 studies, compared to patients cases. Microscopic extension through prostatic capsule treated with 8 months of NHT in the Vancouver43,44 study.

Table 3 Effects of pretreatment PSA and grade on pathologic stage after 8 months of NHT

Serum PSA (mg/l) (%) Gleason gradea (%)

Pathologic stage Overall (%) < 4 4 ± 9.9 10 ± 20 > 20 45,6 7

P0 20 (13) 4 (24) 11 (13) 4 (11) 1 (5) 8 (18) 11 (16) 1 (3) P2a 53 (34) 7 (41) 30 (37) 11 (29) 5 (26) 18 (40) 21 (28) 14 (39) P2b 50 (32) 6 (35) 26 (32) 13 (34) 5 (26) 13 (29) 26 (35) 11 (30) P3 M 7 21 (13) 0 14 (17) 3 (8) 4 (22) 5 (11) 11 (15) 5 (14) P3 M 9 (6) 0 0 6 (15) 3 (16) 1 (2) 4 (5) 4 (9) P 6 N 3 (2) 0 1 (1) 1 (3) 1 (5) 0 1 (1) 2 (5)

aPre-treatment.

Prostate Cancer and Prostatic Diseases Neoadjuvant hormonal therapy ME Gleave et al 141

Figure 4 Kaplan ± Meier survival curves for 156 patients treated with 8 months of NHTand radical prostatectomy, with mean follow-up post-surgery of 54 months. (A) overall, compared to CUOG 0 vs 3 month NHT study33 and Cleveland Clinic series;7 (B) strati®ed by pathologic stage. nadir levels, the nadir PSA value was not predictive of of Medicine8 (Table 5). Differences in PSA recurrence rates pathologic stage in individual patients. are most pronounced when only patients with high-risk Biochemical recurrences occurred in 19 patients features are considered (Table 5). (12.2%), most within the ®rst 2 y of surgery after a mean Meyer et al 46 recently reviewed 680 men followed for 52 months postoperative follow-up (Figure 4A). Risk of 38 months after radical prostatectomy and reported an biochemical recurrence after 8 months of NHT remains overall PSA failure rate of 33% at 5 y. Interestingly, proportional to known risk factors for recurrence, includ- patients who received neoadjuvant androgen ablation ing pretreatment serum PSA and Gleason score, and for longer than 3 months had a signi®cantly lower risk pathologic stage (Table 4). The PSA recurrence rate in of PSA failure than those treated by radical prostatectomy this Phase II study using 8 months of NHT compares alone (hazard ratio 0.52). favorably with recurrence rates reported by the US Inter- Apparent improvement in recurrence rates between group,40 the Canadian Uro-Oncology Group,41 the Cleve- uncontrolled studies from different centers must, how- land Clinic, Washington University,45 and Baylor College ever, be interpreted with caution. Any potential bene®t of

Table 4 Analysis of treatment outcome by preoperative risk factor after 8 months NHT and radical prostatectomy

Any single high risk factor

Total (%) No risk factors cT3 Gleason 7 PSA 10 PSA > 20

No. patients 156 81 12 37 57 19 Mean PSA (mg/l) 10.9 6.1 17.2 14 19.2 29.2 Mean F/U (months) 52.5 53.5 44.5 50.7 52.6 54.3 pT3 M (%) 9 (5.8) 0 3 (25) 4 (10.8) 9 (16) 3 (16) PSA recurrence (%) 19 (12.2) 5 (6.2) 3 (25) 8 (22) 12 (21) 3 (16)

Prostate Cancer and Prostatic Diseases Neoadjuvant hormonal therapy ME Gleave et al

142 Table 5 Comparative biochemical progression-free survival after radical retropubic prostatectomy

VGH CUOG41 Soloway37 Cleveland7 Baylor8 Catalona45 Institution RP and 8 NHT44 RP and NHT RP and NHT RP only RP only RP only

Overall 87% 60% 79% 61% 78% 74% PSA-free survival (5 y) (3 y) (2 y) (5 y) (5 y) (7 y) Gleason 4 93% NA NA 92% 92% 84% 5 or 6 89% 68% NA 77% 83% 68% 7 79% 10%a NA 48% 62% 48%a PSA 10 78% 50% NA 56% 69% 49% > 20 81% 36% NA 26% 44% NA De®nition of PSA PSA PSA PSA PSA Not recurrence > 0.3 > 0.2 > 0.4 > 0.2 > 0.4 de®ned

aGleason 7 tumors included in 5 or 6 category.

longer-term NHT is best con®rmed by Phase III studies, keratin immunostaining to help identify foci of cancer one of which has been recently initiated by the Canadian dif®cult to see with H&E prepared sections did not Uro-Oncology Group. CUOG-P95A randomized 547 men signi®cantly increase positive margin rates.43,48 Third, with clinically con®ned prostate cancer to either 3 or 8 animal model studies using androgen-dependent Shio- months of leuprolide acetate and ¯utamide prior to nogi tumors demonstrate that NHT reduces local recur- radical prostatectomy.47 The study was powered to rence and positive margin rates by 50%.22 Finally, detect a 25% decrease in PSA recurrence, assuming a pathologic stage remains proportional to pre-treatment 30% recurrence in the 3 month arm after 3 y. Patients risk factors (serum PSA, T-stage and Gleason grade), and were equally strati®ed for clinical stage, Gleason grade an important prognostic predictor for biochemical and serum PSA. Interim analysis of biochemical and relapse33,36,44 (Figure 4B). Taken together, these observa- pathologic differences between the 3 and 8 month tions suggest that the changes in pathologic stage induced groups demonstrated signi®cant differences. Mean by NHT are real and not artifactual. serum PSA decreased 89% to 0.12 mg/l after 3 months, with a further 53% to 0.056 mg/l from 3 to 8 months. Pre- surgery PSA nadir levels were < 0.1 mg/l in 35% vs 73%, Is surgery more dif®cult following neoadjuvant and 0.3 mg/l in 37% vs 10% after 3 vs 8 months, therapy? respectively. TRUS-determined prostatic volume decreased 37% from a mean of 40.5 cm3 to 25.7 cm3 after Concerns have been raised by some urologists that NHT 3 months of NHT (P 0.0001) and a further 11% to makes surgery more dif®cult by increasing periprostatic 22.8 cm3 after 8 months (P 0.03). Positive margin rates adhesions. However, this does not appear to be the case. were 23% and 12% in the 3 and 8 month groups, respec- Randomized series that have documented surgical dif®- tively (P < 0.01). This Phase III study con®rms that there culty, blood loss, length of procedure, and postoperative is on-going biochemical and pathological regression of complications found no signi®cant difference in these parameters between NHT-treated and untreated prostate tumors between 3 and 8 months of NHT, sug- 31,33 gesting the optimal duration of NHT is longer than 3 groups. Apoptosis initiated by androgen withdrawal months. Determination of whether longer therapy alters (or other stimuli elsewhere throughout the organism) is a PSA recurrence rates must await longer follow-up. controlled cellular event that is not associated with acute in¯ammation or scarring. However, one technical factor after 8 months of NHT is that the prostate can decrease to a small, fusiform swelling of the urethra which can make Potential pitfalls of neoadjuvant therapy it more dif®cult to palpate the prostatic apex and decide where to take the dorsal bundle. Does NHT result in artifactual pathological understaging? Does tumor progression occur during prolonged Following NHT, residual foci of atrophic glands can be NHT? dif®cult to identify with H&E staining, raising the possi- bility that pT-0 staging or low positive margin rates may Because of possible risk of progression and outgrowth of be an artifact resulting from pathological understaging. androgen-independent clones during prolonged NHT, Some authors argue that downsizing occurs without changes in levels of the proliferation marker Ki-67 and downstaging and any apparent downstaging results the oncoprotein Bcl-2 have been assessed.43,49 Ki-67 stain- from dif®culty in pathologic evaluation of the NHT- ing remained suppressed in most patients after 8 months treated prostatectomy specimen.11 ± 13 Evidence thus far, of NHT.43,49 Similarly, staining for the cyclin-dependent however, overwhelming suggests that the NHT-induced kinase inhibitor, p27, increased in prostate cancer speci- decrease in positive margin rates is a real, rather than an mens after NHT ranging between 3 and 8 months.50 These artifactual, phenomenon. First, all randomized studies observations suggest cell cycle arrest is induced by andro- using a single central and experienced pathologist gen withdrawal in most residual cancer cells surviving report similar (50%) decreases in positive margin rates NHT. Levels of the antiapoptotic oncoprotein, Bcl-2, (see Table 2). Second, reevaluation of prostatectomy speci- increase during the ®rst 3 months of NHT, and remain mens using prostatic acid phosphatase (PAP) and cyto- elevated after 8 months of therapy,49 consistent with its

Prostate Cancer and Prostatic Diseases Neoadjuvant hormonal therapy ME Gleave et al role in prevention of castration-induced apoptosis.51,52 9 Chodak GW et al. Results of conservative management of clini- 143 Taken together, ongoing decreases in serum PSA, the cally localized prostate cancer. New Engl J Med 1994; 330: 242 ± high percentage of organ-con®ned and small-volume 248. 10 Skipper HE, Schabel FM Jr, Wilcox WS. Experimental evaluation tumors, and the absence of increased PCNA and Ki-67 of potential anticancer agents. XII: On the criteria and kinetics immunostaining suggest that progression from out- associated with `curability' of experimental leukemia. Cancer growth of androgen-independent clones during pro- Chemother Rep 1964; 35:1±111. longed NHT is unlikely. 11 Macfarlane MT et al. Neoadjuvant hormonal deprivation inpati- ents with locally advanced prostate cancer. J Urol 1993; 150: 132. 12 Oesterling JE et al. Preoperative androgen deprivation therapy: arti®cial lowering of serum prostate speci®c antigen without Will recurrence rates be altered by neoadjuvant downstaging the tumor. JUrol1992; 149: 779. therapy? 13 Paulson, DF. Neoadjuvant androgen deprivation therapy prior to radical prostatectomy. Con. Urology 1996; 48: 539 ± 540. With the exception of testis, low ano-rectal, and Wilm's 14 Cox RL, Crawford ED. Estrogens in the treatment of prostate tumors, most data do not support use of neoadjuvant cancer. J Urol 1995; 154: 1991 ± 1998. therapies for most solid tumors when survival outcomes 15 The Leuprolide Study Group. Leuprolide versus diethylstilbes- terol for metastatic prostate cancer. New Engl J Med 1984; 311: are compared to local therapy alone or followed by 1281 ± 1286. adjuvant therapy. Indeed, it remains possible that delay 16 Mahler C. Is disease ¯are a problem? Cancer 1993; 72: 3799 ± 3802. in de®nitive local therapy by neoadjuvant treatments may 17 Bruchovsky, N. The metabolism of testosterone and dihydrotes- be detrimental. Lack of bene®t of neoadjuvant chemother- tosterone in an androgen-dependent tumor. A possible correla- apy in other solid tumors, such as cervix,53 small cell tion between dihydrotestosterone and tumor growth in vivo. lung54 and bladder,55 illustrates that use of therapies Biochem J 1972; 127: 561. 18 Tenniswood M. Apoptosis, tumor invasion and prostate cancer. capable of inducing complete responses may not have Br J Urol 1997; 79: 27 ± 34. positive effects on survival when use in a neoadjuvant 19 Bruchovsky N et al. Effects of androgen withdrawl on the stem setting. Determining whether 3 months of NHT reduces cell population of the Shionogi carcinoma. Cancer Res 1990; 50: recurrences after radical prostatectomy will have to await 2275. maturation of the recently completed Phase III 20 Gleave ME et al. Serum PSA levels in mice bearing human studies30 ± 36 (Table 2). However, it is critical to realize prostate LNCaP tumors are determined by tumor volume and that the sample size of these studies was based on endocrine and growth factors. Cancer Res 1992; 50: 1598. 21 Stamey TA, Kabalin JN, Ferrari M, Yang N. Prostate speci®c detecting differences in positive margin rates, and lacks antigen in the diagnosis and treatment of adenocarcinoma of the suf®cient statistical power to detect smaller differences in prostate. IV. Anti-androgen treated patients. J Urol 1989; 141: biochemical recurrence rates. Whether longer-term NHT 1088. is any better is being evaluated by Phase III studies 22 Gleave M et al. Neoadjuvant androgen ablation reduces local through CUOG. Until recurrence rates have been demon- recurrence rates after tumour excision in the Shionogi tumour strated to decrease in randomized studies, NHT should be model. J Urol 1997; 157: 1727 ± 1730. 23 Soloman MH et al. Hormone ablation treatments as neoadjuvant considered investigational and studied in the context of therapy prior to radical prostatectomy. Clin Invest Med 1993; 16: controlled clinical studies. 532 ± 538. 24 Aprikian AG et al. Experience with neoadjuvant diethylstibesterol and radical prostatectomy in patients with localy advanced prostate cancer. Br J Urol 1994; 74: 630 ± 636. References 25 Fair WR et al. The role of neoadjuvant hormonal manipulation in localized prostate cancer. Cancer 1993; 71: 1031. 1 Rosen MA et al. Frequency and location of extracapsular exten- 26 Kava BR et al. Interim follow-up of patients receiving neoadju- sion and positive surgical margins in radical prostatectomy vant hormonal therapy prior to radical prostatectomy at Memor- specimens. J Urol 1992; 148: 331. ial Sloan±Kettering Cancer Center: tumor designation as organ- 2 Jones, EC. Resection margin status in radical prostatectomy con®ned and specimen-con®ned does not appear to result from specimens: relationship to type of operation, tumor size, tumor pathologic understaging. Mol Urol 1997; 1: 141 ± 148. grade and local tumor extension. JUrol1990; 144: 89. 27 Sassine AM and Sculman CC. Neoadjuvant hormonal dep- 3 Trapasso JG, deKernion JB, Smith RB, Dorey F. The incidence and rivation before radical prostatectomy. Eur Urol 1993; 24:46± signi®cance of detectable levels of serum PSA after radical 50. prostatectomy. J Urol 1994; 152: 1816. 28 Schulman CC. Neoadjuvant androgen blockade prior to pros- 4 Partin AW et al. The use of prostate speci®c antigen, clinical stage, tatectomy: a retrospective study and critical review. Prostate 1994; and Gleason score to predict pathologic stage in men with 5:9. localized prostate cancer. JUrol1993; 149: 110. 29 HaÈggman M et al. Neoadjuvant GnRH agonist treatment and 5 Lange PH, Ercole CJ, Vessella RL. Serum prostatic speci®c antigen total prostatectomy. Eur Urol 1993; 24: 456 ± 460. in the management of patients after radical prostatectomy. JUrol 30 Labrie F et al. Down-staging of early stage prostate cancer: the 1988; 139(part 2): 607A. ®rst randomized trial of neoadjuvant combination therapy with 6 Frazier HA, Robertson JE, Humphrey PA, Paulson DF. Is PSA of ¯utamide and a luteinizing hormone-releasing hormone agonist. clinical signi®cance in evaluating outcome after radical prosta- Urology 1995; 44: 29. tectomy. J Urol 1993; 149: 516. 31 Soloway MS et al. Randomized prospective study compring 7 Kupelian P et al. Correlation of clinical and pathologic factors radical prostatectomy alone versus radical prostatectomy pre- with rising prostate-speci®c antigen pro®les after radical prosta- ceded by androgen blockade in clinical stage B2 (T2bN 6 M0) tectomy alone for clinically localized prostate cancer. Urology prostate cancer. JUrol1995; 154: 424 ± 428. 1996; 48:2. 32 Van Poppel H et al. Neoadjuvant hormonal therapy before radical 8 Dillioglugil O et al. Hazard rates for progression after radical prostatectomy decreases the number of positive surgical margins prostatectomy for clinically localized prostate cancer. Urology in stage T2 prostate cancer: interim results of a prospective 1997; 50: 93 ± 99. randomized trial. JUrol1995; 154: 429 ± 434.

Prostate Cancer and Prostatic Diseases Neoadjuvant hormonal therapy ME Gleave et al

144 33 Goldenberg SL, Klotz LH, Jewitt MAS, and the Canadian Urolo- 44 Gleave ME et al. Long-term neoadjuvant hormone therapy prior gic Oncology Group. Randomized controlled study of neoadju- to radical prostatectomy. Analysis of outcome by preoperative vant reversible androgen withdrawal therapy with cyproterone risk factors. Mol Urol 1998; 2: 171 ± 179. acetate in the surgical management of localized prostate cancer. 45 Catalona WJ, Ramos CG, Carvalhal GF. Contemporary results of JUrol1996; 156: 873 ± 877. anatomic radical prostatectomy. CA Cancer J Clin 1999; 49: 34 Fair WR and Kava B. Neoadjuvant hormonal therapy: update on 282 ± 296. memorial Sloan ± Kettering Cancer Center Trials. Mol Uro 1997; 1: 46 Meyer F et al. Neoadjuvant hormonal therapy before radical 135 ± 140. prostatectomy and risk of prostate speci®c antigen failure. JUrol 35 Witjes W, Schulman CC, Debruyne MJ, and the European Study 1999; 162: 2024 ± 2028. Group on Neoadjuvant Treatment for Prostate Cancer. Prelimin- 47 Gleave ME, Goldenberg SL, Warner J and the Canadian Uro- ary results of a prospective randomized study copartin radical Oncology Group. Randomized Comparative Study of 3 vs 8 prostatectomy versus radical prostatectomy associated with months neoadjuvant hormone therapy prior to radical prosta- neoadjuvant hormonal combination therapy in T2-3N0M0 tectomy: biochemical and pathological effects. J Urol 1999; 161: prostatic carcinoma. Urology 1997; 49: 65 ± 69. 154. 36 Hugosson J et al. The risk of malignancy in the surgical margin at 48 Bazinet M et al. Morphologic changes induced by neoadjuvant radical prostatectomy reduced almost 3-fold in patients given androgen ablation may result in underdetection of positive neo-adjuvant hormone treatment. Eur Urol 1996; 29: 413 ± 419. surgical margins and capsular involvement by prostatic adeno- 37 Soloway MS et al. Radical prostatectomy alone vs. radical pros- carcinoma. Urology 1997; 49: 721 ± 725. tatectomy preceeded by androgen blockade in cT2b prostate 49 Patterson R et al. Immunohistochemical analysis of radical pros- cancer: 24 month results. JUrol1997; 157: 160A. tatectomy specimens after 8 months of neoadjuvant hormone 38 Klotz LH et al. CUOG randomized trial of neoadjuvant androgen therapy. Mol Urol 1999; 3: 277 ± 286. ablation before radical prostatectomy: 36 month post treatment 50 Tsihlias J et al. Loss of cyclin-dependent kinase inhibitor p27Kip1 PSA results. Canadian Urologic Oncology Group. Urology 1999; is a novel prognostic factor in localized human prostate adeno- 53: 757 ± 763. carcinoma. Cancer Res 1998; 58: 542 ± 548. 39 Goldenberg SL et al. Intermittent androgen suppression in the 51 McDonnell TJ et al. Expression of bcl-2 oncoprotein and p53 treatment of prostatic carcinoma: a preliminary report. Urology protein accumulation in bone marrow metastases of androgen 1995; 45: 839 ± 845. independent prostate cancer. JUrol1997; 157: 569 ± 574. 40 Sayer J, Ramirez EI, von Eschenbach AC. Retrospective review of 52 Reed JC. Bcl-2 and the regulation of programmed cell death. J Cell prostate cancer patients with lymph node metastases. J Urol 1992; Biol 1994; 124:1±6. 147: 52A. 53 Souhami L et al. A randomized trial of chemotherapy followed by 41 Cheng CWS, Bergstralh EJ, and Zincke H. Stage D1 prostate pelvic radiation therapy in stage IIIB carcinoma of the cervix. cancer. A nonrandomized comparison of conservative treatment J Clin Oncol 1991; 9: 970 ± 977. options versus radical prostatectomy. Cancer 1993; 71: 996 ± 1004. 54 Murray N et al. Importance of timing for thoracic irradiation in 42 Sato N et al. Intermittent androgen suppression delays time to the combined modality treatment of limited stage small cell lung non-androgen regulated prostate speci®c antigen gene expression cancer. J Clin Oncol 1993; 11: 336 ± 344. in the human prostate LNCaP tumour model. J Steroid Biochem 55 Hall RR for the MRC Study Group. Neoadjuvant CMV che- Mol Biol 1996; 58: 139 ± 146. motherapy and cystectomy or radiotherapy in muscle-invasive 43 Gleave ME et al. Biochemical and pathological effects of eight bladder cancer. First analysis of MRC/EORTC Intercontinental months of androgen withdrawal therapy prior to radical prosta- trial. Proc Am Soc Clin Oncol 1996; 15: 244A. tectomy in clinically con®ned prostate cancer. J Urol 1996; 155: 213 ± 219.

Prostate Cancer and Prostatic Diseases