- Email: [email protected]
Duration of neoadjuvant androgen deprivation therapy before radical prostatectomy and disease-free survival in men with prostate cancer
DURATION OF NEOADJUVANT ANDROGEN DEPRIVATION THERAPY BEFORE RADICAL PROSTATECTOMY AND DISEASE-FREE SURVIVAL IN MEN WITH PROSTATE CANCER FRANC¸OIS MEYER, ISABELLE BAIRATI, CAROLINE BE´DARD, LOUIS LACOMBE, BERNARD TEˆTU, AND YVES FRADET
ABSTRACT There is little evidence that neoadjuvant androgen deprivation therapy (ADT) of 3 months’ duration before radical prostatectomy (RP) favorably influences disease-free survival. However, recent data suggest that prolonged treatment may improve outcome. We conducted a prospective cohort study to determine whether ADT of either standard or prolonged duration before RP influences the risk of prostate-specific antigen (PSA) failure. We followed 756 men treated for prostate cancer by RP between 1991 and 1998 in Quebec City. Of these, 240 received combined neoadjuvant ADT for either ⱕ92 days (129 men) or ⱖ93 days (111 men), and 516 were treated by RP alone. Multivariate Cox regression was used to estimate the hazard ratios (HR) of PSA failure (⬎0.3 ng/mL) associated with treatment regimen controlling for age, clinical stage, grade, and initial PSA level. The median duration of follow-up was 4 years. Compared with men treated by RP alone, those who received neoadjuvant ADT for ⱖ93 days had an HR of PSA failure of 0.60. The inverse association with the risk of PSA failure became statistically significant from the third year on, reached its greatest magnitude after 4 years, and was still present 8 years after RP. No association was observed for ADT of ⱕ92 days. These results suggest that neoadjuvant ADT before RP has a real, delayed, and persistent effect on disease-free survival, if and only if ADT is prolonged beyond 3 months. UROLOGY 58 (Suppl 2A): 71–77, 2001. © 2001, Elsevier Science Inc.
N
eoadjuvant androgen deprivation therapy (ADT) before radical prostatectomy (RP) reduces the likelihood of positive margins, extracapsular extension, and lymph node invasion.1,2 However, there is still controversy about the prognostic significance of these pathologic indicators after hormone manipulation. Furthermore, the published results of the randomized, controlled trials of neoadjuvant ADT before RP have not provided strong evidence in favor of the hypothesis that ADT influences disease progression as judged by the rate of prostate-specific antigen (PSA) failure.3–9 However, the duration of ADT in these trials never exceeded 3 months. There is now evidence that, compared with a 3-month standard treatment, a prolonged neoadjuvant ADT of 6 or 8 From the Cancer Research Center and Faculty of Medicine, Laval University, Quebec, Canada This study was conducted with financial support from the Medical Research Council of Canada (Grant No. MOP-36447) Reprint requests: Franc¸ois Meyer, MD, Laval University Cancer Research Center, CHUQ, l’Hoˆtel-Dieu de Que´bec, 11, Coˆte du Palais, Quebec City, Quebec G1R 2J6 Canada. E-mail: francois. [email protected] © 2001, ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED
months further reduces tumor volume10 and PSA nadir levels,11 and it further decreases the proportion of men with positive margins.11 We recently reported that neoadjuvant ADT combining an antiandrogen and a luteinizing hormone–releasing hormone (LHRH) analog was associated with a better disease-free survival, until biochemical failure, when the duration of treatment was ⬎3 months.12 A randomized trial is being conducted to compare disease-specific survival in prostate cancer patients treated by RP after combined ADT of either 3 or 8 months’ duration, but data on PSA failure rates are not yet available.13 We present the results of a prospective cohort study of 756 men treated by RP and observed for a median duration of 4 years in which we assessed whether prolonged combined neoadjuvant ADT was associated with improved biochemical disease–free survival. METHODS Between January 1991 and December 1998, 883 men were treated by RP as their initial treatment for prostate cancer at 0090-4295/01/$20.00 PII S0090-4295(01)01245-6 71
the Hoˆtel-Dieu de Que´bec, Que´bec City University Hospital. Of these, 571 were treated by RP without neoadjuvant ADT, 245 received an ADT, combining an antiandrogen and an LHRH analog before RP, and 67 received an antiandrogen alone before RP. The present article is based on the first 2 groups (a total of 816 patients), and the analyses focus on the effect of neoadjuvant combined ADT. Because treatment outcome is based primarily on biochemical disease–free survival documented by postprostatectomy PSA levels, 60 men were excluded from the study population, because PSA failure could not be assessed (47 were under adjuvant ADT at the end of the follow-up period and 13 were lost to follow-up immediately after RP). Thus, the study population comprised 756 patients, aged 44 to 74 years at the time of RP, of which 516 were treated by RP alone and 240 received a neoadjuvant combined ADT before RP. The combined ADT consisted of leuprolide or goserelin, and flutamide, bicalutamide, nilutamide, or cyproterone acetate. Because, until recently, the standard prescribed duration of neoadjuvant ADT was 3 months, we considered ADT duration in 2 categories: ⱕ92 days versus ⱖ93 days. The first group comprised 129 men with a mean ADT duration of 89 days and is referred to as the 3-month ADT group. The second group comprised 111 men who received ADT for a mean duration of 151 days and is referred to as the 5-month ADT group. Patients’ hospital records were reviewed to abstract clinical, surgical, pathologic, and biochemical data and treatment information up to July 2000. Clinical stage was determined according to the fifth edition of the TNM Classification of Malignant Tumours.14 Tumor grade was assessed according to the Gleason scoring system using the diagnostic prostate biopsy material.15 The initial PSA value was categorized into 4 levels: (1) ⱕ3.9 ng/mL; (2) 4.0 to 9.9 ng/mL; (3) 10.0 to 19.9 ng/mL; and (4) ⱖ20.0 ng/mL. Pathologic stage was recorded according to the fifth revision of the TNM classification14 and surgical margin involvement was noted. All patients had bilateral lymph-node dissection. The RP specimens were processed according to a standardized protocol with 5-mm sections. For each patient, the pathologist systematically examined 25 to 30 blocks and the inked margins. Information on adjuvant therapies (adjuvant ADT and external-beam adjuvant radiation therapy) after RP was recorded. For each patient, the duration of follow-up was calculated as the time elapsed between RP and the time of PSA failure, or the date of the last PSA test. PSA failure was defined as 2 consecutive PSA values (or, for 15 patients, 1 final value) ⬎0.3 ng/mL. Bilateral orchiectomy in the absence of PSA increase was also considered as biochemical failure. Medical ADT for palliative treatment was never instituted without evidence of PSA failure. The relation of neoadjuvant ADT to age, clinical stage, biopsy grade, and initial PSA was evaluated with chi-square tests. Similarly, the distributions of these variables according to categories of ADT duration were compared by chi-square tests. The relation of neoadjuvant ADT to pathologic stage and surgical margin involvement was considered across 3 groups of patients: (1) RP alone; (2) ADT ⱕ3 months’ duration; and (3) ADT ⬎3 months’ duration. Logistic regression was used to assess linear trends across treatment regimens while controlling for age, stage, grade, and initial PSA level. Kaplan-Meier curves were used to present crude survival until PSA failure in the 3 treatment groups. Cox proportional hazard models were used to assess the association between neoadjuvant ADT of either standard or prolonged duration and the relative risk of PSA failure after RP, with RP alone as the reference category.16 Hazard ratios (HR) and their 95% confidence intervals (CI) were adjusted for age, clinical stage, biopsy grade, and initial PSA level. Similar Cox regression models were used to assess the association between ADT regimens and biochemical disease– 72
TABLE I. Percentages of patients with selected characteristics at time of diagnosis according to whether neoadjuvant ADT was used before radical prostatectomy RP ADT ⴙ RP (N ⴝ 516) (N ⴝ 240) P Value Age (years) ⱕ59 60–64 ⱖ65 Clinical stage T1 T2 T3 Biopsy Gleason score 2–4 5,6 7 8–10 PSA (ng/mL) ⱕ3.9 4.0–9.9 10.0–19.9 ⱖ20.0
29.9 30.6 40.5
35.0 37.5 27.5
0.003
18.6 79.3 2.1
10.0 81.2 8.8
0.001
36.2 45.7 13.6 4.5
27.5 52.1 12.9 7.5
0.04
10.7 49.6 27.9 11.8
15.0 43.8 22.5 18.7
0.01
ADT ⫽ androgen deprivation therapy, PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy.
free survival up to each yearly anniversary after RP. Although the preceding analyses assessed the cumulative effect up to a given time in the follow-up period, we also examined the effect of ADT within subsequent periods in the follow-up using other Cox regression models. The magnitude of the association between prolonged neoadjuvant ADT and disease-free survival was also studied by Cox regression models according to categories of determinants at the time of diagnosis: age, stage, grade, and PSA level. The proportionality assumptions of the Cox models were verified. All reported P values are 2 sided. All analyses were performed with statistical analysis system (SAS Institute; Cary, NC) software.
RESULTS Patients’ characteristics at the time of diagnosis are presented in Table I. Men who were prescribed ADT were at statistically significant higher risk of poor outcome on the basis of clinical stage, grade, and PSA level. Table II shows the characteristics of patients treated by ADT according to treatment duration: standard duration (3 months) versus prolonged duration (5 months). No statistical difference was observed in the distribution of baseline characteristics between the 2 groups of patients. The decreasing PSA trend during the first 3 months under ADT was similar in the 2 groups. In men receiving ADT for the standard duration, the mean PSA levels were 15.3, 2.4, and 0.28 ng/mL, respectively at diagnosis, after 6 weeks and after 3 months of ADT. In men receiving prolonged ADT, the corresponding values were 14.7, 2.2, and 0.31 ng/mL. UROLOGY 58 (Supplement 2A), August 2001
TABLE II. Percentages of patients with selected characteristics at time of diagnosis according to duration of neoadjuvant androgen deprivation therapy (ADT) 3-month ADT (N ⴝ 129)
5-month ADT (N ⴝ 111)
P Value
34.1 39.5 26.4
36.0 35.1 28.8
0.8
10.1 83.7 6.2
9.9 78.4 11.7
0.3
33.3 48.8 10.9 7.0
20.7 55.9 15.3 8.1
0.2
17.1 41.8 24.0 17.1
12.6 46.0 20.7 20.7
0.6
Age (years) ⱕ59 60–64 ⱖ65 Clinical stage T1 T2 T3 Biopsy Gleason score 2–4 5,6 7 8–10 PSA (ng/mL) ⱕ3.9 4.0–9.9 10.0–19.9 ⱖ20.0
PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy.
TABLE III. Percentages of patients with pathological indicators of prognosis at the time of radical prostatectomy by treatment regimen: radical prostatectomy alone and neoadjuvant ADT of either standard or prolonged duration followed by radical prostatectomy RP 3-month ADT 5-month ADT P Value (N ⴝ 516) (N ⴝ 129) (N ⴝ 111) for Trend* Positive surgical margins Extracapsular extension Seminal vesicle invasion Lymph node invasion
48.7 51.9 15.5 13.6
29.5 41.1 19.4 11.6
12.6 21.6 7.2 8.1
0.0001 0.0001 0.004 0.002
ADT ⫽ androgen deprivation therapy, RP ⫽ radical prostatectomy. * P values for trend were obtained from logistic regression models including age, clinical stage, biopsy Gleason score, and initial PSA level.
In this group, the average PSA level after 5 months of ADT was 0.25 ng/mL. The effect of ADT and ADT duration on pathologic indicators of prognosis at the time of RP was examined. The proportions of patients with positive surgical margins, extracapsular extension, seminal vesicle invasion, and lymph node invasion were compared across the 3 treatment groups. After controlling for age, clinical stage, Gleason score, and initial PSA, there was a statistically significant trend for each of these indicators, with higher proportions in men treated by RP alone and lower proportions in those treated by prolonged ADT followed by RP (Table III). The median duration of follow-up was 4 years, UROLOGY 58 (Supplement 2A), August 2001
with 33% of the patients being observed for ⬎5 years and 3% for ⬎8 years. Among the 756 patients in the study population, 185 had PSA failure during the follow-up period. Overall, the 5-year biochemical disease–free survival rate was 70.6%. Altogether, 15 men had a bilateral orchiectomy immediately after RP without evidence of PSA failure. Of these men, only 1 had received neoadjuvant ADT. The decision to perform a surgical castration was based mostly on pathologic findings at the time of RP (pT3N⫹). Figure 1 shows the proportion of men without PSA failure during the first 8 years of the follow-up period among men treated by RP alone and among those treated by neoadjuvant ADT of either standard or prolonged duration 73
followed by RP. During the first year of the follow-up period, biochemical disease–free survival was higher for men who received ADT than for those who did not. From the third year after RP and thereafter, the crude biochemical disease–free survival was better for men who received prolonged neoadjuvant ADT than for those in the 2 other treatment groups. However, the validity of the crude analysis of treatment efficacy by KaplanMeier survival curves is questionable, because men treated by ADT had a worse risk profile at the time of diagnosis, as documented in Table I. The results from the multivariate Cox regression analysis confirmed the independent predictive capacity of clinical stage, biopsy grade, and initial PSA level on the risk of biochemical failure (Table IV). After controlling for these known determinants of disease progression, the HR associated with neoadjuvant ADT of a standard 3-months’ duration was 1.01, CI: 0.70 to 1.45, indicating that neoadjuvant ADT of 3-months’ duration did not influence PSA failure rate. In contrast, the HR associated with neoadjuvant ADT of prolonged duration was 0.60, CI: 0.38 to 0.94 — clearly suggesting a beneficial effect of prolonged neoadjuvant ADT on disease-free survival. A total of 43 patients received adjuvant external-beam radiation therapy and 112 received adjuvant ADT. These treatments were not associated with the risk of biochemical failure in the multivariate Cox models: HR of 0.82, CI: 0.46 to 1.47 for radiation therapy and HR of 1.27, CI: 0.87 to 1.86 for adjuvant ADT. When the regression model included adjuvant therapies, the association between neoadjuvant ADT and PSA failure remained similar. To study the association between neoadjuvant ADT (of either standard or prolonged duration) and biochemical disease–free survival cumulatively from RP until specific times in the follow-up period, multivariate Cox regression models, similar to that presented in the main analysis (Table IV) were used, censoring the observations at each yearly anniversary after RP. The HRs of PSA failure are presented in Figure 2. As anticipated, patients who had received neoadjuvant ADT had a lower risk of PSA failure in the first year after RP than those treated by surgery alone. Two years after RP and for the rest of the follow-up period, men treated by 3-month neoadjuvant ADT had a risk of biochemical failure similar to that of men treated by surgery alone. In contrast, for men who received 5-month ADT, the inverse association with the risk of PSA failure became statistically significant from the third year and forward, reached its greatest magnitude after 4 years (HR ⫽ 0.53, CI: 0.31 to 0.89), and was still present 8 years after RP (HR ⫽ 0.60, CI: 0.38 to 0.94). Other multivariate Cox models were used to as74
FIGURE 1. Kaplan-Meier curves for disease-free survival until PSA failure according to treatment regimen: radical prostatectomy alone (RP) and neoadjuvant ADT of either standard (3mADT) or prolonged (5mADT) duration followed by radical prostatectomy. Log-rank test P ⫽ 0.32. 3m ⫽ 3 months; 5m ⫽ 5 months; ADT ⫽ androgen deprivation therapy; PSA ⫽ prostate-specific antigen; RP ⫽ radical prostatectomy.
TABLE IV. Hazard ratios of biochemical failure and 95% confidence intervals associated with known indicators of prognosis and treatment regimen
Age (per 1-year increment) Clinical stage T1† T2 T3 Biopsy Gleason score 2–4† 5,6 7 8–10 PSA (ng/mL) ⱕ3.9† 4.0–9.9 10.0–19.9 ⱖ20.0 Treatment regimen RP alone† 3-month ADT 5-month ADT
Hazard Ratio
95% Confidence Interval*
1.01
0.98–1.04
1.00 1.37 2.11
0.83–2.28 1.05–4.25
1.00 1.61 3.43 3.06
1.09–2.37 2.21–5.34 1.74–5.40
1.00 1.79 2.83 4.45
0.91–3.49 1.43–5.58 2.22–8.96
1.00 1.01 0.60
0.70–1.45 0.38–0.94
ADT ⫽ androgen deprivation therapy, PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy. * Hazard ratios and confidence intervals were obtained from Cox regression models including simultaneously all variables in the table. † Reference category.
sess the HR of PSA failure associated with prolonged ADT during subsequent follow-up time periods (Figure 3). The effect of prolonged neoadjuvant ADT on biochemical disease–free surUROLOGY 58 (Supplement 2A), August 2001
FIGURE 2. Hazard ratios of PSA failure until each yearly anniversary after RP associated with neoadjuvant ADT of either standard or prolonged duration with no ADT before RP as the reference category. The multivariate Cox regression models included age, clinical stage, biopsy Gleason score, and initial PSA level. The vertical bars represent the 95% confidence intervals associated with the HRs for neoadjuvant ADT of prolonged duration. 3m ⫽ 3 months; 5m ⫽ 5 months; ADT ⫽ androgen deprivation therapy; PSA ⫽ prostatespecific antigen; RP ⫽ radical prostatectomy.
FIGURE 3. Hazard ratios of PSA failure during subsequent time periods in the follow-up after RP associated with neoadjuvant ADT of prolonged duration using RP alone as the reference category. The multivariate Cox regression models included age, clinical stage, biopsy Gleason score, and initial PSA level. Because the number of events was small, years 6, 7, and 8 were considered together. ADT ⫽ androgen deprivation therapy; PSA ⫽ prostate-specific antigen; RP ⫽ radical prostatectomy.
vival occurred mostly during the third year (HR ⫽ 0.22, CI: 0.05 to 0.95) and the fourth year after RP (HR ⫽ 0.26, CI: 0.05 to 1.31). Additional analyses were conducted to examine whether the association between prolonged ADT UROLOGY 58 (Supplement 2A), August 2001
and biochemical disease–free survival was stronger among subgroups of patients. We further investigated the relation between prolonged neoadjuvant ADT and biochemical disease–free survival, by applying the same multivariate Cox regression models to subgroups of the study population formed on the basis of baseline risk factors. Table V presents the HRs for PSA failure for men treated by prolonged neoadjuvant ADT in comparison with men treated by RP alone by category of age, clinical stage, Gleason score, and initial PSA level. This analysis suggests that the potential beneficial effect of prolonged neoadjuvant ADT was relatively less important for patients in the highest category of risk for all variables considered. In particular, there was significant trend for tumor grade (P ⫽ 0.04) showing that the magnitude of the association decreased with increasing risk. DISCUSSION This is 1 of few reports on the effect of duration of neoadjuvant ADT on biochemical failure rates after RP. We observed that patients who received a combined neoadjuvant ADT for ⬎3 months before RP had a statistically significant biochemical survival advantage compared with those treated by RP alone or with those who received a similar treatment for ⱕ3 months. This effect was progressive, became fully established 4 years after RP, and was still present 8 years after RP. Our study has many strengths, including its large size, the relatively long follow-up period, and the multivariate analysis used in the assessment of treatment regimens and disease-free survival. The main concern in interpreting the results of our cohort study comes from the difference in baseline risk between patients treated by RP alone and those who received neoadjuvant ADT of either standard or prolonged duration followed by RP. The decision to initiate ADT was influenced by the severity and aggressiveness of prostate cancer as judged by clinical stage, biopsy grade, and initial PSA. Men at higher risk of progression were more likely to receive ADT. This suggests that clinicians believed in the efficacy of neoadjuvant ADT, and it underscores the need of multivariate analysis to control confounding by baseline level of risk in the assessment of the effect of ADT on disease-free survival. On the other hand, there was no association between clinical stage, grade, or PSA level and the duration of ADT, suggesting that the decision to prolong ADT for ⬎3 months was not related to risk of PSA failure inferred from patient and tumor characteristics at the time of diagnosis. Because the decreasing PSA trend under ADT was similar in the 2 groups, it is unlikely that the decision to prolong ADT was based on patients’ drug tolerance or re75
TABLE V. Hazard ratios of biochemical failure and 95% confidence intervals associated with prolonged neoadjuvant ADT by category of baseline indicators of prognosis
Entire study population Age (years) ⱕ64 ⱖ65 Clinical stage T1, T2 T3 Biopsy Gleason score 2–4 5,6 7 8–10 PSA (ng/mL) ⱕ9.9 10.0–19.9 ⱖ20.0
Hazard Ratio
95% Confidence Interval*
0.60
0.38–0.94
0.44 0.86
0.25–0.79 0.43–1.73
0.57 0.62
0.34–0.94 0.17–2.26
0.28 0.43 0.83 1.20
0.04–2.09 0.22–0.87 0.37–1.87 0.32–4.53
0.44 0.51 0.62
0.20–0.97 0.20–1.31 0.29–1.32
ADT ⫽ androgen deprivation therapy, PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy. * Hazard ratios and confidence intervals were obtained from Cox regression models restricted to men within a category of a given variable and including simultaneously all other variables in the table.
sponse to therapy. We systematically used multivariate statistical analyses to enhance the comparability of the 3 treatment groups by controlling the confounding effect of well-established determinants of the risk of progression: clinical stage, tumor grade, and PSA level. It is therefore unlikely that differences in the baseline risk profile have affected the results of the multivariate analyses. This study enabled us to examine the effect of neoadjuvant ADT of either standard or prolonged duration over the course of 8 years of follow-up after RP. Our results confirmed the well-documented effect of ADT on pathologic prognostic factors at the time of RP.1,2 After controlling for the confounding effect of clinical stage, grade, and initial PSA level, we observed that men who received ADT had reduced rates of positive margins, extracapsular extension, and positive lymph nodes, compared with men treated by RP alone. Furthermore, the longer the ADT duration, the lower was the risk of positive margins, extracapsular extension, and lymph node involvement. The better pathologic profile of men treated by prolonged ADT compared with those receiving a standard treatment of 3 months has also been observed in an ongoing randomized trial.13 The long-term prognostic significance of the pathologic changes observed after ADT is not clear, because most randomized trials comparing ADT of 76
3 months’ duration with RP alone have demonstrated a strong effect on pathologic indicators but have not shown a clear survival advantage for men treated by neoadjuvant ADT.3–9 In each of the 5 randomized trials3–5,7,8 where combined neoadjuvant ADT was given for 3 months before RP, the PSA failure rate was slightly lower for men in the neoadjuvant ADT arm than in the RP-alone arm. The difference in biochemical disease–free survival failed to reach statistical significance in each individual trial and even in a meta-analysis combining the data from these trials.12 Our data suggest that neoadjuvant ADT may have an immediate and transient effect without prognostic significance. The lower risk of PSA failure during the first year after RP among men treated by neoadjuvant ADT could be explained by both biologic and management factors: (1) the persisting influence of ADT during the months after its discontinuation could mask or delay biochemical failure; (2) treating surgeons made a management decision, based on the findings at the time of RP, which are affected by ADT (downstaging), and they proceeded to a bilateral orchiectomy without waiting for the occurrence of PSA failure. Because we categorized these patients as experiencing PSA failure, the hazard of biochemical failure among men treated by ADT was artificially lowered. Both of these explanations appear to have played a role in our study population. On the other hand, the results of this study strongly suggest that neoadjuvant ADT has a delayed, persistent, and real effect on biochemical disease–free survival. This effect is observed only among men treated by neoadjuvant ADT for a prolonged duration. It is clearly observed during the third and fourth year after RP, and it has long-lasting consequences on overall biochemical disease–free survival up to 8 years after RP. We reported previously that patients treated by prolonged ADT— combining an LHRH analog and an antiandrogen— had an improved biochemical disease–free survival rate compared with those treated by RP alone or by ADT of standard duration before RP.12 In the present article, we identify the effect of timing of ADT on biochemical disease–free survival during the course of follow-up after RP. As neoadjuvant hormone therapy for prostate cancer delays the time of surgical radical treatment, prolonged ADT would be justified if a significant survival advantage was obtained for patients with poor prognosis as inferred from grade, stage, and PSA at the time of diagnosis. Our data suggest that, in relative terms, less benefit is observed for patients at higher risk of biochemical failure. In particular, patients with a high Gleason score seem to benefit less from prolonged ADT than other patients. If other studies confirm this observation, it UROLOGY 58 (Supplement 2A), August 2001
would probably weigh against the use of prolonged ADT before RP. CONCLUSION Neoadjuvant ADT before RP had been adopted with enthusiasm after the demonstration of its effect on pathologic indicators at the time of RP. After the publication of the initial survival data from the randomized trials addressing the longterm efficacy of ADT, this management strategy has fallen into disuse. Our results suggest that neoadjuvant ADT has a real beneficial effect on biochemical disease–free survival among patients treated for prostate cancer by RP. This effect appears to be delayed and is observed only when ADT is prolonged for ⬎3 months. However, the benefit of prolonged neoadjuvant ADT could be demonstrated only in a randomized, controlled trial, such as the trial underway. REFERENCES 1. Bonney WW, Schned AR, and Timberlake DS: Neoadjuvant androgen ablation for localized prostatic cancer: pathology methods, surgical endpoints and meta-analysis of randomized trials. J Urol 160: 1754 –1760, 1998. 2. Wieder JA, and Soloway MS: Incidence, etiology, location, prevention and treatment of positive surgical margins after radical prostatectomy for prostate cancer. J Urol 160: 299 –315, 1998. 3. Witjes WPJ, Schulman CC, Debruyne FMJ, for the European Study Group on Neoadjuvant Treatment of Prostate Cancer: Preliminary results of a prospective randomized study comparing radical prostatectomy versus radical prostatectomy associated with neoadjuvant hormonal combination therapy in T2–3 N0 M0 prostatic carcinoma. Urology 49(suppl 3A): 65– 69, 1997. 4. Aus G, Abrahamsson PA, Ahlgren G, et al: Hormonal treatment before radical prostatectomy: a 3-year follow-up. J Urol 159: 2013–2017, 1998. 5. Homma Y: Neoadjuvant androgen deprivation preced-
ing to radical prostatectomy: its role in short-term and longterm outcomes. Nippon Rinsho 58: 2162–2166, 1998. 6. Klotz LH, Goldenberg SL, Jewett M, et al: CUOG randomized trial of neoadjuvant androgen ablation before radical prostatectomy: 36 months post-treatment PSA results. Urology 53: 757–763, 1999. 7. Wildschutz T, Louis L, Hourriez L, et al: Neoadjuvant hormonal treatment prior to radical prostatectomy: follow-up of a prospective randomized study [abstract]. Eur Urol 30: 210, 1996, Abstract 774. 8. Soloway M, Sharifi R, Wajsman Z, et al, and the Lupron Depot Neoadjuvant Study Group: Radical prostatectomy alone vs radical prostatectomy preceded by androgen blockade in cT2b prostate cancer: 24 month results [abstract]. J Urol 157: 160, 1997, Abstract 619. 9. Baert LV, Goethuys HJ, De Ridder DJ, et al: Neoadjuvant treatment before radical prostatectomy decreases the number of positive margins in cT2–T3 but has no impact on PSA progression or survival [abstract]. J Urol 159: 61, 1998, Abstract 229. 10. Van der Kwast TH, Tetu B, Candas B, et al: Prolonged neoadjuvant combined androgen blockade leads to a further reduction of prostatic tumor volume: three versus six months of endocrine therapy. Urology 53: 523–529, 1999. 11. Gleave ME, Goldenberg SL, Jones EC, et al: Biochemical and pathological effects of 8 months of neoadjuvant androgen withdrawal therapy before radical prostatectomy in patients with clinically confined prostate cancer. J Urol 155: 213–219, 1996. 12. Meyer F, Moore L, Bairati I, et al: Neoadjuvant hormonal therapy before radical prostatectomy and risk of prostate specific antigen failure. J Urol 162: 2024 –2028, 1999. 13. Gleave M, Goldenberg SL, Warner J, et al: Randomized comparative study of 3 vs 8 months of neoadjuvant hormonal therapy prior to radical prostatectomy: biochemical and pathological effects [abstract]. J Urol 161: 154, 1999, Abstract 592. 14. Sobin LH, and Wittekind CH (Eds): International Union Against Cancer (UICC). TNM Classification of Malignant Tumours. 5th edition. New York, John Wiley and Sons, 1997. 15. Bostwick DG: Grading prostate cancer. Am J Clin Pathol 102(suppl 1): S38 –S56, 1994. 16. Clayton D, and Hills M: Statistical Models in Epidemiology. Oxford, Oxford University Press, 1993.
DISCUSSION FOLLOWING DR. FRANC¸OIS MEYER’S PRESENTATION Harry W. Daniell, MD (Redding, California): Your 5-month group actually had therapy from 3 months and 1 day to a year or some other longer period. Did you break that group up into shorter and longer subgroups to see if there was any relation? Franc¸ois Meyer, MD, DSc (Quebec City, Quebec, Canada): No. We felt that the numbers were not sufficient. It is not only the number of patients in each group, but it is also the number of failures that is important.
UROLOGY 58 (Supplement 2A), August 2001
Dr. Daniell: Was the division at 3 months made before the data were analyzed, or was that a cut-off point that was partly determined by the results that you got? Dr. Meyer: Both. Most of the randomized trials have 3 months of treatment in their protocols, so we considered 3 months as the reference, the standard duration of treatment. We also needed to find a cut point to have different groups according to duration. If we wanted to have groups of similar size, that would also be at 3 months.
77
ABSTRACT There is little evidence that neoadjuvant androgen deprivation therapy (ADT) of 3 months’ duration before radical prostatectomy (RP) favorably influences disease-free survival. However, recent data suggest that prolonged treatment may improve outcome. We conducted a prospective cohort study to determine whether ADT of either standard or prolonged duration before RP influences the risk of prostate-specific antigen (PSA) failure. We followed 756 men treated for prostate cancer by RP between 1991 and 1998 in Quebec City. Of these, 240 received combined neoadjuvant ADT for either ⱕ92 days (129 men) or ⱖ93 days (111 men), and 516 were treated by RP alone. Multivariate Cox regression was used to estimate the hazard ratios (HR) of PSA failure (⬎0.3 ng/mL) associated with treatment regimen controlling for age, clinical stage, grade, and initial PSA level. The median duration of follow-up was 4 years. Compared with men treated by RP alone, those who received neoadjuvant ADT for ⱖ93 days had an HR of PSA failure of 0.60. The inverse association with the risk of PSA failure became statistically significant from the third year on, reached its greatest magnitude after 4 years, and was still present 8 years after RP. No association was observed for ADT of ⱕ92 days. These results suggest that neoadjuvant ADT before RP has a real, delayed, and persistent effect on disease-free survival, if and only if ADT is prolonged beyond 3 months. UROLOGY 58 (Suppl 2A): 71–77, 2001. © 2001, Elsevier Science Inc.
N
eoadjuvant androgen deprivation therapy (ADT) before radical prostatectomy (RP) reduces the likelihood of positive margins, extracapsular extension, and lymph node invasion.1,2 However, there is still controversy about the prognostic significance of these pathologic indicators after hormone manipulation. Furthermore, the published results of the randomized, controlled trials of neoadjuvant ADT before RP have not provided strong evidence in favor of the hypothesis that ADT influences disease progression as judged by the rate of prostate-specific antigen (PSA) failure.3–9 However, the duration of ADT in these trials never exceeded 3 months. There is now evidence that, compared with a 3-month standard treatment, a prolonged neoadjuvant ADT of 6 or 8 From the Cancer Research Center and Faculty of Medicine, Laval University, Quebec, Canada This study was conducted with financial support from the Medical Research Council of Canada (Grant No. MOP-36447) Reprint requests: Franc¸ois Meyer, MD, Laval University Cancer Research Center, CHUQ, l’Hoˆtel-Dieu de Que´bec, 11, Coˆte du Palais, Quebec City, Quebec G1R 2J6 Canada. E-mail: francois. [email protected] © 2001, ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED
months further reduces tumor volume10 and PSA nadir levels,11 and it further decreases the proportion of men with positive margins.11 We recently reported that neoadjuvant ADT combining an antiandrogen and a luteinizing hormone–releasing hormone (LHRH) analog was associated with a better disease-free survival, until biochemical failure, when the duration of treatment was ⬎3 months.12 A randomized trial is being conducted to compare disease-specific survival in prostate cancer patients treated by RP after combined ADT of either 3 or 8 months’ duration, but data on PSA failure rates are not yet available.13 We present the results of a prospective cohort study of 756 men treated by RP and observed for a median duration of 4 years in which we assessed whether prolonged combined neoadjuvant ADT was associated with improved biochemical disease–free survival. METHODS Between January 1991 and December 1998, 883 men were treated by RP as their initial treatment for prostate cancer at 0090-4295/01/$20.00 PII S0090-4295(01)01245-6 71
the Hoˆtel-Dieu de Que´bec, Que´bec City University Hospital. Of these, 571 were treated by RP without neoadjuvant ADT, 245 received an ADT, combining an antiandrogen and an LHRH analog before RP, and 67 received an antiandrogen alone before RP. The present article is based on the first 2 groups (a total of 816 patients), and the analyses focus on the effect of neoadjuvant combined ADT. Because treatment outcome is based primarily on biochemical disease–free survival documented by postprostatectomy PSA levels, 60 men were excluded from the study population, because PSA failure could not be assessed (47 were under adjuvant ADT at the end of the follow-up period and 13 were lost to follow-up immediately after RP). Thus, the study population comprised 756 patients, aged 44 to 74 years at the time of RP, of which 516 were treated by RP alone and 240 received a neoadjuvant combined ADT before RP. The combined ADT consisted of leuprolide or goserelin, and flutamide, bicalutamide, nilutamide, or cyproterone acetate. Because, until recently, the standard prescribed duration of neoadjuvant ADT was 3 months, we considered ADT duration in 2 categories: ⱕ92 days versus ⱖ93 days. The first group comprised 129 men with a mean ADT duration of 89 days and is referred to as the 3-month ADT group. The second group comprised 111 men who received ADT for a mean duration of 151 days and is referred to as the 5-month ADT group. Patients’ hospital records were reviewed to abstract clinical, surgical, pathologic, and biochemical data and treatment information up to July 2000. Clinical stage was determined according to the fifth edition of the TNM Classification of Malignant Tumours.14 Tumor grade was assessed according to the Gleason scoring system using the diagnostic prostate biopsy material.15 The initial PSA value was categorized into 4 levels: (1) ⱕ3.9 ng/mL; (2) 4.0 to 9.9 ng/mL; (3) 10.0 to 19.9 ng/mL; and (4) ⱖ20.0 ng/mL. Pathologic stage was recorded according to the fifth revision of the TNM classification14 and surgical margin involvement was noted. All patients had bilateral lymph-node dissection. The RP specimens were processed according to a standardized protocol with 5-mm sections. For each patient, the pathologist systematically examined 25 to 30 blocks and the inked margins. Information on adjuvant therapies (adjuvant ADT and external-beam adjuvant radiation therapy) after RP was recorded. For each patient, the duration of follow-up was calculated as the time elapsed between RP and the time of PSA failure, or the date of the last PSA test. PSA failure was defined as 2 consecutive PSA values (or, for 15 patients, 1 final value) ⬎0.3 ng/mL. Bilateral orchiectomy in the absence of PSA increase was also considered as biochemical failure. Medical ADT for palliative treatment was never instituted without evidence of PSA failure. The relation of neoadjuvant ADT to age, clinical stage, biopsy grade, and initial PSA was evaluated with chi-square tests. Similarly, the distributions of these variables according to categories of ADT duration were compared by chi-square tests. The relation of neoadjuvant ADT to pathologic stage and surgical margin involvement was considered across 3 groups of patients: (1) RP alone; (2) ADT ⱕ3 months’ duration; and (3) ADT ⬎3 months’ duration. Logistic regression was used to assess linear trends across treatment regimens while controlling for age, stage, grade, and initial PSA level. Kaplan-Meier curves were used to present crude survival until PSA failure in the 3 treatment groups. Cox proportional hazard models were used to assess the association between neoadjuvant ADT of either standard or prolonged duration and the relative risk of PSA failure after RP, with RP alone as the reference category.16 Hazard ratios (HR) and their 95% confidence intervals (CI) were adjusted for age, clinical stage, biopsy grade, and initial PSA level. Similar Cox regression models were used to assess the association between ADT regimens and biochemical disease– 72
TABLE I. Percentages of patients with selected characteristics at time of diagnosis according to whether neoadjuvant ADT was used before radical prostatectomy RP ADT ⴙ RP (N ⴝ 516) (N ⴝ 240) P Value Age (years) ⱕ59 60–64 ⱖ65 Clinical stage T1 T2 T3 Biopsy Gleason score 2–4 5,6 7 8–10 PSA (ng/mL) ⱕ3.9 4.0–9.9 10.0–19.9 ⱖ20.0
29.9 30.6 40.5
35.0 37.5 27.5
0.003
18.6 79.3 2.1
10.0 81.2 8.8
0.001
36.2 45.7 13.6 4.5
27.5 52.1 12.9 7.5
0.04
10.7 49.6 27.9 11.8
15.0 43.8 22.5 18.7
0.01
ADT ⫽ androgen deprivation therapy, PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy.
free survival up to each yearly anniversary after RP. Although the preceding analyses assessed the cumulative effect up to a given time in the follow-up period, we also examined the effect of ADT within subsequent periods in the follow-up using other Cox regression models. The magnitude of the association between prolonged neoadjuvant ADT and disease-free survival was also studied by Cox regression models according to categories of determinants at the time of diagnosis: age, stage, grade, and PSA level. The proportionality assumptions of the Cox models were verified. All reported P values are 2 sided. All analyses were performed with statistical analysis system (SAS Institute; Cary, NC) software.
RESULTS Patients’ characteristics at the time of diagnosis are presented in Table I. Men who were prescribed ADT were at statistically significant higher risk of poor outcome on the basis of clinical stage, grade, and PSA level. Table II shows the characteristics of patients treated by ADT according to treatment duration: standard duration (3 months) versus prolonged duration (5 months). No statistical difference was observed in the distribution of baseline characteristics between the 2 groups of patients. The decreasing PSA trend during the first 3 months under ADT was similar in the 2 groups. In men receiving ADT for the standard duration, the mean PSA levels were 15.3, 2.4, and 0.28 ng/mL, respectively at diagnosis, after 6 weeks and after 3 months of ADT. In men receiving prolonged ADT, the corresponding values were 14.7, 2.2, and 0.31 ng/mL. UROLOGY 58 (Supplement 2A), August 2001
TABLE II. Percentages of patients with selected characteristics at time of diagnosis according to duration of neoadjuvant androgen deprivation therapy (ADT) 3-month ADT (N ⴝ 129)
5-month ADT (N ⴝ 111)
P Value
34.1 39.5 26.4
36.0 35.1 28.8
0.8
10.1 83.7 6.2
9.9 78.4 11.7
0.3
33.3 48.8 10.9 7.0
20.7 55.9 15.3 8.1
0.2
17.1 41.8 24.0 17.1
12.6 46.0 20.7 20.7
0.6
Age (years) ⱕ59 60–64 ⱖ65 Clinical stage T1 T2 T3 Biopsy Gleason score 2–4 5,6 7 8–10 PSA (ng/mL) ⱕ3.9 4.0–9.9 10.0–19.9 ⱖ20.0
PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy.
TABLE III. Percentages of patients with pathological indicators of prognosis at the time of radical prostatectomy by treatment regimen: radical prostatectomy alone and neoadjuvant ADT of either standard or prolonged duration followed by radical prostatectomy RP 3-month ADT 5-month ADT P Value (N ⴝ 516) (N ⴝ 129) (N ⴝ 111) for Trend* Positive surgical margins Extracapsular extension Seminal vesicle invasion Lymph node invasion
48.7 51.9 15.5 13.6
29.5 41.1 19.4 11.6
12.6 21.6 7.2 8.1
0.0001 0.0001 0.004 0.002
ADT ⫽ androgen deprivation therapy, RP ⫽ radical prostatectomy. * P values for trend were obtained from logistic regression models including age, clinical stage, biopsy Gleason score, and initial PSA level.
In this group, the average PSA level after 5 months of ADT was 0.25 ng/mL. The effect of ADT and ADT duration on pathologic indicators of prognosis at the time of RP was examined. The proportions of patients with positive surgical margins, extracapsular extension, seminal vesicle invasion, and lymph node invasion were compared across the 3 treatment groups. After controlling for age, clinical stage, Gleason score, and initial PSA, there was a statistically significant trend for each of these indicators, with higher proportions in men treated by RP alone and lower proportions in those treated by prolonged ADT followed by RP (Table III). The median duration of follow-up was 4 years, UROLOGY 58 (Supplement 2A), August 2001
with 33% of the patients being observed for ⬎5 years and 3% for ⬎8 years. Among the 756 patients in the study population, 185 had PSA failure during the follow-up period. Overall, the 5-year biochemical disease–free survival rate was 70.6%. Altogether, 15 men had a bilateral orchiectomy immediately after RP without evidence of PSA failure. Of these men, only 1 had received neoadjuvant ADT. The decision to perform a surgical castration was based mostly on pathologic findings at the time of RP (pT3N⫹). Figure 1 shows the proportion of men without PSA failure during the first 8 years of the follow-up period among men treated by RP alone and among those treated by neoadjuvant ADT of either standard or prolonged duration 73
followed by RP. During the first year of the follow-up period, biochemical disease–free survival was higher for men who received ADT than for those who did not. From the third year after RP and thereafter, the crude biochemical disease–free survival was better for men who received prolonged neoadjuvant ADT than for those in the 2 other treatment groups. However, the validity of the crude analysis of treatment efficacy by KaplanMeier survival curves is questionable, because men treated by ADT had a worse risk profile at the time of diagnosis, as documented in Table I. The results from the multivariate Cox regression analysis confirmed the independent predictive capacity of clinical stage, biopsy grade, and initial PSA level on the risk of biochemical failure (Table IV). After controlling for these known determinants of disease progression, the HR associated with neoadjuvant ADT of a standard 3-months’ duration was 1.01, CI: 0.70 to 1.45, indicating that neoadjuvant ADT of 3-months’ duration did not influence PSA failure rate. In contrast, the HR associated with neoadjuvant ADT of prolonged duration was 0.60, CI: 0.38 to 0.94 — clearly suggesting a beneficial effect of prolonged neoadjuvant ADT on disease-free survival. A total of 43 patients received adjuvant external-beam radiation therapy and 112 received adjuvant ADT. These treatments were not associated with the risk of biochemical failure in the multivariate Cox models: HR of 0.82, CI: 0.46 to 1.47 for radiation therapy and HR of 1.27, CI: 0.87 to 1.86 for adjuvant ADT. When the regression model included adjuvant therapies, the association between neoadjuvant ADT and PSA failure remained similar. To study the association between neoadjuvant ADT (of either standard or prolonged duration) and biochemical disease–free survival cumulatively from RP until specific times in the follow-up period, multivariate Cox regression models, similar to that presented in the main analysis (Table IV) were used, censoring the observations at each yearly anniversary after RP. The HRs of PSA failure are presented in Figure 2. As anticipated, patients who had received neoadjuvant ADT had a lower risk of PSA failure in the first year after RP than those treated by surgery alone. Two years after RP and for the rest of the follow-up period, men treated by 3-month neoadjuvant ADT had a risk of biochemical failure similar to that of men treated by surgery alone. In contrast, for men who received 5-month ADT, the inverse association with the risk of PSA failure became statistically significant from the third year and forward, reached its greatest magnitude after 4 years (HR ⫽ 0.53, CI: 0.31 to 0.89), and was still present 8 years after RP (HR ⫽ 0.60, CI: 0.38 to 0.94). Other multivariate Cox models were used to as74
FIGURE 1. Kaplan-Meier curves for disease-free survival until PSA failure according to treatment regimen: radical prostatectomy alone (RP) and neoadjuvant ADT of either standard (3mADT) or prolonged (5mADT) duration followed by radical prostatectomy. Log-rank test P ⫽ 0.32. 3m ⫽ 3 months; 5m ⫽ 5 months; ADT ⫽ androgen deprivation therapy; PSA ⫽ prostate-specific antigen; RP ⫽ radical prostatectomy.
TABLE IV. Hazard ratios of biochemical failure and 95% confidence intervals associated with known indicators of prognosis and treatment regimen
Age (per 1-year increment) Clinical stage T1† T2 T3 Biopsy Gleason score 2–4† 5,6 7 8–10 PSA (ng/mL) ⱕ3.9† 4.0–9.9 10.0–19.9 ⱖ20.0 Treatment regimen RP alone† 3-month ADT 5-month ADT
Hazard Ratio
95% Confidence Interval*
1.01
0.98–1.04
1.00 1.37 2.11
0.83–2.28 1.05–4.25
1.00 1.61 3.43 3.06
1.09–2.37 2.21–5.34 1.74–5.40
1.00 1.79 2.83 4.45
0.91–3.49 1.43–5.58 2.22–8.96
1.00 1.01 0.60
0.70–1.45 0.38–0.94
ADT ⫽ androgen deprivation therapy, PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy. * Hazard ratios and confidence intervals were obtained from Cox regression models including simultaneously all variables in the table. † Reference category.
sess the HR of PSA failure associated with prolonged ADT during subsequent follow-up time periods (Figure 3). The effect of prolonged neoadjuvant ADT on biochemical disease–free surUROLOGY 58 (Supplement 2A), August 2001
FIGURE 2. Hazard ratios of PSA failure until each yearly anniversary after RP associated with neoadjuvant ADT of either standard or prolonged duration with no ADT before RP as the reference category. The multivariate Cox regression models included age, clinical stage, biopsy Gleason score, and initial PSA level. The vertical bars represent the 95% confidence intervals associated with the HRs for neoadjuvant ADT of prolonged duration. 3m ⫽ 3 months; 5m ⫽ 5 months; ADT ⫽ androgen deprivation therapy; PSA ⫽ prostatespecific antigen; RP ⫽ radical prostatectomy.
FIGURE 3. Hazard ratios of PSA failure during subsequent time periods in the follow-up after RP associated with neoadjuvant ADT of prolonged duration using RP alone as the reference category. The multivariate Cox regression models included age, clinical stage, biopsy Gleason score, and initial PSA level. Because the number of events was small, years 6, 7, and 8 were considered together. ADT ⫽ androgen deprivation therapy; PSA ⫽ prostate-specific antigen; RP ⫽ radical prostatectomy.
vival occurred mostly during the third year (HR ⫽ 0.22, CI: 0.05 to 0.95) and the fourth year after RP (HR ⫽ 0.26, CI: 0.05 to 1.31). Additional analyses were conducted to examine whether the association between prolonged ADT UROLOGY 58 (Supplement 2A), August 2001
and biochemical disease–free survival was stronger among subgroups of patients. We further investigated the relation between prolonged neoadjuvant ADT and biochemical disease–free survival, by applying the same multivariate Cox regression models to subgroups of the study population formed on the basis of baseline risk factors. Table V presents the HRs for PSA failure for men treated by prolonged neoadjuvant ADT in comparison with men treated by RP alone by category of age, clinical stage, Gleason score, and initial PSA level. This analysis suggests that the potential beneficial effect of prolonged neoadjuvant ADT was relatively less important for patients in the highest category of risk for all variables considered. In particular, there was significant trend for tumor grade (P ⫽ 0.04) showing that the magnitude of the association decreased with increasing risk. DISCUSSION This is 1 of few reports on the effect of duration of neoadjuvant ADT on biochemical failure rates after RP. We observed that patients who received a combined neoadjuvant ADT for ⬎3 months before RP had a statistically significant biochemical survival advantage compared with those treated by RP alone or with those who received a similar treatment for ⱕ3 months. This effect was progressive, became fully established 4 years after RP, and was still present 8 years after RP. Our study has many strengths, including its large size, the relatively long follow-up period, and the multivariate analysis used in the assessment of treatment regimens and disease-free survival. The main concern in interpreting the results of our cohort study comes from the difference in baseline risk between patients treated by RP alone and those who received neoadjuvant ADT of either standard or prolonged duration followed by RP. The decision to initiate ADT was influenced by the severity and aggressiveness of prostate cancer as judged by clinical stage, biopsy grade, and initial PSA. Men at higher risk of progression were more likely to receive ADT. This suggests that clinicians believed in the efficacy of neoadjuvant ADT, and it underscores the need of multivariate analysis to control confounding by baseline level of risk in the assessment of the effect of ADT on disease-free survival. On the other hand, there was no association between clinical stage, grade, or PSA level and the duration of ADT, suggesting that the decision to prolong ADT for ⬎3 months was not related to risk of PSA failure inferred from patient and tumor characteristics at the time of diagnosis. Because the decreasing PSA trend under ADT was similar in the 2 groups, it is unlikely that the decision to prolong ADT was based on patients’ drug tolerance or re75
TABLE V. Hazard ratios of biochemical failure and 95% confidence intervals associated with prolonged neoadjuvant ADT by category of baseline indicators of prognosis
Entire study population Age (years) ⱕ64 ⱖ65 Clinical stage T1, T2 T3 Biopsy Gleason score 2–4 5,6 7 8–10 PSA (ng/mL) ⱕ9.9 10.0–19.9 ⱖ20.0
Hazard Ratio
95% Confidence Interval*
0.60
0.38–0.94
0.44 0.86
0.25–0.79 0.43–1.73
0.57 0.62
0.34–0.94 0.17–2.26
0.28 0.43 0.83 1.20
0.04–2.09 0.22–0.87 0.37–1.87 0.32–4.53
0.44 0.51 0.62
0.20–0.97 0.20–1.31 0.29–1.32
ADT ⫽ androgen deprivation therapy, PSA ⫽ prostate-specific antigen, RP ⫽ radical prostatectomy. * Hazard ratios and confidence intervals were obtained from Cox regression models restricted to men within a category of a given variable and including simultaneously all other variables in the table.
sponse to therapy. We systematically used multivariate statistical analyses to enhance the comparability of the 3 treatment groups by controlling the confounding effect of well-established determinants of the risk of progression: clinical stage, tumor grade, and PSA level. It is therefore unlikely that differences in the baseline risk profile have affected the results of the multivariate analyses. This study enabled us to examine the effect of neoadjuvant ADT of either standard or prolonged duration over the course of 8 years of follow-up after RP. Our results confirmed the well-documented effect of ADT on pathologic prognostic factors at the time of RP.1,2 After controlling for the confounding effect of clinical stage, grade, and initial PSA level, we observed that men who received ADT had reduced rates of positive margins, extracapsular extension, and positive lymph nodes, compared with men treated by RP alone. Furthermore, the longer the ADT duration, the lower was the risk of positive margins, extracapsular extension, and lymph node involvement. The better pathologic profile of men treated by prolonged ADT compared with those receiving a standard treatment of 3 months has also been observed in an ongoing randomized trial.13 The long-term prognostic significance of the pathologic changes observed after ADT is not clear, because most randomized trials comparing ADT of 76
3 months’ duration with RP alone have demonstrated a strong effect on pathologic indicators but have not shown a clear survival advantage for men treated by neoadjuvant ADT.3–9 In each of the 5 randomized trials3–5,7,8 where combined neoadjuvant ADT was given for 3 months before RP, the PSA failure rate was slightly lower for men in the neoadjuvant ADT arm than in the RP-alone arm. The difference in biochemical disease–free survival failed to reach statistical significance in each individual trial and even in a meta-analysis combining the data from these trials.12 Our data suggest that neoadjuvant ADT may have an immediate and transient effect without prognostic significance. The lower risk of PSA failure during the first year after RP among men treated by neoadjuvant ADT could be explained by both biologic and management factors: (1) the persisting influence of ADT during the months after its discontinuation could mask or delay biochemical failure; (2) treating surgeons made a management decision, based on the findings at the time of RP, which are affected by ADT (downstaging), and they proceeded to a bilateral orchiectomy without waiting for the occurrence of PSA failure. Because we categorized these patients as experiencing PSA failure, the hazard of biochemical failure among men treated by ADT was artificially lowered. Both of these explanations appear to have played a role in our study population. On the other hand, the results of this study strongly suggest that neoadjuvant ADT has a delayed, persistent, and real effect on biochemical disease–free survival. This effect is observed only among men treated by neoadjuvant ADT for a prolonged duration. It is clearly observed during the third and fourth year after RP, and it has long-lasting consequences on overall biochemical disease–free survival up to 8 years after RP. We reported previously that patients treated by prolonged ADT— combining an LHRH analog and an antiandrogen— had an improved biochemical disease–free survival rate compared with those treated by RP alone or by ADT of standard duration before RP.12 In the present article, we identify the effect of timing of ADT on biochemical disease–free survival during the course of follow-up after RP. As neoadjuvant hormone therapy for prostate cancer delays the time of surgical radical treatment, prolonged ADT would be justified if a significant survival advantage was obtained for patients with poor prognosis as inferred from grade, stage, and PSA at the time of diagnosis. Our data suggest that, in relative terms, less benefit is observed for patients at higher risk of biochemical failure. In particular, patients with a high Gleason score seem to benefit less from prolonged ADT than other patients. If other studies confirm this observation, it UROLOGY 58 (Supplement 2A), August 2001
would probably weigh against the use of prolonged ADT before RP. CONCLUSION Neoadjuvant ADT before RP had been adopted with enthusiasm after the demonstration of its effect on pathologic indicators at the time of RP. After the publication of the initial survival data from the randomized trials addressing the longterm efficacy of ADT, this management strategy has fallen into disuse. Our results suggest that neoadjuvant ADT has a real beneficial effect on biochemical disease–free survival among patients treated for prostate cancer by RP. This effect appears to be delayed and is observed only when ADT is prolonged for ⬎3 months. However, the benefit of prolonged neoadjuvant ADT could be demonstrated only in a randomized, controlled trial, such as the trial underway. REFERENCES 1. Bonney WW, Schned AR, and Timberlake DS: Neoadjuvant androgen ablation for localized prostatic cancer: pathology methods, surgical endpoints and meta-analysis of randomized trials. J Urol 160: 1754 –1760, 1998. 2. Wieder JA, and Soloway MS: Incidence, etiology, location, prevention and treatment of positive surgical margins after radical prostatectomy for prostate cancer. J Urol 160: 299 –315, 1998. 3. Witjes WPJ, Schulman CC, Debruyne FMJ, for the European Study Group on Neoadjuvant Treatment of Prostate Cancer: Preliminary results of a prospective randomized study comparing radical prostatectomy versus radical prostatectomy associated with neoadjuvant hormonal combination therapy in T2–3 N0 M0 prostatic carcinoma. Urology 49(suppl 3A): 65– 69, 1997. 4. Aus G, Abrahamsson PA, Ahlgren G, et al: Hormonal treatment before radical prostatectomy: a 3-year follow-up. J Urol 159: 2013–2017, 1998. 5. Homma Y: Neoadjuvant androgen deprivation preced-
ing to radical prostatectomy: its role in short-term and longterm outcomes. Nippon Rinsho 58: 2162–2166, 1998. 6. Klotz LH, Goldenberg SL, Jewett M, et al: CUOG randomized trial of neoadjuvant androgen ablation before radical prostatectomy: 36 months post-treatment PSA results. Urology 53: 757–763, 1999. 7. Wildschutz T, Louis L, Hourriez L, et al: Neoadjuvant hormonal treatment prior to radical prostatectomy: follow-up of a prospective randomized study [abstract]. Eur Urol 30: 210, 1996, Abstract 774. 8. Soloway M, Sharifi R, Wajsman Z, et al, and the Lupron Depot Neoadjuvant Study Group: Radical prostatectomy alone vs radical prostatectomy preceded by androgen blockade in cT2b prostate cancer: 24 month results [abstract]. J Urol 157: 160, 1997, Abstract 619. 9. Baert LV, Goethuys HJ, De Ridder DJ, et al: Neoadjuvant treatment before radical prostatectomy decreases the number of positive margins in cT2–T3 but has no impact on PSA progression or survival [abstract]. J Urol 159: 61, 1998, Abstract 229. 10. Van der Kwast TH, Tetu B, Candas B, et al: Prolonged neoadjuvant combined androgen blockade leads to a further reduction of prostatic tumor volume: three versus six months of endocrine therapy. Urology 53: 523–529, 1999. 11. Gleave ME, Goldenberg SL, Jones EC, et al: Biochemical and pathological effects of 8 months of neoadjuvant androgen withdrawal therapy before radical prostatectomy in patients with clinically confined prostate cancer. J Urol 155: 213–219, 1996. 12. Meyer F, Moore L, Bairati I, et al: Neoadjuvant hormonal therapy before radical prostatectomy and risk of prostate specific antigen failure. J Urol 162: 2024 –2028, 1999. 13. Gleave M, Goldenberg SL, Warner J, et al: Randomized comparative study of 3 vs 8 months of neoadjuvant hormonal therapy prior to radical prostatectomy: biochemical and pathological effects [abstract]. J Urol 161: 154, 1999, Abstract 592. 14. Sobin LH, and Wittekind CH (Eds): International Union Against Cancer (UICC). TNM Classification of Malignant Tumours. 5th edition. New York, John Wiley and Sons, 1997. 15. Bostwick DG: Grading prostate cancer. Am J Clin Pathol 102(suppl 1): S38 –S56, 1994. 16. Clayton D, and Hills M: Statistical Models in Epidemiology. Oxford, Oxford University Press, 1993.
DISCUSSION FOLLOWING DR. FRANC¸OIS MEYER’S PRESENTATION Harry W. Daniell, MD (Redding, California): Your 5-month group actually had therapy from 3 months and 1 day to a year or some other longer period. Did you break that group up into shorter and longer subgroups to see if there was any relation? Franc¸ois Meyer, MD, DSc (Quebec City, Quebec, Canada): No. We felt that the numbers were not sufficient. It is not only the number of patients in each group, but it is also the number of failures that is important.
UROLOGY 58 (Supplement 2A), August 2001
Dr. Daniell: Was the division at 3 months made before the data were analyzed, or was that a cut-off point that was partly determined by the results that you got? Dr. Meyer: Both. Most of the randomized trials have 3 months of treatment in their protocols, so we considered 3 months as the reference, the standard duration of treatment. We also needed to find a cut point to have different groups according to duration. If we wanted to have groups of similar size, that would also be at 3 months.
77