Impact of prostate-specific antigen testing on the clinical and pathological outcomes after radical prostatectomy for Gleason 8–10 cancers

Authors


Michael L. Blute, 200 First Street SW, Rochester, MN 55905, USA.
e-mail: blute.michael@mayo.edu

Abstract

OBJECTIVE

To investigate whether the clinical and pathological outcomes after radical retropubic prostatectomy (RRP) have changed since the advent of prostate-specific antigen (PSA) testing for patients with Gleason 8–10 cancers.

PATIENTS AND METHODS

We identified 584 men treated with RRP between 1988 and 2001 for pathological Gleason 8–10 tumours. Patients were divided for analysis by year of surgery, i.e. early (1988–93), mid (1994–97) and late PSA era (1998–2001). Survival rates after RRP were estimated using the Kaplan-Meier method, and the effect of clinicopathological factors on outcome was analysed using Cox proportional hazard regression models.

RESULTS

The median preoperative PSA level decreased from 15 ng/mL in the early to 10 ng/mL in the late PSA era (P < 0.001), while the rate of organ-confined disease increased from 22.9% to 35.1% (P = 0.007). However, the 7-year biochemical recurrence-free (37% vs 45%, P = 0.087) and cancer-specific survival (89% to 91%, P = 0.73) did not change significantly from the early to the late PSA era. Increased preoperative PSA level (P < 0.001), seminal vesicle invasion (P < 0.001) and positive lymph nodes (P = 0.02) were associated with biochemical recurrence. Seminal vesicle invasion (P = 0.005), positive nodes (P < 0.001) and positive surgical margins (P = 0.03) predicted death from cancer.

CONCLUSION

Although the pathological features of Gleason 8–10 cancers have become more favourable over the PSA era, survival has not changed. This lack of improvement in clinical outcome probably reflects the inherent biological aggressiveness of these cancers. While RRP provides long-term cancer control in a subset of these patients, continued investigation of multi-modal treatment options is warranted.

Abbreviations
BCR

biochemical recurrence

RRP

radical retropubic prostatectomy

AHT

adjuvant hormonal therapy

RT

radiotherapy

CSS

cancer-specific survival

SVI

seminal vesicle invasion

HR

hazard ratio.

INTRODUCTION

The widespread use of PSA testing has affected the clinical and demographic characteristics of men with newly diagnosed prostate cancer. Specifically, patients present at a younger age, with a lower serum PSA level and a higher proportion of organ-confined disease [1]. However, the effect of PSA testing on patients defined as high risk by the classification of D’Amico et al.[2] has been less predictable. A recent multi-institutional study found that the outcome for patients with high-risk prostate cancer after radical retropubic prostatectomy (RRP) has not changed significantly over the PSA era [3]. This lack of improvement in survival for high-risk patients might be due in part to the fact that, while it has been shown that high-risk cancers are now detected at an earlier stage [3,4] and lower PSA level [3], nevertheless there are still men considered to have high-risk prostate cancer based on a high Gleason score, which has been consistently identified as an independent risk factor for death from prostate cancer after RRP [5–7].

At the same time, a recent investigation of prostate cancer treatment patterns from the Cancer of the Prostate Strategic Urological Research Endeavor disease registry, that is comprised primarily of community practices across the USA, showed a significant increase in the use of radiation therapy (RT) and androgen deprivation in patients with high-risk prostate cancer, and a corresponding decrease in surgical therapy [8]. Specifically, patients with Gleason 8–10 cancers were 1.9 times more likely to have external beam RT and 3.9 times more likely to receive primary androgen deprivation [8]. This practice pattern might be based in part on data from studies of patients with Gleason 8–10 disease from cohorts before and early in the PSA era, when the overall clinicopathological features of prostate cancer tended to be more advanced [1].

Thus we investigated whether the clinical and pathological outcomes of patients with Gleason 8–10 cancers treated with RRP at our institution have changed since the advent of PSA testing. In addition, we identified risk factors for disease progression in these tumours, which we think will help to optimize the individualized treatment of these patients.

PATIENTS AND METHODS

After Institutional Review Board approval was obtained, we reviewed the records of 9081 consecutive patients who had RRP and bilateral pelvic lymph node dissection between 1988 and 2001 at the Mayo Clinic. Men who received hormonal treatment or RT before RRP were excluded from the analysis. Patients with pathological Gleason 8–10 cancers were then identified. These patients were divided for analysis by year of RRP into early (RRP 1988–1993), mid (1994–97) and late PSA era (1998–2001).

The RRP was performed by different surgeons but using standardized techniques. The 1997 TNM classification system was used for staging, and the Gleason system was used for grading. DNA ploidy was assessed by flow cytometry. Adjuvant hormonal treatment (AHT) consisted of treatment received within 90 days of RRP, and was given at the discretion of the treating physician. Medical hormone deprivation therapy was generally intended to be life-long, but given the retrospective nature of the present study, it is uncertain whether patients discontinued treatment after a period of androgen deprivation.

Assessments before RRP included a physical examination and serum PSA measurement, quarterly for the initial 2 years, semi-annually for a further 2 years, and annually thereafter. Radiographic evaluation was used as indicated clinically. Biochemical recurrence (BCR) was defined here as a PSA level of ≥ 0.4 ng/mL. Local recurrence was defined as cancer detected on biopsy of the prostatic bed, or receipt of salvage RT to the prostatic bed with no evidence of systemic recurrence. Systemic progression involved detectable metastatic deposits on radio-isotopic bone scan or on biopsies other than biopsy of the prostatic bed. The cause of death was identified from death certificates or physician correspondence. For patients followed elsewhere, the Mayo Clinic Prostatectomy Registry monitors outcomes annually by correspondence.

The clinicopathological features from patients with Gleason 8–10 cancers in the early, mid and late PSA eras were compared using the chi-square and Wilcoxon rank-sum tests, as appropriate. Survival after RRP for patients with Gleason 8–10 cancers from each era was estimated using the Kaplan-Meier method, with patients censored at the last follow-up or death, if the endpoint of interest had not been attained. The effect of various clinicopathological factors on outcome for patients with Gleason 8–10 tumours was then analysed using Cox proportional hazard regression models; hazard ratios (HRs) and 95% CIs were estimated using the proportional hazards model. All tests were two-sided with P ≤ 0.05 considered to indicate significance.

RESULTS

We identified 584 patients with Gleason 8–10 disease at RRP between 1988 and 2001; the patient demographics, separated by PSA era, are listed in Table 1. The incidence of Gleason 8–10 disease did not change over time, from 6.4% (223/3507) of all RRPs in 1988–1993 to 6.2% (187/3028) in 1994–1997 and 6.8% (174/2546) in 1998–2001 (P = 0.59). However, Gleason 8–10 tumours from the late PSA era were associated with lower preoperative PSA levels, and were more likely to be clinically and pathologically organ-confined.

Table 1.  The patients’ demographics
FeatureEra of RRPP
1988–931994–971998–2001
No. of men223187174 
Median (interquartile range):
 age at surgery, years 67.0 (63,71) 67.0 (63,71) 66.0 (61,70)0.05
 preop PSA, ng/mL 15.0 (6.9,28.6)  9.4 (6.6,15.8) 10.0 (6.1,17.8)<0.001
n (%)
Clinical stage   <0.001
 T1 23 (10.3) 46 (24.7) 51 (30.2) 
 T2 141 (63.2)120 (64.5) 99 (58.6) 
 T3/4 59 (26.5) 20 (10.8) 19 (11.2) 
Pathological tumour stage   0.007
 T2 51 (22.9) 66 (35.3) 61 (35.1) 
 T3/4 72 (77.1) 121 (64.7) 113 (64.9) 
SVI   0.008
 No108 (48.4) 113 (60.4)109 (62.6) 
 Yes 115 (51.6) 74 (39.6) 65 (37.4) 
Positive surgical margin   0.002
 No 113 (50.7) 63 (33.7) 78 (44.8) 
 Yes 110 (49.3)124 (66.3) 96 (55.2) 
Lymph node metastases   <0.001
 No154 (69.1)155 (82.9)153 (87.9) 
 Yes 69 (30.9) 32 (17.1) 21 (12.1) 
Tumour DNA ploidy   0.13
 Diploid 97 (44.9) 90 (48.1) 65 (37.8) 
 Non-diploid 119 (55.1) 97 (51.9)107 (62.2) 

After RRP, 224/584 (38.4%) patients with Gleason 8–10 tumours received AHT. The use of AHT did not change over time, from 84/223 (37.7%) patients in the early to 70/187 (37.4%) in the mid and 70/174 (40.2%) in the late PSA era (P = 0.62). Patients treated with AHT had significantly worse clinicopathological features than those who did not receive AHT (Table 2). Moreover, 162 patients received salvage hormonal therapy, including 91 treated at BCR, and 71 treated at the time of systemic progression. In addition, 70/584 (12%) patients were treated with adjuvant RT, while 103 (17.6%) received salvage RT.

Table 2.  The clinicopathological characteristics of patients, by receipt of AHT
FeatureAHTNo AHTP
No. of men224360 
Median (interquartile range):
 age at surgery, years  66.0 (62–70) 67.0 (63–71)0.09
n (%):
Preoperative PSA, ng/mL  <0.001
 <10 82 (36.6)199 (55.3) 
 10–19.9 49 (21.9) 88 (24.4) 
 ≥20 93 (41.5) 73 (20.3) 
Pathological Gleason score  0.006
 8 94 (42)193 (53.6) 
 9–10130 (58)167 (46.4) 
Pathological tumour stage  <0.001
 T2 32 (14.3)146 (40.6) 
 T3/T4192 (85.7)214 (59.4) 
SVI  <0.001
 No 78 (34.8)252 (70) 
 Yes146 (65.2)108 (30) 
Positive nodes  <0.001
 No106 (47.3)356 (98.9) 
 Yes 118 (52.7)  4 (1.1) 
Positive surgical margin  <0.001
 No 53 (23.7)201 (55.8) 
 Yes 171 (76.3)159 (44.2) 
Tumour DNA ploidy  0.03
 Diploid 84 (38.2)187 (52.7) 
 Non-diploid136 (61.8)168 (47.3) 

At a median (interquartile range) follow-up after RRP of 8.3 (5.3–11.5) years, 332 patients with Gleason 8–10 disease had BCR, 89 had local recurrence, 133 relapsed systemically, and 206 died, with 87 dying from prostate cancer. Kaplan-Meier survival curves for these patients are shown in Fig. 1. The 7-year event-free survival for these patients was 42%, 86%, 82% and 89% for BCR, local recurrence, systemic progression and cancer death, respectively.

Figure 1.

The outcome after RRP for patients with Gleason 8–10 tumours.

Interestingly, although the associated pathological features of patients with Gleason 8–10 tumours improved over the PSA era (Table 1), the clinical outcome of these patients remained relatively stable over time, i.e. the 7-year BCR-free rate for the early, mid and late PSA era was 37%, 47% and 45%, respectively (P = 0.087; Fig. 2a). Similarly, the 7-year local recurrence-free survival for each group was 84%, 89% and 79% (P = 0.78; Fig. 2b), while the systemic progression-free survival was 82%, 81% and 79% (P = 0.93; Fig. 2c); the 7-year cancer-specific survival (CSS) for the early PSA era was 89%, vs 88% for mid and 91% for the late PSA era (P = 0.73; Fig. 2d).

Figure 2.

The BCR-free (A), local recurrence-free (B) systemic progression-free (C) and CSS (D) after RRP for patients with Gleason 8–10 tumours, according to the era of surgery.

We next examined the risk factors for disease progression in patients with Gleason 8–10 cancers. In a multivariate Cox model, elevated preoperative PSA level (P < 0.001), seminal vesicle invasion (SVI) (P < 0.001) and positive lymph nodes (P = 0.02) predicted BCR (Table 3). Meanwhile, SVI, nodal disease, and non-diploid tumour ploidy were associated with both systemic progression and death from prostate cancer for patients with Gleason 8–10 cancers. The presence of a positive surgical margin also predicted death from prostate cancer (P = 0.01). Use of AHT was associated with lower risks of PSA progression and local recurrence (both P < 0.001), but did not significantly affect systemic progression or CSS (Fig. 3).

Table 3.  Multivariate analysis of risk factors associated with outcomes after RRP for patients with Gleason 8–10 prostate cancer
Risk factorHR (95% CI), chi-square P
BCRLocal recurrenceSystemic progressionCSS
Age at RRP1.01 (0.99–1.02), 0.470.98 (0.94–1.0), 0.170.99 (0.96–1.02), 0.401.00 (0.96–1.03), 0.77
Era of surgery1.02 (0.76–1.30), 0.921.03 (0.56–1.9), 0.911.04 (0.62–1.74), 0.890.89 (0.43–1.85), 0.75
Preoperative PSA1.17 (1.07–1.3), <0.0011.08 (0.91–1.3), 0.380.92 (0.80–1.06), 0.230.90 (0.76–1.08), 0.25
Gleason score (9/10 vs 8)1.20 (0.96–1.5), 0.11.39 (0.9–2.2), 0.141.31 (0.92–1.87), 0.131.32 (0.85–2.1), 0.21
Extraprostatic extension1.14 (0.84–1.54), 0.411.02 (0.55–1.9), 0.941.35 (0.77–2.4), 0.29 1.65 (0.77–3.54), 0.20
SVI1.75 (1.3–2.4), <0.0011.38 (0.78–2.4), 0.272.20 (1.4–3.7), 0.0022.67 (1.34–5.31), 0.005
Lymph node metastases1.71 (1.1–2.7), 0.022.15 (0.83–5.6), 0.112.68 (1.34–5.38), 0.0054.60 (1.88–11.3), <0.001
Positive surgical margin1.12 (0.88–1.42), 0.370.82 (0.52–1.3), 0.411.38 (0.94–2.02), 0.101.72 (1.05–2.82), 0.03
Non-diploid (vs diploid)1.21 (0.97–1.5), 0.091.52 (0.97–2.4), 0.071.54 (1.06–2.21), 0.021.57 (0.99–2.47), 0.05
AHT (vs no AHT)0.30 (0.21–0.42), <0.0010.29 (0.14–0.64), 0.0020.64 (0.39–1.08), 0.090.65 (0.33–1.26), 0.20
Figure 3.

The outcome after RRP for patients with Gleason 8–10 tumours, according to receipt of AHT; (A) BCR-free survival; (B) local recurrence-free survival; (C) systemic progression-free survival; (D) CSS.

DISCUSSION

While the results for RRP in patients with Gleason 8–10 cancers were investigated previously [4,9–14], the studies to date involved relatively few patients, with a short or intermediate follow-up, and primarily reported outcome in terms of BCR. We present here what is, to our knowledge, the largest series to date of patients with Gleason 8–10 disease treated during the PSA era. In addition, with a median follow-up of 8.3 years, we report the effect of high Gleason grade on local recurrence, systemic progression and CSS. Our institution’s experience with RRP for high-grade prostate cancer was previously reported in patients in the early and mid-PSA era [15]. However, as PSA screening has affected the clinical and pathological features of patients diagnosed with prostate cancer [1], we have updated our experience, with additional follow-up, and including patients treated during what might be characterized as a more recent PSA era.

Our finding of an increase rate of organ-confined disease over the PSA era supports the results from a previous study of patients with Gleason 8–10 cancers by Mian et al.[11], who found that 45% of patients had organ-confined disease at RRP in 1995–1998, compared to 19% who had RRP in 1987–1994. Similarly, Perrotti et al.[4] reported a 30% rate of organ-confined tumours in patients who had RRP in 1992–1996 for tumours that were detected primarily through PSA screening. Despite a decreased incidence of locally advanced and metastatic tumours, we found that adverse tumour pathology, in the form of SVI and lymph node metastases, affected CSS. Similarly, Perrotti et al.[4] found that tumour stage predicted BCR in a study of 64 men with Gleason 8–10 tumours who had RRP, while Mian et al.[11] noted that pathological tumour stage was the most significant predictor of disease recurrence after RRP for Gleason 8–10 cancers.

At the same time, a study of patients with Gleason 8–10 tumours treated with RRP or RT, alone or combined, found that a pretreatment PSA level of >20 ng/mL and SVI predicted BCR, although only SVI predicted systemic failure [9]. While other investigators reported the effect of preoperative PSA level on BCR after RRP [12–16], in the present study an increased preoperative PSA level predicted BCR, but was not associated with systemic progression or CSS. Positive surgical margins were also significantly associated with death from prostate cancer for patients with Gleason 8–10 tumours. This was reported previously in patients with Gleason 8–10 disease [12,13,15].

Although the associated pathological features of Gleason 8–10 tumours improved from the early to late PSA era, nevertheless the patients’ outcomes after RRP did not significantly change over time. Indeed, while there was a slight trend to an improved BCR rate (P = 0.087) there were no differences in the rates of systemic progression or death from prostate cancer. There is an inverse relationship between the Gleason grade and the PSA content of prostate cancers [17], that might limit the effect of PSA testing on the outcome of these patients. Moreover, the lack of effect on the improvement in survival after surgery might be due in part to the presence of a high Gleason score reflecting an inherent biological aggressiveness of the cancer, independent of other pathological variables. Indeed, a recent multi-institutional investigation similarly reported that the outcome of high-risk prostate cancer has not changed significantly over the PSA era, further suggesting the inability of surgery alone to treat most such patients [3].

The potential value of a multi-modal approach to high-grade prostate cancer is suggested in the results of Do et al.[9], who found better 5-year rates of biochemical and clinical progression-free survival in patients with Gleason 8–10 tumours treated with RRP and RT (65% and 80%) than in patients treated with RRP (20% and 35%) or RT (30% and 60%) alone. Similarly, the addition of long-term AHT has been shown to improve locoregional control, CSS and overall survival in patients treated with RT for Gleason 8–10 tumours [18].

While the fact that nearly 40% of the present patients received AHT might be criticised as a confounder of the effect of RRP alone on high-grade prostate cancer, we assert that this reflects a practice not uncommon among urologists at present. Meng et al.[8] reported that a Gleason score of 8–10 was a significant predictor for receiving AHT in a multivariate analysis. In the present study AHT decreased the risk of BCR but did not significantly affect systemic progression or CSS. However, AHT was administered at the discretion of the treating physician. As such, patients who received AHT had significantly worse clinicopathological features than those who did not, which might have limited our ability to discern any effect of immediate androgen deprivation in high-grade cancers. In addition, an effect of AHT on systemic progression and CSS might become evident with a longer follow-up.

Comparisons of the present outcomes for patients with Gleason 8–10 cancers after RRP to published series are made difficult by the varied patient populations that have been studied, the different lengths of follow-up, and disparate definitions of failure. Nevertheless, our overall 7-year BCR-free rate of 42% in 584 patients with Gleason 8–10 cancers, of whom 122 (21%) had positive lymph nodes, compares favourably with the results from Mian et al.[11], who reported a 55% 7-year disease-free rate in 188 patients with Gleason 8–10 tumours, of whom only 6% had positive nodes. A separate investigation of 54 patients with pathological Gleason 8–10 cancers who did not receive AHT reported a 41% BCR rate at a median follow-up of 49 months [12]. At the same time, the present 89% 7-year overall CSS for patients with Gleason 8–10 cancers, of whom 65–77% from each PSA era had pT3/T4 disease, compares favourably with the 69% 10-year disease-specific survival reported by Sciarra et al.[13] in 41 patients with Gleason 8–10 pT3 tumours, none of whom received AHT.

Further studies, ideally in the setting of clinical trials, are needed to better define the role of androgen deprivation, potentially together with chemotherapeutic regimens [19], in men with Gleason 8–10 cancers. A future challenge will be to individualize the timing and selection of systemic treatments in patients with high-risk prostate cancer, based on an assessment of the individual patient’s risk for progression and likelihood of treatment response. For example, one potential predictor of patients’ response to AHT is DNA tumour ploidy [20]. In the present study men with non-diploid tumours had a greater risk of systemic progression and death from prostate cancer, suggesting that these patients might benefit from other systemic treatments, such as chemotherapy or investigational agents.

Although our study represents a retrospective, single-centre experience in which the treatment after RRP was not standardized, given the relative infrequency of Gleason 8–10 cancers, prospective randomized trials of these patients remain difficult. Therefore, large, retrospective series with a long-term follow-up remain an important tool for risk stratification and assessing treatment outcomes.

CONFLICT OF INTEREST

None declared.

Ancillary