Radical prostatectomy for clinically advanced (cT3) prostate cancer since the advent of prostate-specific antigen testing: 15-year outcome


John F. Ward, 620 John Paul Jones Circle, Portsmouth, VA 23708, USA.
e-mail: jfward@mar.med.navy.mil


In the first paper in this section, authors from the Mayo Clinic describe their experience and 15-year outcomes in the controversial subject of radical prostatectomy in patients with clinical T3 prostate cancer. The findings were interesting in many respects, but the authors concluded that radical prostatectomy as part of multimodal treatment for patients with clinical T3 disease offers cancer control and good survival rates.

There follows a series of papers on both prostate cancer and bladder cancer, but the final paper in this section from the UK attempts to define the accuracy of urologists and oncologists in assessing patient life-expectancy. Using various methods they found that, rather disappointingly, doctors were poor at predicting 10-year survival, leading to the possible outcome that some patients may be denied treatment after a pessimistic assessment of life-expectancy.


To report a long-term experience with extirpative surgery in patients presenting with locally advanced (cT3) prostate cancer, as the best management of such patients remains a problem.


In a single-institution retrospective study identifying 5652 men who had radical prostatectomy (RP) for histologically confirmed prostate cancer since the advent of prostate-specific antigen (PSA) testing (1987–97), 15% (842) had RP for cT3 disease. The median follow-up of these men was 10.3 years. Cancer-specific, overall and disease-free survival was plotted and compared with those of patients having RP for cT2 disease during the same period. Perioperative morbidity, continence and erectile function rates were examined, with a multivariate analysis for risk factors of disease recurrence.


Freedom from local or systemic disease at 5, 10, and 15 years after RP for cT3 disease was 85%, 73% and 67%; the respective cancer-specific survival rates were 95%, 90% and 79%. Significantly many men who did not receive neoadjuvant therapy (27%) were clinically over-staged (pT2) and most men with pT3 disease (78%) received adjuvant therapy. The mean time to adjuvant therapy after RP was not significantly different between men with cT3 and cT2 disease (4.0 and 4.3 years). Pathological grade (≥7), positive surgical margins, and nondiploid chromatin were all independently associated with a significant risk for clinical disease recurrence, while preoperative PSA level had little effect on outcome. Complications and continence rates after RP in patients with cT3 mirrored those in patients with cT2 disease.


Significantly many patients with cT3 prostate cancer are overstaged (pT2) in the PSA era. RP as part of a multimodal treatment strategy for patients with cT3 disease offers cancer control and survival rates approaching those achieved for cT2 disease. Pathological grade, ploidy and margin status are all significant predictors of outcome after RP. Complications and incontinence rates in patients with cT3 disease mirror those after RP for cT2 disease.


radical prostatectomy


cancer-specific survival


hormonal therapy


Radiation Therapy Oncology Group


The presentational characteristics of prostate cancer have changed dramatically within the USA in the last few years, coincident with the widespread use of PSA testing as a screening tool. At our institution, where we have advocated radical prostatectomy (RP) for patients with cT3 prostate cancer for over 20 years [1,2], the proportion of RPs in men with cT3 disease has declined significantly, from 25.3% in 1987 to 2.8% in 2001 (Fig. 1). While we hope that stage migration in this period has accounted for much of this decline, lingering biases about the best management of cT3 disease may also be skewing the referral of such patients. It was reported that when the effect of upstaging in patients undergoing RP is excluded from the Surveillance, Epidemiology and End Results database, the number of those presenting with cT3 disease has remained remarkably stable for 20 years [3].

Figure 1.

Clinical stage migration of 5652 newly diagnosed prostate cancers (T1c, green bars; T3-4, red) since the advent of PSA testing (1987–2001).

The best management of patients with clinically advanced prostate cancer remains controversial. At this stage the tumour appears to extend beyond the prostatic capsule, but distant metastases are not yet detectable. In the USA, surgery rates are 30% for patients with newly diagnosed cT1–2 disease, while only 6% of patients staged cT3 undergo RP [4]. Even in men with prolonged life-expectancy the RP rates are ≈ 67% for the youngest with cT1–2 prostate cancer but only 19% for the youngest with cT3 disease [4]. Thus we present our retrospective single-institution experience with RP as primary therapy for patients with cT3 prostate cancer.


Conduct of this study was approved by the Institutional Review Board (♯1989–02) and constitutes a minimal-risk investigation. Consent for the use of medical records for research was obtained from all patients before starting this analysis. A single-institution retrospective study was conducted using the referral-based longitudinal Mayo Clinic Prostate Cancer Registry. Men (5662) who had RP with pelvic lymph node dissection during the 11-year (1987–97) period since the advent of PSA testing were identified. Clinical staging consisted of a DRE by two clinicians, and defined using the 1997 American Joint Committee on Cancer guidelines. Bone scintigraphy, CT of the pelvis and cysto-urethroscopy before RP was discretionary. No patient with confirmed distant metastasis underwent RP.

Locally advanced prostate cancer (cT3) with no distant metastasis, pelvic side-wall extension or involvement of the trigone was present in 841 (15%) of the identified patients. The clinical characteristics of all patients in the study period (cT2 and cT3) are presented in Table 1. There was a significant difference between these groups in preoperative PSA and biopsy grade. Neoadjuvant therapy was given to 23% and 5% of those staged cT3 and cT2, respectively.

Table 1.  Comparison of all patients undergoing RP during the 11-year investigation period
VariableMedian (25–75th percentile)
Age at RP, years  66 (61–70) 66 (61–70)
Pre-op PSA, ng/mL   7.2 (4.4–11.9) 10.2 (4.7–23.7)
Clinical grade   6 (5–7)  7 (6–7)
Clinical grade ≥ 7, %  31 54
Neoadjuvant therapy   5 23

The excised prostate glands were evaluated at the time of surgery by a standardized, limited-sampling protocol using frozen-section techniques. The following day the prostates were evaluated using haematoxylin and eosin-stained permanent sections. The apex and base of the prostate were amputated and submitted as en fasce margins, followed by serial sectioning perpendicular to the long axis of the gland from apex to the tip of the seminal vesicles. On average, 14 prostate blocks were examined per patient. Pathological extraprostatic extension was defined as tumour extending into extraprostatic tissue (pT3). The primary tumour was graded according to the Gleason system. The lymph nodes were totally embedded for histological evaluation.

The median (range) follow-up was 10.3 (0.1–16.7) years. To date, adjuvant (≤90 days after RP) or salvage (>90 days after RP) therapies were administered to 48% and 41% of patients with cT3, and 21% and 22% of patients with cT2 disease.

Clinical progress was assessed at regular intervals either at our institution or by the referring physician, from the time of surgery throughout the follow-up. Clinical failure was defined as a serum PSA level of ≥0.4 ng/mL after RP, or demonstrable metastatic disease or local disease, or the initiation of salvage therapy (radiotherapy or hormonal therapy, HT) >90 days after RP. Disease outcome, perioperative and late treatment complications were retrieved by extensive record review and maintained within the registry.

The primary endpoints were times to death, prostate cancer death, clinical recurrence and biochemical failure (defined as a PSA of ≥ 0.4 ng/mL). Survival curves were generated using the method of Kaplan and Meier. Univariate and multivariate assessment of survival associations was conducted using the log-rank test and Cox proportional hazard models.


The characteristics of the extirpated prostates are detailed in Table 2. Of the 661 men with cT3 disease who did not receive neoadjuvant HT, 27% were clinically over-staged, harbouring organ-confined prostate cancer (pT2). Nodal metastases (TxN+) were present in 27% of patients (37% after neoadjuvant HT, 27% without). The biopsy grade reviewed at the Mayo concurred with pathological grade in all but 6% of patients who had higher pathological grade disease. There was a significant difference in the chromatin content between patients with pT2 and pT3 disease (diploid, tetraploid, aneuploid, 71%, 22%, 7% vs 51%, 35%, 14%, respectively).

Table 2.  Histological characterization of extirpated prostates for cT3 prostate cancer
GroupAllHT before surgery, % (n)P
Pathological stage (841)   0.064
pT227 (223)27 (174)27 (49)
pT3/446 (391)49 (326)36 (65)
TxN+27 (227)24 (161)37 (66)
Pathological Gleason score (738)   0.037
≤641 (305)42 (249)38 (56)
741 (305)42 (250)37 (55)
≥818 (128)15 (91)25 (37)
Chromosome content (816)   0.022
Diploid51 (420)53 (332)46 (88)
Tetraploid35 (284)35 (218)35 (66)
Aneuploid14 (112)12 (76)19 (36)
Surgical margins   0.97
Positive56 (471)56 (361)56 (110)
Negative44 (371)44 (284)44 (87)

Neoadjuvant HT was administered to 21% of patients (Table 2); rates of organ confinement (27%) and negative surgical margins (44%) were identical regardless of the preoperative administration of HT. Patients receiving neoadjuvant HT had a higher rate of pathological Gleason ≥ 8 (25% vs 15%), N+ disease (37% vs 24%) and aneuploid DNA content (19% vs 12%) than patients not receiving HT.

The perioperative morbidity in patients with cT3 disease (Table 3) was similar to that previously reported for patients with cT2 disease undergoing RP at our institution [5]. There was a parallel decrease in hospitalized blood transfusions in patients with cT2 or cT3 over the study period. After RP, 75% of the reporting patients had no erectile function, reflecting the infrequent use of a nerve-sparing technique (12% bilateral, 14% unilateral nerve preservation, 74% wide excision of both neurovascular bundles) [6]. Urinary continence (completely dry or security pad seldom moist) at 1 year was achieved in 79% of men staged cT3 (84% cT2), with few (6%) patients having severe incontinence (≥2 pads/day) and 0.5% requiring an artificial urinary sphincter.

Table 3.  Morbidity associated with RP in patients with cT3 disease
ComplicationRate, % (n)
Rectal injury 1.6 (14)
Intraoperative haemorrhage 1.8 (15)
Hospitalized blood transfusion29.0 (241)
Hernia 2.6 (22)
Bladder neck contracture 11.2 (93)
Lymphocele 1.0 (8)
Urethral stricture 3.2 (27)
Deep vein thrombosis 8.0 (15)
Pulmonary embolism 1.2 (10)
Erectile dysfunction75.3 (532 of 706 known)

Of patients staged cT3 and cT2, 78% and 41% received HT, radiotherapy or both at some point after RP (Table 4). There was no significant difference in the time (mean, median) to initiate secondary therapy between cT3 (4.0, 3.5 years) and cT2 (4.3, 3.5 years).

Table 4.  Adjuvant and salvage therapies administered to patients with cT3 prostate cancer after RP, segregated by pathological stage
Pathological stage (n)HT, n (%)Radiotherapy, n (%)
  • *

    1997 TNM Revision.

  • Adjuvant therapies were initiated ≤ 90 days after RP.

  • Salvage therapies were initiated > 90 days after RP.

  • ¶5.2% of cT3 patients received both HT and radiotherapy at some time after RP.

  • §

    §13.2% of cT3 patients received adjunctive therapy before and after RP.

T2N0 (223) 64 (28.7) 40 (17.9) 24 (10.8) 27 (12.1)
T3/4N0 (391)149 (38.1)128 (32.7) 73 (18.7) 75 (19.2)
TxN+ (227)216 (95.2) 51 (22.5) 36 (15.9) 23 (10.1)
Total (841)§429 (51.0)219 (26.0)133 (15.8)125 (14.9)

At 5, 10 and 15 years after RP for cT3 disease, 85%, 73% and 67% of patients were free of local or systemic disease recurrence. Figure 2A compares this outcome with patients undergoing RP during the same period for cT2 disease. Freedom from biochemical recurrence for cT3 and cT2 at 5, 10 and 15 years had a similar relationship (58%, 43% and 38% for cT3, vs 74%, 61% and 52% for cT2). The overall (90%, 76%, 53%) and cancer-specific survival (CSS) (95%, 90%, 79%) for patients with cT3 disease at 5, 10 and 15 years was only moderately lower than that in patients with cT2 disease (95%, 82% and 61%, and 99, 96% and 92%, respectively) during the same period (Fig. 2B). Stratified by pathological stage (Fig. 3), patients staged cT3 had a statistically significantly different CSS, but those staged pT3/4 had a 10- and 15-year CSS of 89% and 80% with adjuvant therapy.

Figure 2.

A, The time free of local or systemic failure and B, CSS (segregated by clinical stage; cT2 green; cT3, red) after RP for patients with prostate cancer.

Figure 3.

The CSS for patients with cT3 disease segregated by final pathological stage (green solid line T2N0; red dotted line, T3-4N0, and light green dashed line, TxN+).

The multivariate analysis of risk factors for clinical disease recurrence after RP for cT3 prostate cancer is shown in Table 5. Pathological grade (≥7), positive surgical margins and nondiploid chromatin content were all independently associated with a significant risk of clinical recurrence, while preoperative PSA level had little impact on this endpoint.

Table 5.  Multivariate analysis with hazard rates for clinical disease recurrence after RP in patients with cT3 prostate cancer (systemic or local disease)
VariableHazard ratio (95% CI)P
Pathological grade ≥ 71.27 (1.03–1.56)0.026
Pre-op PSA (doubling)1.13 (0.96–1.34)0.154
Ploidy (non-diploid)1.85 (1.18–2.91)0.008
Positive surgical margin1.77 (1.12–2.79)0.015
Seminal vesicle invasion1.49 (0.92–2.43)0.108


Within this cohort of patients presenting to a single referral centre with cT3 prostate cancer, clinical over-staging occurred in a significant proportion of hormone-naïve patients (27%); for these men, monotherapy with RP was potentially curative. Consistent with previous studies of the effect of neoadjuvant HT in patients with cT3 prostate cancer, HT given to 21% of the study cohort had little effect on grade, stage or rates of margin positivity, and did not influence progression-free or CSS. The morbidity of RP in these patients was no greater than that reported by us and others with RP for cT2 prostate cancer. However, the maintenance of erectile function was low in the present men (25% whose status was known), reflecting the wide resection of one or both neurovascular bundles in 88% of patients. Nonetheless, this erectile dysfunction rate compares favourably to those after RT for cT3 disease. For patients with cT3 disease, RP was part of a multimodal approach to disease eradication or control, which included HT or radiotherapy at some time after RP in 58% and 27%, respectively. Interestingly, the median time from surgery to secondary therapy for patients with cT3 disease was not significantly different from those with cT2. This probably reflects the biology of prostate cancer at the time of RP, regardless of clinical stage. As part of this multimodal approach to patients who once felt doomed to die from the disease, RP achieved CSS rates which approach those in patients with cT2 disease undergoing RP (90% vs 96% at 10 years and 79% vs 92% at 15 years).

This series represents the largest single-institution experience of the surgical management of cT3 prostate cancer, with a long-term follow-up (15 years) unmatched by any other therapy. Eliminating the prostate reduces the potential for late dissemination of radioresistant prostate cancer cells and simplifies the use of serum PSA levels in the follow-up. For a quarter of patients who are over-staged clinically, it eliminates the overtreatment of organ-confined disease with combined hormonal and radiotherapy, which is the current standard treatment for cT3 disease. With close follow-up and serial PSA measurements, the remaining patients harbouring pT3 disease may avoid castration and the deleterious effect that this has on quality of life, until the PSA becomes detectable. On the other hand, patients with N+ disease, which unless bulky is difficult to detect with modern imaging techniques, can initiate early HT, which has been found to improve survival over delayed HT in patients with N+ disease [7].

Opponents of surgical treatment have cited a lack of benefit if the prostate is not completely excised [8], an increased incidence of micrometastasis [1], and increased surgical morbidity [9]. Wide-field irradiation has therefore become the standard accepted treatment. However, as a monotherapy, radiotherapy has had limited long-term success.

Prostate biopsy studies after radiotherapy showed persistent prostate cancer in 14–91% of patients [10,11]. Coen et al.[12] evaluated 1469 men with biopsy-confirmed prostate cancer treated with radiotherapy, and found not only an independent association between delayed metastasis and the local persistence of the cancer on biopsy, but also a temporally increasing hazard rate. They postulated that a biologically altered prostate cancer after radiotherapy resulted in a late wave of metastatic seeding, possibly worsening the outcome.

To improve the problem of local control with radiotherapy, radiotherapists have used a multimodal approach (radiotherapy and HT) for treating cT3 prostate cancer. Laverdière et al.[13] conducted a prospective study of patients with cT2-cT4 prostate cancer, randomly assigning patients to one of three treatment arms, i.e. radiotherapy alone, neoadjuvant HT and radiotherapy, or neoadjuvant HT + radiotherapy + adjuvant HT. As a monotherapy, radiotherapy inadequately controlled the cancer, with two-thirds of prostate biopsies 2 years after treatment positive for residual prostate cancer. However, with HT, the rate of positive biopsies improved to less than a third for neoadjuvant HT + radiotherapy, and <5% in the three-treatment arm.

Bolla et al.[14] conducted a prospective, randomized trial comparing radiotherapy alone vs radiotherapy + 3 years of adjuvant HT in men with advanced prostate cancer (367 cT3/cT4, 34 T1/T2 Grade 3, 14 N+). With a median follow-up of 5.5 years, the combined arm had better clinical disease-free (74% vs 40%) and overall survival (78% vs 62%, P < 0.001) than the radiotherapy-only arm. The clinical disease-free survival rate prospectively achieved for combined treatment compares with the 85% rate in the present retrospective study of RP and secondary therapy for cT3 disease.

Finally, the Radiation Therapy Oncology Group (RTOG 86–10) reported the results of a 15-year, prospective, randomized investigation of radiotherapy vs radiotherapy with 4 months of adjuvant HT in 471 men with cT3 prostate cancer (± lymph node involvement) [15]. At a median follow-up of 6.7 years, patients with Gleason score ≤ 6 prostate cancer receiving combined therapy had better local and distant disease control. However, there was no significant advantage to combined therapy (locoregional, distant metastasis or survival) in patients with Gleason 7–10 carcinomas. While the cancer grading for the RTOG study was based on a review of needle-biopsy specimens, we found a concordance between final pathological Gleason score and biopsy Gleason score in the present surgically treated patients (6% up-graded). With nearly two-thirds of the present patients with cT3 having Gleason ≥ 7 disease, the findings of RTOG 86–10 are worrisome.

The use and outcome of neoadjuvant HT followed by surgery has been examined prospectively but with significantly fewer patients than radiotherapy studies. The South-west Oncology Group (SWOG-9109) examined 55 patients staged cT3 who received 4-months of neoadjuvant total androgen blockade (goserelin + flutamide) with RP [16]. The 5-year progression-free and CSS estimate in this group, reported at a median follow-up of 6.1 years, was 70% and 90%, respectively. Similarly, Gleave et al.[17] found a 75% progression-free survival in such patients treated with 8-months of total androgen blockade followed by RP. Because of the infrequent use of neoadjuvant and frequent use of adjuvant HT within the present study we were unable to discern an effect of neoadjuvant HT. However, neoadjuvant HT did not affect the surgical margin status, which, with tumour ploidy, was the most significant predictor of clinical disease recurrence.

The optimum treatment strategy for high-risk/locally advanced prostate cancer remains unknown. However, definitive local treatments as monotherapy cure only a minority of patients in both radiotherapy and RP series. Multimodal therapy which includes androgen suppression and RP or radiotherapy clearly improves the outcome in men with locally advanced prostate cancer. Although the optimum timing of HT is not clear, short courses appear inferior to long-term therapy [18]. However, the effects of HT on quality of life are not insignificant. Over a quarter of the present men (pT2) would have been unnecessarily exposed to the adverse effects of HT if empirically treated. Another two-thirds of cT3 prostate cancers examined contained grade ≥ 7 cancer, a pattern for which radiotherapy ± HT regimens were less effective.

While HT has been the mainstay of combined therapy, recent efforts have focused on the delivery of chemotherapy and/or HT around RP. Konety et al.[19] conducted a phase I/II study evaluating 36 patients with high-risk prostate cancer (cT3/4, cT1/2 with Gleason 8–10 and/or PSA >20 ng/mL) who received neoadjuvant HT and four cycles of paclitaxel/carboplatin/estramustine before RP. Although the effect of this therapy on clinical progression-free survival is not yet known, the morbidity of such therapy was low and the positive surgical margin rate lower than in the present study (22% vs 56%).

The Cancer and Leukaemia Group B is conducting a prospective trial (CALGB-90203) comparing RP with estramustine and docetaxel before RP [20]. An ambitious accrual goal of 700 men during a 48-month period has been set. The primary study endpoint is a decrease in 5-year recurrence rates, with secondary outcomes comparing safety, tolerability and the impact of neoadjuvant therapy on the pathological specimen. Results of this study are not expected until after 2011.

While long-term progression-free and CSS is the ultimate goal of any treatment strategy, preventing the significant morbidity of local prostate cancer progression (bleeding, urethral/ureteric obstruction, pain) is also necessary. Within the present patients, none had any symptoms associated with local tumour recurrence or progression. Tomlinson et al.[21] assessed patients with newly diagnosed cT3 prostate cancer treated with either extirpative (perineal RP, 24) or less than extirpative surgery (TURP, 26) or simple RP (two). Non-extirpative surgery failed to ameliorate the local morbidity of in-situ tumour progression and 75% later developed BOO. Also, ureteric obstruction (40%), infection (80%) and gross haematuria (45%) were more frequent in the nonextirpative groups than in those patients undergoing RP (4%, 26%, and 9%, respectively).

Finally, a quarter of the present patients with cT3 were N+; we have long advocated the early introduction of HT for this stage of disease [7,22]. However, the ability to closely monitor serum PSA after RP for pT3 prostate cancer, with our increased awareness of the quality-of-life issues surrounding HT, means that we now closely observe patients with pT3N0 disease and initiate salvage therapy only after the patient has had a significant PSA doubling time after RP (<1 year) or clinical evidence of disease recurrence [23]. This is identical to our practice for patients with pT2N0 disease.

As presented here, the CSS and overall survival was excellent using this strategy in patients once considered incurable. However, this report has many limitations, beginning with its retrospective view, and an uncontrolled bias for the initiation and timing of adjuvant therapy before and/or after surgery. Second, as a tertiary centre, there was probably referral-pattern selection bias, which prevents an assessment of all men with newly diagnosed cT3 prostate cancer. Clinical trials comparing surgery or radiotherapy as part of a multimodal treatment regimen which includes HT and/or chemotherapy are needed. Until then, the findings of this and other surgical series suggest that RP when combined with adjuvant HT compares favourably with current radiotherapy/HT strategies.


None declared.