World Journal of Urology

, Volume 29, Issue 5, pp 665–670

Pathological findings and oncological control afforded by radical prostatectomy in men with high-risk prostate cancer: a single-centre study

Authors

  • Alexandra Masson-Lecomte
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
  • Vincent Hupertan
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
  • Eva Comperat
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
  • Christophe Vaessen
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
  • Emmanuel Chartier-Kastler
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
  • Olivier Cussenot
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
  • Marc-Olivier Bitker
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
    • The Academic Department of Urology, Pathology and Statistics of La Pitié-SalpétrièreGroupe Hospitalo-Universitaire EST, Assistance-Publique Hôpitaux de Paris
    • Faculté de Médecine Pierre et Marie CurieUniversity Paris VI
    • CeRePP group, Centre d’Etudes et de Recherche sur les Pathologies Prostatiques
Original Article

DOI: 10.1007/s00345-010-0608-5

Cite this article as:
Masson-Lecomte, A., Hupertan, V., Comperat, E. et al. World J Urol (2011) 29: 665. doi:10.1007/s00345-010-0608-5

Abstract

Objective

To assess pathological findings and oncological control afforded by radical prostatectomy (RP) in high-risk prostate cancers (PCa) at our institution.

Materials and methods

We performed a retrospective review of prostate cancer patients who underwent RP between 1995 and 2006 for a high-risk prostate cancer (i.e., PSA >20 or biopsy Gleason ≥8 or clinical stage ≥T2c). Biochemical recurrence was defined as a single rise in PSA levels over 0.2 ng/ml after surgery. Survival curves were elaborated by the Kaplan–Meier method and Cox proportional hazard regression analysis. For each patient, a prognostic score for recurrence was estimated, and a prediction model was then constructed.

Results

Overall, 138 patients were included and followed for a median time of 53 months. Mean age at diagnosis was 63.4 years (range 39–80) and mean pre-operative PSA was 15.5 ng/ml (range 7.4–31). The median follow-up was 53 months (range 6–166). Overall, 82 patients (59%) had biochemical recurrence. The five-year PSA recurrence-free survival rate was 40%. In univariate analysis, clinically palpable tumours (T2–T3) (P = 0.032), biopsy Gleason score ≥8 (P = 0.031), seminal vesicle invasion (pT3b), positive margins and positive lymph nodes (P < 0.001) were significantly associated with recurrence. In multivariate analysis, the biopsy Gleason score ≥8, seminal vesicle invasion, positive margins and positive lymph nodes predicted recurrence (P < 0.05).

Conclusions

RP affords an acceptable oncological control at first-line treatment of selected patients with high-risk PCa. However, in certain cases, surgery alone might not be sufficient and may be part of a multimodal treatment including either adjuvant radiotherapy or androgen deprivation.

Keywords

Prostatic neoplasmsProstate-specific antigenOutcomeBiochemical recurrenceRadical prostatectomyGleason score

Introduction

Because of the widespread use of PSA, an increasing number of prostate cancers (PCa) with localized tumours and a low risk of progression have been diagnosed [1, 2]. Early diagnosis of tumours of low volume that are likely indolent sometimes leads to aggressive therapeutic use, which is not devoid of side effects and often leads to overtreatment [3, 4]. According to the D’Amico classification, high-risk PCas are defined by a PSA >20 ng/ml and/or a Gleason score on biopsy ≥8 and/or a clinical stage ≥T2c [5]. These patients are considered to be at high risk for treatment failure after radical prostatectomy (RP) or radiotherapy. External beam radiotherapy combined to hormonal therapy has progressively emerged as the gold standard treatment for locally advanced disease [6]. At the same time, the number of RPs for high-risk PCa has decreased from 34% in the early 1990s to 14% in recent years [7]. However, radiotherapy alone has never formally demonstrated its superiority over surgery in these cases [5, 8]. Few teams have even reported satisfactory oncological outcomes after surgery in high-risk PCa, even without adjuvant treatment [7, 9, 10]. These teams have shown a biochemical-free survival rate as high as 50% and a cancer-specific survival rate of up to 90% at 10 years. The objective of our study was to evaluate the oncological outcome after RP in our own population of high-risk PCa patients who have undergone surgery as a first-line treatment at our institution.

Materials and methods

Patient population

We retrospectively reviewed the data from 1,167 patients treated by RP at our institution between 1995 and 2006. We selected all PCa patients who were treated surgically and had a high-risk profile according to D’Amico’s classification (i.e., PSA >20 ng/ml and/or a Gleason score on biopsy ≥8 and/or a clinical stage ≥T2c) [5]. All patients had undergone prostatic biopsies because of an elevation of the PSA (normal level <4 ng/mL) and/or an abnormality discovered on a digital rectal examination (DRE). None of the patients received neoadjuvant treatment. We collected the following data: patient age at the time of surgery, clinical presentation, pre-operative PSA level, Gleason score (from biopsy and pathological specimen) and 2009 TNM stage, type of surgery (open or laparoscopic approach), pathological data, margin status, lymph node status and disease progression (either biochemical or clinical). Pre-operative evaluation always included abdominal CT to detect suspicious lymph nodes and a bone scan. Only a few patients from the current study had an MRI as part of the pre-operative work-up because MRI was introduced at our centre in 2003. The clinical stage was assigned by the attending surgeon according to pre-operative DRE findings. Patients were excluded from surgery if evidence was found of lymph node involvement or metastasis on the pre-operative work-up.

Operative technique

RP was proposed as a first-line curative treatment regardless of the Gleason score from the biopsies in the following cases: age <75 years and/or a strong desire by the patient to undergo surgery rather than external beam radiation as a first-line treatment. The patients had either open retropubic or laparoscopic RP, depending on the choice and on the experience of the surgeon. Two seasoned surgeons were involved in each of these procedures (i.e., >400 open RP cases and >250 laparoscopic RP). RP were performed according to a non-conservative technique of the neurovascular bundles and extensive removal of the tumour, as previously described [7, 10]. A systematic pelvic lymphadenectomy was performed in all cases included in the current study.

Pathological findings

RP specimens were fixed intact in 10% neutral buffered formalin and then sectioned transversally at regular intervals. Tumour differentiation was given by the surgical Gleason score: high (2–6), moderate (7) and poor (8–10) according to the WHO consensus [11]. During macroscopic examination (orientation, weight, size, description) of the surgical specimen, the entire surface of the RP specimen was inked to evaluate it in an accurate manner (right side green, left side black). When evaluating the RP specimen, we proceeded in the following manner: the prostate was fixed in 10% formalin for at least 24 h. Inclusion was according to the Stanford protocol [12]. The positive margin length was given by the greatest tumour length in millimetres reaching the ink in the histological sections. Surgical margins were positive when tumour foci were found by microscopy in the absence of macroscopically visible abnormalities. Lymph node involvement was detected by gross and microscopic examination. All slides were reviewed by one experienced uropathologist (EC).

Follow-up

All patients were seen three and six months after surgery and at least once a year thereafter. Each visit included a DRE and PSA measurement. The biochemical recurrence of the disease was defined as a single measurement of a PSA level greater than 0.2 ng/ml after RP. All patients with lymph node invasion received immediate hormone therapy. In other cases, the decision for adjuvant treatment was always discussed case by case.

Statistical analysis

The PSA-free survival was determined from the date of surgery to the last follow-up visit. Survival curves were established using the Kaplan–Meier method and compared by the log-rank test. Prognostic factors were established by univariate analysis and those that were significant were entered into multivariate analysis using the Cox stepwise regression method. For each patient, a prognostic score (predicted probability) for recurrence-free survival at five years was estimated, and a prediction model was then constructed. The discriminating ability of the model was assessed by Harrell’s concordance index (c-index) and graphical calibration. Bootstrapping was used to assess confidence intervals. A P < 0.05 was considered significant. All analyses were performed with R™ version 2.10.1 software with the Design and Hmisc libraries added.

Results

Population

Overall, 138 patients were included. The mean age at diagnosis was 63.4 ± 6.5 years (range 39–80). The mean pre-operative PSA value was 15.5 ± 12 (range 7.4–31) ng/ml. Fifty patients (36%) had a PSA level ≥20 ng/ml. In the clinical stage, 29 (21%) patients had T2 PCa, and 90 (66%) patients had T3 PCa. Regarding the biopsy Gleason score, only 19 patients (14%) had a score ≥8. The patients’ characteristics (clinical and pathological) are summarized in Table 1. Overall, 80% underwent an open RP (N = 110) and 20% underwent a laparoscopic RP (N = 28).
Table 1

Clinical and pathological characteristics of the population of patients with high-risk prostate cancer (N = 138)

Characteristics

Value

Mean age, years (range)

63.4 (39–80)

Follow-up, months

 Mean (SD)

57.3 ± 38.8

 Median (range)

53 (6–166)

PSA level, ng/ml

 Mean (SD)

15.5 ± 12

PSA group (%)

 <20 ng/ml

88 (64)

 ≥20 ng/ml

50 (36)

Clinical stage (%)

 T1c

17 (12)

 T2

29 (21)

 T3

90 (66)

Biopsy Gleason score (%)

 <8

119 (86)

 ≥8

19 (14)

Pathological stage (%)

 pT2

46 (33)

 pT3a

59 (43)

 pT3b

33 (24)

Specimen Gleason score (%)

 <8

99 (72)

 ≥8

39 (28)

Margins (%)

 M+

67 (49)

 M−

71 (51)

Lymph nodes (%)

 N+

21 (15)

 N−

117 (85)

For the pathological specimen, 67 patients (49%) had positive margins, and 21 patients (15%) had positive lymph nodes. The median number of nodes removed was 14. There was seminal vesicle invasion in 33 patients (24%). DRE was predictable for extracapsular extension on the specimen analysis (66% of T3 after DRE vs. 67% of pT3), but the biopsy Gleason score was underestimated, as 28% of the patients had a Gleason score ≥8 on the specimen versus 14% on the biopsy.

Survival

The median follow-up was 53 months (range 6–166). According to the Kaplan–Meier analysis, the overall, five-year, PSA recurrence-free survival rate was 40% (Fig. 1). Overall, 82 patients had a biochemical recurrence (59%) and 45 had been treated at the end of the current study. After recurrence, 31 patients (22.5%) were treated with salvage radiotherapy, 4 patients (3%) had hormonal therapy, and 10 (7.2%) patients had combination of both.
https://static-content.springer.com/image/art%3A10.1007%2Fs00345-010-0608-5/MediaObjects/345_2010_608_Fig1_HTML.gif
Fig. 1

Kaplan–Meier curves for PSA-free survival after RP: in all patients (n = 138) with high-risk prostate cancer (a); according to the clinical status (T1c vs. T2 vs. T3) (b); according to the biopsy Gleason score (<8 vs. ≥8) (c); according to the margin status (positive vs. negative) (d); according to the pathological stage (pT2 vs. pT3a vs. pT3b) (e); according to the lymph nodes status (positive vs. negative) (f)

In univariate analysis, two pre-operative indicators were associated with recurrence: a biopsy Gleason score ≥8 (P = 0.031) and a palpable tumour on DRE (P = 0.032). In contrast, extracapsular extension suspected by the DRE was not significant (P = 0.086). Concerning the post-operative data, a specimen Gleason score ≥8 (P = 0.009), seminal vesicle invasion (P < 0.001), positive margins (P < 0.001) and positive lymph nodes (P < 0.001) were associated with recurrence (Fig. 1). In multivariate analysis, only a biopsy Gleason score ≥8 (P = 0.014), positive margins (P = 0.013), lymph nodes (P < 0.001) and seminal vesicle invasion (P = 0.005) were independently associated with recurrence. During follow-up, 17 patients died of PCa and 5 died of other causes. Upon bootstrapping, our prediction model had a c-index (95% CI), which describes the predictive accuracy, of 0.74 (0.59–0.79). A graphical representation of the prediction model is presented in Fig. 2.
https://static-content.springer.com/image/art%3A10.1007%2Fs00345-010-0608-5/MediaObjects/345_2010_608_Fig2_HTML.gif
Fig. 2

Scatter plot (graphic presentation) of our institution’s model to predict PSA recurrence adjusted according to the main prognostic factors: margin status (a); pathological stage (b); clinical stage (c); biopsy Gleason score (d). NB: dashed lines represent confidence intervals

Discussion

The optimal treatment strategy for high-risk PCa remains particularly unclear and is still a matter of debate, but some recent studies advocate aggressive surgical resection [7, 9, 13, 14]. Historically, high-risk patients have not been considered good candidates for RP because of the high incidence of positive pelvic lymph nodes and poor long-term survival rates [15, 16]. In our study, we found only 15% of patients with positive lymph nodes. In our whole population, the oncological outcome was acceptable because the five-year PSA-free survival rate was estimated to be 40% with RP alone.

Early PSA failure is related to an increased risk of metastatic disease and PCa-specific mortality [17]. Therefore, the immediate and effective oncological control is essential. In regard to our results, RP must be considered a part of a multimodal treatment including hormone therapy if there is evidence of positive lymph nodes or even adjuvant radiotherapy in the case of positive margins [18, 19]. In the case of positive lymph nodes, there is a high level of evidence showing the necessity of an immediate adjuvant hormone treatment [20]. The exact and appropriate schedule for launching adjuvant treatment and the nature of this treatment itself remain a moot point in every other case [19, 21, 22]. In our study, immediate adjuvant radiotherapy was a predictive factor of recurrence only in univariate analysis. Objectively, it is clear that the patients who received adjuvant radiotherapy had a worse pathological prognosis than the patients who did not need immediate adjuvant treatment.

Surgery allows an acceptable local control and offers the ability to obtain specific anatomopathological information that is likely to accurately predict the outcome and thus pinpoint the patients who are likely to benefit from adjuvant therapy [14]. In the Capsure study, patients treated by radiotherapy instead of surgery for high-risk cancer had a 3.5 times greater chance of receiving adjuvant hormonal therapy [21]. It has been demonstrated that 14–91% of biopsies are still positive after radiotherapy, which is linked to a risk of a late wave of metastasis [8]. All these findings support the idea of better local control with RP than with radiotherapy alone.

In our study, all post-operative data collected were predictive factors of recurrence, as previously demonstrated [9]. We trust that the pathological findings are much more reliable than pre-operative factors in predicting the outcome, as demonstrated by Kattan et al. when they compared their pre- and post-operative predictive models [23, 24]. Several oncological tools have been previously reported, and nomograms currently appear to be the best predictive tools [25]. Thus, we decided to elaborate a prediction model from our database that we could use with the patients in our own institution (Fig. 2). We can use this scatter plot representation of the outcomes for our high-risk patients in our daily practice. Our model performed well (c-index >0.7) in predicting PCa recurrence, but the accuracy of the nomogram might depend on factors related to catchment areas. It is known that the nomogram’s reliability depends on the initial database, and the predictive accuracy of existing tools has been questioned as soon as they were used outside the original dataset of patients [25].

The risk of overestimating or even underestimating the clinical stage of the disease has been linked with the decision to choose the surgical option [26]. The accuracy of DRE has been questioned in assessing accurately the local extension of the PCa because small cT3s are good candidates for surgery, and almost 21% of patients can be overstaged by DRE [10]. In addition, the biopsy Gleason score is always re-interpreted from the definitive pathological specimen after surgery. For instance, in our study, the biopsy Gleason score was underestimated (19% of Gleason score equal to eight on the biopsy vs. 39% on the specimen). Thus, there is an urgent need for a more accurate pre-operative work-up to select good candidates for surgical removal of the prostate. MRI has produced promising results since we started to perform it in 2003, and some T3-MRIs have already been described in the literature [26, 27]. These findings emphasize the potential role of MRI in determining whether there is an extracapsular invasion. Thus, the relevance of considering pre-operative T3 only on DRE findings may be of limited accuracy and interest [26, 27]. When it concerns high-risk patients, surgery has to be standardized. All of our patients were treated with the same extensive surgery: systematic pelvic lymphadenectomy and non-conservative ablation of the prostate. Despite this surgical behaviour, it appears that the surgical morbidity is similar to low-risk patients [28]. Regarding urinary continence, functional outcomes are in line with those of the RP for localized PCa [29]. Some high-risk patients have organ-confined disease. The D’Amico pre-operative classification remains heterogeneous as it includes in the same category patients with T2 cancers and high PSA. However, the high PSA level may simply be due to voluminous benign prostate hyperplasia, extracapsular extension or seminal vesicle invasion [30]. For some of the high-risk patients, RP appears to be a good therapeutic option. Further studies are needed to define pre-operative criteria that might lead to systematic surgical options.

Conclusion

RP affords an acceptable oncological control and remains a possible therapeutic option in certain patients with a high-risk PCa. However, in certain cases, surgery alone is not sufficient and must be part of a multimodal treatment, including adjuvant radiotherapy or hormones. High-risk patients must be warned that surgery alone might not be sufficient to obtain disease control and that surgery may be a part of a multimodal treatment.

Conflict of interest

The authors declare that they have no conflict of interest.

Copyright information

© Springer-Verlag 2010