Inverse stage migration in patients undergoing radical prostatectomy: results of 8916 European patients treated within the last decade


Lars Budäus, Martiniclinic, Prostate Cancer Center University Hospital Hamburg-Eppendorf, Martinistrasse 52, 20246 Hamburg, Germany. e-mail:


Study Type – Therapy (case series)

Level of Evidence 4

What’s known on the subject? and What does the study add?

The widespread use of PSA testing resulted in a stage migration towards clinical organ-confined prostate cancers at diagnosis during the last decade. However, our study of a large cohort demonstrates an increasing proportion of patients with non-organ confined cancers after radical prostatectomy. These findings may be related to the introduction of new, non-established treatment options for low-risk prostate cancer patients during the last years and the growing adoption of RP in a multimodal treatment setting for locally advanced tumours.


• To investigate the stage migration patterns during the last decade in European men treated with radical prostatectomy (RP).


• Between 2000 and 2009, RP was performed in 8916 patients at a single European tertiary-care institution.

• Age at diagnosis, clinical and pathological data were prospectively collected, and trends and proportions of preoperative and pathological findings were analysed over time.


• The median (mean) age of patients increased from 62 (62) to 63 (65) years between 2000 and 2009 (P < 0.001).

• When patients were stratified based on their clinical findings according to the D’Amico risk groups for disease progression, the proportion of low-risk patients dropped from 66% in 2004 to 35% (P= 0.016) in the final year of the study period.

• Similarly, histopathological evaluation of RP specimens showed a decrease of favourable disease (organ confinement and Gleason 3 + 3 grade) from 53 to 17% (P= 0.008).

• This trend was accompanied by an increase in the number of patients with non-organ-confined prostate cancer (PCa) from 19% in 2003 to 33% in 2009 (P= 0.008).

• The restriction of the analyses in the present study to a single tertiary-care centre could limit the generalizeability of the results.


• During the last decade, we observed an inverse stage migration trend in those European patients with PCa who were treated with RP.

• The recorded increase in patients with non-organ-confined disease after RP could be related to changes in patient selection and the growing adoption of RP in multimodal treatment settings for locally advanced tumours as well as the availability of new treatment alternatives for low-risk disease.


high-frequency focused ultrasonography


lymph node invasion


prostate cancer


radical prostatectomy


seminal vesicle invasion


The incidence of prostate cancer (PCa) has increased substantially since the introduction of PSA screening in the late 1980s [1]. Moreover, a significant decrease in the mean age, PSA level and clinical stage at diagnosis has subsequently been reported in the United States [2]. According to this stage migration, the proportion of patients with organ-confined disease at final pathology after radical prostatectomy (RP) also increased during the past two decades in the USA [3].

Although PSA testing for early PCa detection is not routinely recommended in Europe, a stage migration towards organ-confined disease has also been described for European patients undergoing RP [4]. In a recent report, the stage and grade migration patterns were compared between European and American patients treated with RP from 1988 to 2005 [5]. Intriguingly, the proportion of American patients with organ-confined disease after RP increased to approximately 80% in 2001, followed by a subsequent decrease towards 60% organ confinement in 2005. Conversely, the stage migration pattern in Europe (represented by one Italian, one Austrian and one German tertiary referral centre) has been reported to follow a sustained increase, culminating in a 75% rate of organ-confined tumours in 2005 [5].

To date, no report has addressed the subsequent development of disease characteristics in European populations of patients who have undergone RP. Therefore, we examined the temporal trends of patient and tumour characteristics of 8916 men treated between 2000 and 2009. Moreover, we analysed the annual pathological stage distribution after RP and tested the hypothesis that, similar to trends recorded in the USA, PSA testing and new treatment regimens could also result in changes in pathological tumour characteristics in European patients with PCa.


Between 2000 and 2009, 9098 patients underwent RP for treatment of localized PCa at the Department of Urology, University Hospital Hamburg Eppendorf, and at the Martini-Clinic Prostate Cancer Centre Hamburg Eppendorf. The open intrafascial nerve-sparing RP technique was performed as previously described by 14 different surgeons [6,7].

The preoperative evaluation of patients included the assessment of clinical stage, PSA, percentage free PSA and rectal ultrasonography. The pretreatment PSA level (AxSYM; Abbott Diagnostics, Abbott Park, IL, USA) was measured before DRE and TRUS. Clinical stage was assigned by the attending urologist according to the 2002 TNM system. For patients diagnosed before 2002, the 1997 TNM system was converted to the 2002 TNM system. All clinical and pathological data were prospectively recorded in an institutional database.

Approximately 70% of the patients who underwent RP had a prostate biopsy at our institution (initial or repeat biopsy). Among the 30% who presented with an external biopsy result, two-thirds underwent a confirmatory in-house sextant prostate biopsy and/or their histology result was reviewed by our institutional uropathologist. This was done to reduce the risk of under-estimating grade and stage, which has implications for planning of the individual surgical approach (e.g. nerve-sparing, frozen section analyses, lymphadenectomy). The biopsy Gleason score for the remaining third of patients was assigned by members of a regional uropathology quality circle comprising 10 dedicated uropathologists, who were trained at the Department of Pathology, University Hospital Hamburg Eppendorf, to achieve consistent histology quality.

At prostate biopsy and RP specimen, the primary and secondary biopsy Gleason scores were assigned by a dedicated genitourinary-pathology service. RP specimens were surface-inked by the operating surgeon and processed by uropathologists using 3 mm serial step sections according to the Stanford protocol [8]. The 1992 TNM classification was converted to the 2002 TNM system for patients diagnosed and treated before 2002. The pathological stage was defined according to the four categories reported by Partin et al. [9].

For statistical analysis, clinical and pathological variables were stratified according to the year of surgery. Descriptive statistics consisted of tests of means, medians, proportions and trends that focused on all procedures performed during the study period. The significance of differences across categorical variables was assessed with the Mantel–Haenszel test. Temporal trends were assessed by comparing the first and second 5 years of the present study using the Mann–Whitney U-test. Risk of PCa was summarized using the D’Amico classification, which stratifies patients into low, intermediate and high risk based on the PSA level, biopsy Gleason score and clinical T stage [10].


Patients with unavailable postoperative data (n= 182), such as pathological Gleason score (six), clinical stage (40), PSA level (130) and pathology results (six), were excluded from the study. This resulted in 8916 evaluable patients undergoing RP at a single institution. The annual number of RPs increased 4.5-fold from 381 in 2000 to 1752 in 2009 (Fig. 1). The median (mean) age at presentation was 64.0 (62.8) years. Complete clinical and pathological data for the whole RP population are given in Table 1.

Figure 1.

Annual number of RPs performed.

Table 1.  Descriptive characteristics of 8916 patients who underwent RP between 2000 and 2009
 n (%) (unless noted otherwise)
  1. IAS, invasion adjacent structures (pT4). *Values missing to 100%: Gleason grading not applicable due to hormonal manipulation.

Age (mean, median range), years  62.8, 64.0, 36–79
Clinical stage (%) 
 T1c7128 (79.9)
 T21618 (18.2)
 T3 170 (1.9)
Mean (median) PSA level, ng/mL   8.50 (6.51)
Biopsy Gleason grade* 
 ≤3 + 34334 (48.6)
 3 + 42838 (32.0)
 4 + 31028 (11.6)
 ≥4 + 4 663 (7.5)
Pathological Gleason grade* 
 ≤3 + 33025 (33.9)
 3 + 44354 (48.8)
 4 + 3 1139 (12.8)
 ≥4 + 4 327 (3.7)
Organ-confined6297 (70.6)
Extracapsular extension2621 (29.4)
SVI  811 (9.1)
LNI 385 (4.3)
IAS  25 (0.3)

During the study period, median (mean) age at presentation increased from 62 (62) years in 2000 to 63 (65) years in 2009 (P < 0.001; Fig. 2). Conversely, median (mean) PSA levels at presentation decreased from 7.4 (10.2) ng/mL in 2000 to 6.6 (9.1) ng/mL in 2009 (P < 0.001; Fig. 3).

Figure 2.

Preoperative median (mean) age according to year of treatment (chi-squared trend, P < 0.001).

Figure 3.

Preoperative median (mean) PSA levels according to year of treatment (chi-squared trend for median PSA levels, P= 0.08; chi-squared trend for mean PSA levels, P < 0.001).

When patients were stratified according to the D’Amico risk groups for disease progression based on their clinical findings, there was an increase in the number of patients in the low-risk group from 46% in 2000 to 66% in 2004, and a subsequent decrease to 35% in 2009 (P= 0.016; Fig. 4). Likewise, the proportion of patients classified as intermediate and high risk decreased from 39 and 16%, respectively, in 2000 to 28 and 6% in 2004, followed by an opposite trend to 49 and 17% in 2009 (P= 0.008).

Figure 4.

Proportion of 8916 patients stratified according to D’Amico risk groups (low, intermediate and high) according to year of treatment (chi-squared trend, P= 0.008).

At surgical pathology, the proportion of non-organ-confined disease decreased from 29% in 2000 to 19% in 2003, and subsequently increased steadily to 33% in 2009 (P= 0.008; Fig. 5). During the same period, the proportion of patients with favourable disease (organ confinement and Gleason score 3 + 3) steadily decreased from 48% in 2000 to 17% in 2009 (P= 0.008).

Figure 5.

Proportion of organ-confined (OC) and non-organ-confined (non-OC) pathological stages in 8916 patients undergoing RP according to year of treatment (chi-squared trend, P= 0.008).

No statistically significant change was detected during the study period in the proportion of patients with seminal vesicle invasion (SVI) or pT4 tumour at final pathology (Fig. 6). The percentage of patients with invasion of the lymph node decreased from 5.5% in 2000 to 1.2% in 2002 and further migrated towards 5.6% in 2009 (Fig. 6).

Figure 6.

Rates of SVI, LNI and proportion of tumours invading adjacent structures (pT4) in 8916 patients who underwent RP between 2000 and 2009 according to year of treatment (chi-squared trend, P= 0.6).


A stage migration in patients with PCa towards younger age and earlier stage at diagnosis during the 1990s has been reported in several studies [3,11,12]. Most of these publications evaluated patients in North America where routine PSA testing is widespread used [10]. In patients undergoing RP for treatment of PCa, a recent report showed an inverse trend towards locally advanced tumours for patients from the USA during the last decade [5]. Accordingly, the rate of patients with organ-confined PCa after RP decreased from 80% in 2001 to 60% in 2005. Moreover, this trend was accompanied by an increase in patients with high-grade PCa (Gleason score ≥7) at clinical and surgical pathology [5].

Disease characteristics of European populations who have undergone RP have been suggested to differ from their North American counterparts [13,14]. The onset and distribution of PSA testing have, to a large extent, explained these observed differences [13]. Nonetheless, a significant trend towards organ-confined tumours after RP has also been seen in European patients, with 75% having organ-confined disease after RP in 2005 [5]. However, neither the subsequent trend of pathological stages nor detailed clinical variables have ever been reported during the last decade [5].

Owing to the lack of more recent data addressing this topic specifically in European patients, in the present study we analysed the temporal stage migration pattern in RP patients from a single German tertiary-care centre. We hypothesized that PSA testing and new treatment regimens could result in changes in pathological tumour characteristics in patients with PCa in Europe, similarly to those trends reported in patients from the USA.

The analysis of these data produced several important findings. First, when patients were stratified according to the D’Amico criteria before surgery, there was a decline in the proportion of low-risk patients, from 66% in 2004 to 35% in 2009. Conversely, the proportion of patients in the intermediate- and high-risk groups increased between 2004 and 2009, from 28 to 49% and 6 to 17%, respectively. Second, there was an increase in the proportion of patients with non-organ-confined disease at surgical pathology, from 29 to 33%. Furthermore, the proportion of patients with a better prognosis (organ-confined disease and Gleason score 3 + 3) decreased from 53% in 2004 to 17% in 2009. Finally, the proportion of patients with SVI and pT4 PCa remained virtually stable during the study period, while the presence of lymph node invasion (LNI) increased from 1.2% in 2002 to 5.6% in 2009.

The recorded trend towards higher age and more aggressive tumour characteristics in European patients undergoing RP in the present study closely follows the recently reported trends in patients undergoing RP in the USA, with a delay of 5–8 years.

These findings might be explained by several factors. First, a paradigm shift in treating high-risk PCa has been observed within the urological community. Historically, surgeons were reluctant to perform RP in locally advanced PCa such as clinical T3 or lymph node-positive disease [15]. With evolving surgical technique, open and robotically assisted RP can be performed safely even in locally advanced disease [16,17]. Second, several retrospective studies have reported that RP is also a valid option for patients with locally advanced disease [18–20]. For example, since patients with clinical T3 PCa are found to be often over-staged, RP as part of a multimodal treatment strategy can offer cancer control and survival rates, approaching those for clinical T2 disease [21]. Similarly, Freedland et al. [18] reported that RP alone for T3 disease provides a 15-year PCa-specific survival rate of 84% and long-term cancer control is achieved in about half of these patients.

Moreover, in the case of biochemical recurrence, salvage therapy could provide cure for a substantial proportion of patients. Van Der Poel et al. [22] recently reported the effectiveness and morbidity of sequential treatment of patients with recurrent localized PCa. It was shown that 10-year cancer control rates did not differ significantly between patients who underwent radiation therapy followed by salvage RP and those who underwent RP followed by salvage radiation therapy. However, urinary continence and erectile function were better preserved in patients undergoing salvage radiotherapy after surgery than in patients undergoing salvage prostatectomy after radiation [22]. This finding of a better quality of life for those patients treated with surgery followed by adjunctive radiotherapy could also have contributed to a higher acceptance of RP for treating high-risk patients’ tumours.

The recorded increase of the patients’ median age during the study period might be explained by the implementation of prediction tools and nomograms in daily clinical practice instead of a strict age limit. These tools allow better identification of those patients suitable for RP with a life expectancy greater than 10 years [23,24].

Additionally, an array of focal and non-surgical treatment alternatives for low- and intermediate-risk PCa evolved during the last decade. These techniques could also contribute to the observed inverse trend of organ-confined tumours in the RP population in the present study [25]. For example, an increasing proportion of patients are treated with focal treatment regimens such as high-frequency focused ultrasonography (HIFU) therapy, cryo-ablation therapy or brachytherapy [25]. However, since sufficient long-term oncological results have only been reported for patients treated with brachytherapy, other treatments such as HIFU and cryo-ablation therapy should be considered as experimental according to current treatment guidelines [26]. Moreover, an increasing proportion of patients with low-risk disease selected deferred treatment according to active surveillance protocols, with the possibility to offer a curative treatment once the PCa seemed to progress [27].

In addition to these changes in treatment selection and patients’ characteristics, the 2005 modification of the Gleason classification of prostate biopsies might also explain some of the findings of the present study. For example, since 2005, more aggressive tumour biopsy fractions are reported even if they were less than 5% of the biopsy core and should replace a less aggressive secondary Gleason score, because of the assumption that this would reflect the aggressiveness of the whole tumour more accurately [28,29]. Albertsen et al. [30] recently examined the potential impact of this modification of the Gleason classification on the shift towards higher Gleason scores that has appeared during recent years. The authors showed that changes in the interpretation of prostate biopsy and specimen rather than selective identification of more aggressive tumours by PSA testing have been responsible for the shift in Gleason score observed [30]. However, these changes in clinical practice do not occur instantly. Moreover, the decrease in the number of patients with Gleason score 3 + 3 at final pathology or biopsy pattern is closely paralleled by an increase in the number of patients with non-organ-confined tumours, which might weaken the effect attributable to changes in the Gleason score classification.

The present study has several limitations. First, the analyses were restricted to a single tertiary-care centre in Germany and might not reflect the changes observed in rural areas or other geographic European regions. Second, the database only allows the analysis of the temporal trends in those patients who underwent RP. Therefore the reported inverse stage migration reflects changes in treatment selection and treatment decision and not changes in patients diagnosed with PCa. Moreover, the decision regarding which established treatment for PCa to adopt, such as active surveillance protocols, external beam radiation, low-dose and high-dose brachytherapy and RP, is taken by an interdisciplinary tumour board, incorporating radio-oncologists and medical oncologists. Hence, the observed trends might be different in other centres, especially if not all treatment regimens for PCa are offered or if the counselling physician prefers a specific treatment regimen.

Unfortunately no peer-reviewed population-based data are available that address the temporal trends of the proportion of patients undergoing different treatment regimens, such as focal or non-surgical treatment alternatives. Currently, the best data on this subject are reported by the Federal state of Brandenburg tumour registry. Between 2003 and 2005, 70% of patients <70 years old underwent RP, whereas 15% of patients selected radiotherapy alone [31,32].

Similarly, no data are available addressing the population-based use of HIFU as an example of focal therapy in Germany. Since full public reimbursement by public health care payers for HIFU is only available in Germany and Italy, adopting data from other countries could be non-representative [33].

Finally, no data are available on the distribution of active surveillance for selected patients in Germany. Only one German health service research study, begun in 2008, addresses the quality of life and oncological outcome in patients with PCa who underwent different treatment options, such as radiotherapy, focal therapy or active surveillance. Currently, 854 patients between 18 and 99 years old are recruited. The first results of this study are expected in 2012 [34].

Based on a growing body of published evidence, active surveillance has been introduced as a treatment option for selected patients in European and American treatment guidelines. Given the increasing number of patients following active surveillance within research protocols [e.g. the Prostate Cancer Research International Active Surveillance (PRIAS) study and the German HAROW study], we assume that there is also wide acceptance of active surveillance as a guideline compliant strategy among German patients with Pca.

Several studies have reported a stage migration towards more favourable PCa characteristics at diagnosis. The analyses in the present study of European patients who have undergone RP showed an inverse stage migration and an increasing proportion of patients with non-organ-confined disease after RP. This trend in European patients who have undergone RP closely follows the trend recently reported in patients undergoing RP in the USA, with a delay of 5–8 years. This trend could be indicative of the broader use of alternative treatment regimens to RP in low-risk patients with PCa, such as brachytherapy, focal therapy and active surveillance protocols. Moreover, the recorded trends could also reflect the increasing importance of RP in treating high-risk patients with PCa.


None declared. Source of Funding: Lars Budäus is supported by the German Federal Ministry of Education and Science in the framework of the program for medical genome research (FKZ:01GS08189).