Study Type – Therapy (prospective cohort)
Level of Evidence 2b
Study Type – Therapy (prospective cohort)
To evaluate the short-term outcomes of the prospective international Prostate Cancer Research International: Active Surveillance (‘PRIAS’) study (Dutch Trial Register NTR1718), as active surveillance (AS) for early prostate cancer might provide a partial solution to the current overtreatment dilemma in this disease.
PATIENTS AND METHODS
The first 500 (of >950) participants with asymptomatic T1c/T2 prostate cancer, with a prostate-specific antigen (PSA) level of ≤10.0 ng/mL, a PSA density of <0.2 ng/mL/mL, a Gleason score of ≤3 + 3 = 6, and one or two positive biopsy cores, were analysed. The follow-up protocol consisted of frequent PSA measurements, digital rectal examinations, and standard repeat biopsies (the first after 1 year). The primary outcome is survival free of active therapy; the secondary endpoints are reasons for stopping AS, findings in 1-year repeat biopsies, and outcomes after radical prostatectomy (RP).
Patients were included between December 2006 and July 2008. The median (25–75th percentile) follow-up after diagnosis was 1.02 (0.6–1.5) years. The 2-year survival rate free from active therapy was 73%. Of the 82 men who changed to active therapy during the follow-up, 68 (83%) did so based on the protocol. Of the 261 repeat biopsies available for analysis, 90 (34%) showed no cancer, while 57 (22%) showed a Gleason score of >6 or more than two positive biopsy cores. There was a relatively unfavourable PSA doubling time of 0–10 years in 53% (102/194) and 62% (33/53) of men with favourable and unfavourable re-biopsy results, respectively. After RP, four of 24 (17%) men had T3 disease and 12 (50%) had a Gleason score of >6.
AS seems feasible, but mortality outcomes are unknown. A strict follow-up protocol including standard 1-year repeat biopsies resulted in a quarter of men stopping AS after 2 years.
active therapy (-free survival)
- PRIAS, Prostate Cancer Research International
European Randomized Study of Screening for Prostate Cancer
Screening for prostate cancer has the potential to decrease disease-specific mortality, but also results in over-diagnosis . Many men who are currently being diagnosed with prostate cancer will never develop symptoms during their lifetime, even if left untreated [2,3]. Radical treatment, which brings an immediate risk of side-effects, is thus not always indicated .
Active surveillance (AS) has emerged as an alternative treatment option . AS aims to avoid overtreatment in men with small, localized, well-differentiated prostate cancer, by initially withholding radical treatment. Instead, the tumour is closely monitored with the purpose of switching to active local therapy with curative intent if there is progression.
AS programmes have been initiated to acquire evidence for the currently used selection and surveillance criteria, which are based on retrospective data [6–11]. The ‘Prostate Cancer Research International: Active Surveillance (PRIAS) study  is an international prospective observational study originating from the European Randomized Study of Screening for Prostate Cancer (ERSPC) . PRIAS offers a protocol for the inclusion and follow-up of men, which is applied by a web-based instrument .
The objective of the present report was to assess the effect on survival free of active therapy (AT-free survival, ATFS) of applying the PRIAS protocol for AS on 500 men diagnosed with early prostate cancer.
PATIENTS AND METHODS
The PRIAS study centre accrual and patient inclusion started in December 2006 and is ongoing. Currently (September 2009), >950 patients had been included. This report presents the first interim analysis, based on the first 500 study inclusions. The main outcome variable was ATFS; secondary endpoints included reasons for stopping AS, findings in the standard 1-year repeat biopsies, and outcomes after radical prostatectomy (RP). Furthermore, the PSA doubling time (DT) distribution, findings at a DRE, and preliminary survival details are discussed.
The criteria for eligibility for PRIAS aim to select asymptomatic, small, localized, well-differentiated prostate cancer and consist of: histologically confirmed adenocarcinoma of the prostate, no previous treatment, fit for curative treatment, clinical stage T1C or T2, a PSA level of ≤10.0 ng/mL, a PSA density of <0.2 ng/mL/mL, a Gleason score of 3 + 3 = 6, or more favourable, and one or two biopsy cores invaded with prostate cancer .
The PRIAS follow-up scheme consists of PSA measurements every 3 months and a DRE every 6 months during the first 2 years after diagnosis; thereafter PSA measurements every 6 months and a yearly DRE. Repeat biopsies are standard and are taken after 1, 4 and 7 years. Whenever during the follow-up the PSA-DT is 0–3 years, the clinical stage > T2, or the re-biopsies show more than two positive cores or a Gleason score of >3 + 3 = 6, the protocol advises a change to AT. When the PSA-DT is in the intermediate range of 3–10 years, yearly repeat biopsies are advised, instead of the standard schedule. The PSA-DT is advised to be used only after 1 year of follow-up, with five separate PSA measurements available. Whenever the PSA level is >20 ng/mL, a bone scan is advised.
A prostate volume-dependent number of random biopsy cores is advised, but is not obligatory, i.e. if <40 mL then eight; 40–60 mL then 10, and >60 mL then 12 biopsies . Biopsy cores containing high-grade prostatic intraepithelial neoplasia or atypical lesions are considered to be negative.
The medical ethical committee of the Erasmus University Medical Centre and, dependent of local regulations, local committees, approved the PRIAS study (MEC number 2004–339). All participants provided written informed consent.
The PRIAS study website is at http://www.prias-project.org. Study documents, screen-shots and general information are open for the public. Physicians log in using a personal account and use the website to include and follow-up their patients on AS. The PSA-DT is calculated automatically and presented in graphs. Furthermore, automatic individualized recommendations based on the protocol are provided (Fig. 1). The coordinating study centre (Erasmus MC, Rotterdam, the Netherlands) monitors and checks the website database every 3 months.
ATFS was assessed using Kaplan-Meier analysis. Reasons for stopping AS and treatments chosen are entered by the treating physicians. Potential predictors of adverse findings in repeat biopsies were explored using multivariate logistic regression analysis. The PSA-DT was calculated by plotting the base-2 logarithm of the PSA value against time since diagnosis. The DT can be calculated as the reciprocal value of the slope of the regression line through these points. PSA values that were measured during infection or shortly after biopsy (as mentioned by the treating physician) were excluded from calculation. Local pathologists reviewed RP specimens.
Men who were already under surveillance, or had already stopped surveillance, at the moment of inclusion were only included if the diagnosis was ≤2 years earlier and if the PRIAS follow-up protocol was applied in this period. Patients using 5α-reductase inhibitors (5-ARIs) were excluded from the current analysis.
Table 1 lists those centres that included ≥10 of the first 500 patients in PRIAS; Table 2 lists the baseline and follow-up characteristics of the study cohort. Patients were included between December 2006 and July 2008. The median (25th−75th percentile) follow-up was 1.02 (0.6–1.5) years. Figure 2 shows the total ATFS and those stratified for reasons for stopping AS (PSA-DT-related, biopsy-related, or other). There was a prominent decrease in the biopsy-related, PSA-DT related, and total curve at ≈1 year of follow-up, which corresponds with the standard 1-year re-biopsy and the use of the PSA-DT as an exclusion factor when five measurements are available after 1 year, as indicated by the protocol.
|Rotterdam||Erasmus University Medical Centre||63|
|Amsterdam||Netherlands Cancer Institute, Antoni van Leeuwenhoek Hospital||55|
|Dordrecht||Albert Schweitzer Hospital||44|
|Rotterdam||St. Franciscus Hospital||41|
|Hengelo||Hospital Group Twente||33|
|Delft||Reinier de Graaf Hospital||11|
|Vancouver||British Columbia Cancer Agency||37|
|Milan||Fondazione IRCCS, Istituto Nazionale dei Tumori||31|
|Reims||Centre Hospitalier Universitaire Robert Debré||10|
|Characteristic||Median (25–75th percentile)|
|Age, years||66.0 (60.7–70.4)|
|PSA, ng/mL||5.3 (3.9–6.7)|
|Prostate volume. mL||42.6 (35.0–56.0)|
|PSA density, ng/mL/mL||0.12 (0.09–0.16)|
|Total biopsy cores, n||8.0 (6.0–11.0)|
|Positive biopsy cores, %|
|Gleason score, %|
|3 + 3 = 6||95.0|
|Follow-up (year)||1.02 (0.6–1.5)|
|Number of follow-up visits||4 (2–6)|
Table 3 lists the specific reasons for 82 men stopping AS, as well as the ATs they selected. Men stopped AS based on medical variables according to protocol in 83% (68) and based on patient anxiety and/or request in 17% (14). There was a biopsy-related reason for stopping AS in 56% (38/68) of the protocol-based decisions, a PSA-DT-related reason in 62% (42), and in 13 of these men both factors were simultaneous reasons for stopping. All patients received radical treatments; palliative hormonal therapy-only was not given.
|Reasons for switching to AT||Deferred ATs|
|PSA-DT, re-biopsy n positive cores and re-biopsy Gleason score||1||–||–||–||1|
|PSA-DT and re-biopsy n positive cores||4||2||–||1||7|
|PSA-DT and re-biopsy Gleason score||1||2||1||1||5|
|Re-biopsy n positive cores and Gleason score||5||5||1||1||12|
|Re-biopsy n positive cores||6||1||5||–||12|
|Re-biopsy Gleason score||2||–||3||–||5|
|T stage only||–||–||–||1||1*|
|Did not want follow-up anymore||–||2||1||1||4|
Of the 170 men with >1.25 years of follow-up (taking into account repeat biopsy being ‘late’ in one follow-up visit), 14% (24) did not comply with having a repeat biopsy. In all, 261 men had standard repeat biopsies, shown in Table 4 by the PSA-DT calculated using the available PSA values at that time. Biopsies were taken at a median (25–75 percentile) of 1.02 (1.0–1.1) years after diagnosis. The median number of cores in the re-biopsy was 10 (8–12). No histological prostate cancer, ‘PRIAS-suitable’ prostate cancer (one or two positive biopsies, Gleason score <3 + 3 = 6), and more unfavourable prostate cancer (more than two positive biopsies and/or Gleason score >3 + 3 = 6) were found in 34% (90/261), 44% (114), and 22% (57), respectively. The PSA-DT was unknown in 5% (14) of these patients. In 194 men with favourable re-biopsy findings the PSA-DT was 0–3, 3–10, >10 years, and negative in 28% (54), 25% (48), 8% (15) and 40% (77), respectively. In 53 men with unfavourable re-biopsy findings the PSA-DT was 0–3, 3–10, >10 years, and negative in 28% (15), 34% (18), 6% (three), and 33% (17), respectively. The PSA-DT was not significantly different between men who showed upgrading in the biopsy characteristics or not (Mann–Whitney P = 0.411).
|Favourable findings, 194 men|
|A, no prostate cancer||24 (28)||18 (21)||7 (8)||37 (43)||4||90 (34)|
|B, ‘PRIAS suitable’ (1–2 biopsy cores, Gleason ≤3 + 3 = 6)||30 (28)||30 (28)||8 (7)||40 (37)||6||114 (44)|
|A + B||54 (28)||48 (25)||15 (8)||77 (40)||10||204 (78)|
|Unfavourable findings, 53 men|
|C, >2 biopsy cores, Gleason ≤3 + 3 = 6||7 (24)||12 (41)||1 (3)||9 (31)||3||32 (12)|
|D, 1–2 biopsy cores, Gleason >3 + 3 = 6||5||1||2||2||0||10 (4)|
|E, >2 biopsy cores and Gleason >3 + 3 = 6||3||5||0||6||1||15 (6)|
|C + D + E||15 (28)||18 (34)||3 (6)||17 (33)||4||57 (22)|
|Total||69 (28)||66 (27)||18 (7)||94 (39)||14||261 (100)|
In univariate analysis (data not shown) the number of positive biopsy cores at the initial diagnostic biopsy (two vs one) was the only variable showing a positive relation to unfavourable repeat biopsy findings. Re-biopsy findings were unfavourable in 17% (32/183) of men with one positive biopsy at diagnosis and in 32% (25/78) of men with two positive biopsies (chi-square P = 0.014). Age, clinical stage, PSA, PSA-DT (neither slope of the regression line, PSA-DT divided in groups, nor PSA-DT with negative values set at 50 years), prostate volume, time to re-biopsy, number of biopsy cores taken at diagnosis, at re-biopsy, difference in number of biopsy cores, or the ratio between the initial number of biopsy cores and the repeat number of cores (even though a median number of two more cores was taken at re-biopsy) showed no significant association.
Of the 27 men who had RP after a median of 1.0 (0.5–1.1) years after starting on AS within PRIAS, the histopathological outcomes were available in 24 (89%), as shown in Table 5. These patients chose RP based on: PSA-DT only (six), re-biopsy findings only (10), a combination (six), or other reasons (five). In the four T3 tumours the re-biopsy already showed both more than two positive biopsies and a Gleason score of >6 in three cases, and more than two positive biopsies in one case. Of the 12 cases showing Gleason upgrading in the RP specimen, re-biopsies were taken previously in 11, and in all these unfavourable characteristics were found (Gleason >6 in seven, more than two positive biopsies in eight, and a combination in four). There was no significant association for any inclusion variables or PSA-DT with adverse findings (T3 or Gleason score >6) on RP.
|pT-stage (TNM 2002)|
|3 + 3 = 6||12 (50)|
|3 + 4 = 7||10 (42)|
|4 + 3 = 7||1 (4)|
|3 + 5 = 8||1 (4)|
|Lymph node stage|
|First PSA level after RP, ng/mL|
In all, 2080 PSA measurements were made during the follow-up (mean 4.2 per patient). Considering the cumulative follow-up of the study cohort, accounting for the difference in frequency of PSA measurements between 0 and 2 years after diagnosis and in the period thereafter, 2026 PSA measurements could be expected (mean 4.1 per patient). With more PSA estimates available for PSA-DT calculation per patient, there were decreasing proportions of very unfavourable (0–3 year) and very favourable (negative) PSA-DTs, due to a regression-to-the-mean effect and due to the protocol advice of switching to AT if there was a rapid increase in PSA level (data not shown).
In all, 23 men showed upstaging in clinical stage on the DRE, of cT1c to cT2, during the follow-up, with none to cT3. Men with DRE upstaging had a longer follow-up after diagnosis (t-test P = 0.009), but were not significantly different from other men in baseline characteristics (all P > 0.05). Re-biopsies (in 19 of 23 men) showed no prostate cancer, PRIAS-suitable prostate cancer, or upgrading/upstaging in 26% (five), 47% (nine) and 26% (five), respectively. PSA-DTs in this group were 0–3, 3–10, >10 years, or negative in 43% (10), 22% (five), 22% (five), and 13% (three), respectively.
The present data do not allow for a mortality analysis. During the available follow-up, no man died from prostate cancer; two patients died from other causes (abdominal aneurysm 1.4 years after diagnosis and myocardial infarction 2.2 years after diagnosis); lymph node metastases were detected in one patient at 1 year after diagnosis.
This study presents the largest prospectively analysed cohort of patients in an AS programme, and with good protocol compliance. We found that applying a strict AS follow-up protocol in selected patients with early prostate cancer resulted in a quarter of men stopping AS after 2 years of follow-up. An important reason for stopping AS was the finding of adverse characteristics in standard repeat biopsies in 20% of patients. There was no difference between PSA-DTs at the moment of re-biopsies that had favourable and unfavourable outcomes. The men with unfavourable findings at RP already had an unfavourable re-biopsy result.
A solution to the overtreatment dilemma of over-diagnosed prostate cancer, as is aimed by AS, has gained further importance since the ERSPC recently reported that screening has the potential to decrease prostate cancer mortality rates . The short-term observations presented in the present study are important to monitor the effects and safety of applying, in men with early prostate cancer, a selection and follow-up AS protocol based on retrospective studies [3,17]. However, a longer follow-up is needed to assess mortality outcomes.
A considerable number of men stopped AS during the follow-up (ATFS 73% after 2 years). The rate of men stopping AS appears to decline after an initial period of selection between tumours with a favourable and tumours with an unfavourable clinical behaviour, but a longer follow-up is needed. The 2-year ATFS rates as reported by other AS study groups vary from 67% to as high as 95%, depending on patient selection and follow-up criteria, and the prospective or more retrospective nature of these studies [6,11]. Although the aim of AS is to eventually avoid AT, switching to radical treatment during the follow-up when progression occurs is an essential part of this strategy.
Psychological factors might influence decisions on receiving treatment in men with expectantly managed prostate cancer . The main reason for stopping AS in our cohort was protocol-based and not due to psychological factors. Other studies have reported a larger role for non-medical reasons to stop AS . Anxiety and distress levels in (Dutch) men participating in the PRIAS study have been previously reported to be mostly favourable . Anxiety of the physician was not recorded as a reason for stopping AS in our cohort .
Biological progression might be a reason for standard repeat biopsy upgrading, but is mainly caused by initial under-sampling. Choo et al. studied the upgrading in Gleason scores and number of positive biopsies in standard repeat biopsies during AS. The median number of prostate biopsy cores was six on both initial and repeat biopsy, compared to eight and 10 in PRIAS. Compliance with the repeat biopsy was also lower at 64%, vs 86% in PRIAS. There was upgrading in Gleason scores to >3 + 3 = 6 in men with an initial Gleason score of ≤3 + 3 = 6, such as included in the PRIAS study, in 38% (32/84), and an increase in the number of positive biopsies to more than two in 37% (39/105). These proportions were only 10% (25/261) and 18% (47/261) in PRIAS. The difference might be explained by the fewer biopsy cores and by the lower re-biopsy compliance rate. Probably, mostly men with unfavourable characteristics were re-biopsied, and men with a favourable disease status were not. By contrast with our findings, another study found a relationship between PSA kinetics and adverse findings in repeat biopsies . However, the repeat biopsies in that study were not standard, but depended on clinical changes. We found no association between the number of biopsy cores obtained at initial and repeat biopsy and the numbers of biopsy upgrading during the follow-up .
Although there were too few RP for definitive analysis, histopathological findings after deferred RP are in line with the outcomes reported after immediate RP in similar selections of patients. Suardi et al. studied the pathological features of RP specimens between 1992 and 2007 in 2345 men with prostate cancer at diagnosis graded T1c/T2a, a Gleason score of ≤6, and a PSA level of ≤10.0 ng/mL. Although the patient characteristics differ from the PRIAS criteria (PSA density and number of biopsy cores was not assessed, and T2b/T2c disease was not included), that study gives an indication of the true nature of prostate cancers included in PRIAS. The Gleason score was >6 in 35.3%; there was extracapsular extension, seminal vesicle invasion, and lymph node involvement in, respectively, 13.5%, 2.9% and 0.9%. In a similar study design, Conti et al. found rates of Gleason upgrading of 23–35% and extracapsular extension of 7–11% when applying AS criteria mainly corresponding to PRIAS. AS might result in a selection process in which mainly men with unfavourable follow-up characteristics receive surgery and therefore show more adverse outcomes after RP, while the patients with a favourable follow-up remain untreated .
The present study has some weaknesses. First, a randomized control group receiving initial radical treatment is not included. Recently, the START trial has been initiated in North America, which aims to study differences between expectant management and radical treatment in men with early prostate cancer, in a randomized fashion . Difficulties in setting up a randomized trial include: (i) the expected survival benefit of RP and radiotherapy is probably small ; (ii) the number of patients needed is very large; and (iii) the follow-up needed is very long. Mortality outcomes could not be analysed in the present study as the follow-up was too short. Further, there might be a selection bias in some of our patients, because physicians might tend to delay inclusion of a patient into PRIAS until, e.g. a second favourable PSA measurement is available. Finally, the percentage of tumour involvement in the biopsy cores was not included in the inclusion and follow-up criteria, to widen the applicability of our protocol to peripheral hospitals where this is not a standard measurement, although at the cost of the predictive value for adverse findings after RP . For this reason, nomogram scores for harbouring indolent prostate cancer  cannot be applied in all patients. However, a future subgroup analysis is possible.
A strength of the present study is that there were many patients; PRIAS has included more patients than any other prospective AS study reported. Furthermore, the protocol adherence is strong, possibly as a result of web-based inclusion and follow-up.
Based on the current analysis, there are no direct indications for patient safety, to tighten the criteria for AS as currently used in PRIAS. If a higher ATFS is desirable, possible adaptations to the protocol might be an option. Patients with only one positive biopsy core have a lower chance of adverse findings at repeat biopsy and might therefore remain on AS for longer. Eggener et al. have also reported that this variable is associated with ATFS. Another possibility is to include a standard repeat biopsy in the inclusion criteria for AS. The importance of repeat biopsies during AS has been emphasized by other study groups [22,31].
Definitive outcomes of AS will not only be defined by tumour, but also by patient characteristics, in relation to life-expectancy and comorbidity scores. Future research, which can partly be done within the PRIAS study, should aim to study long-term mortality outcomes of AS when compared to other treatment options, optimize AS inclusion and follow-up criteria, clarify the role of PSA kinetics during AS, assess the effect of 5ARIs during AS, incorporate new biomarkers and imaging techniques, and study the effect on quality of life during AS when compared to other treatments.
In conclusion, AS for early prostate cancer using a web-based tool is a feasible strategy to avoid overtreatment in the short term after diagnosis. Applying the strict PRIAS inclusion and follow-up protocol results in a quarter of men who start on AS switching to AT within 2 years after diagnosis, mainly because 20% of standard repeat biopsies at 1 year after diagnosis show adverse findings, which is independent of the PSA-DT. Based on the current short-term analysis, the protocol seems not to comprise curability. In the future, PRIAS could provide data to optimize AS protocols, as well as long-term mortality and quality of life outcomes of AS.
The authors thank all patients participating in PRIAS, as well as all physicians and data managers who contributed to the PRIAS study, especially Alex Briede (Vancouver, Canada), Tiziana Magnani (Milan, Italy), Erik van Muilekom (Amsterdam, the Netherlands), and Heleen Versteeg-Leenheer (Rotterdam, the Netherlands).
CONFLICT OF INTEREST
None declared. Source of Funding: Prostate Cancer Research Foundation (SWOP) Rotterdam, the Netherlands.