Pathological and biochemical outcomes after radical prostatectomy in men with low-risk prostate cancer meeting the Prostate Cancer International: Active Surveillance criteria

Authors


Abstract

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

  • Active surveillance has been widely accepted as a treatment tool for low-risk prostate cancer, and use of the Prostate Cancer Research International: Active Surveillance (PRIAS) criteria can select smaller and less aggressive tumours in low-risk disease.
  • The study shows the pathological outcomes of radical prostatectomy for patients with low-risk disease who met the PRIAS criteria. It found that ∼20% had unfavourable pathological features and only 30% satisfied insignificant cancer criteria with pT2 stage, a Gleason score ≤6 and tumour volume <2.5 mL. It concludes that close follow-up including repeat biopsy or MRI is necessary to minimize unexpected progression of disease.

Objective

  • To assess the effectiveness of the Prostate Cancer Research International Active Surveillance (PRIAS) criteria in identifying indolent cancer.

Patients and Methods

  • Data from 1268 patients undergoing radical prostatectomy without neoadjuvant therapy were retrospectively reviewed.
  • Within this cohort, patients with low-risk disease (n = 211) were classified according to whether they met (Group A, n = 87) or did not meet (Group B, n = 124) the PRIAS criteria.
  • Pathological upstaging, upgrading, tumour volume and 5-year prostate-specific antigen (PSA) recurrence-free survival were compared between the two groups, and factors that predicted upstaging, upgrading and PSA recurrence were analysed by univariate and multivariate methods.

Results

  • Pathological T3 stage was present in 10.3% of patients in Group A and in 18.5% of patients in Group B (P = 0.08). Gleason score upgrading to 4+3 or greater was seen in 19.5% of Group A and in 29.9% of Group B (P = 0.01).
  • The mean (range) tumour volume was 0.81 (0.03–5.09) mL in Group A and 1.40 (0.04–8.21) mL in Group B (P < 0.01). The rates of insignificant cancer with total tumour volume <2.5 mL, Gleason score ≤6 and stage pT2 were 30.6% in Group A and 15.4% in Group B (P = 0.07).
  • With a median follow-up of 44 months, the 5-year PSA recurrence-free survival rates were 91.2% in Group A and 86.4% in Group B (P = 0.47).
  • In multivariate analysis, PSA density and the PRIAS criteria were independent factors that predicted upstaging.

Conclusions

  • Although use of the PRIAS criteria could select more favourable tumours even in low-risk prostate cancer, about one in five men had unfavourable pathological outcomes and only three in ten had insignificant cancer.
  • Close and careful follow-up is necessary to avoid misclassification or progression of disease, especially during the first few years of active surveillance.
Abbreviations
PRIAS

Prostate Cancer International: Active Surveillance

RP

radical prostatectomy

AS

active surveillance

GS

Gleason score

ERSPC

European Randomized Study of Screening for Prostate Cancer

OR

odds ratio

Introduction

The 5-year cancer-specific survival for men with low-risk prostate cancer is nearly 100%, regardless of the treatment they receive (radical prostatectomy [RP], external beam radiation, brachytherapy, androgen deprivation therapy or cryotherapy). The use of PSA screening has been helpful for some men with prostate cancer, but may result in overtreatment of patients with low-risk disease. Indeed, Albertsen et al. [1] reported that men with low grade prostate cancer have a minimal risk of dying from prostate cancer during 20 years of follow-up. Furthermore, Welch and Albertsen [2] reported that over a million men were definitively treated for prostate cancer and that 56 000 avoided cancer death; however, >20 patients with prostate cancer were treated to prevent one prostate cancer death. This is a significant finding, in that overtreatment for low-risk prostate cancer can result in considerable morbidity, including incontinence, erectile dysfunction or depression.

Determination of whether a patient should be treated for low-risk prostate cancer remains controversial. Several studies, including the Prostate Cancer International: Active Surveillance (PRIAS) trial, which was conducted in Europe and Japan, have attempted to assess the value of active surveillance (AS), but there are still unresolved problems with regard to selecting patients or the appropriate follow-up protocol. Thaxton et al. [3] reported that of >3000 men with prostate cancer undergoing RP who met the AS criteria, 3–4% had a Gleason score (GS) of 8–10, 16–19% had positive surgical margins, 15–18% had extracapsular tumour extension, 3–5% had seminal vesicle invasion and 0.4–1% had lymph node metastasis. They concluded that a substantial proportion of men who might have been considered potential candidates for AS had aggressive tumour features at RP and that the accurate identification of patients with truly indolent cancer at the time of diagnosis remains challenging. The aim of the present study, therefore, was to assess the validity of the PRIAS selecting criteria by comparing pathological and biochemical outcomes after RP between men who retrospectively were or were not candidates for the PRIAS strategy.

Patients and Methods

The study was approved by our institutional review board, with all participating sites providing the necessary institutional data-sharing agreements before initiation of the study. The database comprised 1286 consecutive patients with clinically localized prostate cancer, who underwent RP at Tohoku University, Akita University, Hirosaki University or the Miyagi Cancer Centre between January 2000 and December 2009. None of the patients had a history of prostate surgery, nor did they undergo neoadjuvant hormonal or radiation therapy. Of these patients, the low-risk group (defined by D'Amico criteria as PSA <10 ng/mL, GS ≤6 and clinical stage T1c or T2a) underwent further analysis.

Patients were categorized as either meeting (Group A) or not meeting (Group B) the PRIAS criteria (Fig. 1). The PRIAS criteria consisted of PSA level ≤10 ng/mL at diagnosis, PSA density <0.2 ng/mL/cc, clinical stage T1c or T2, GS ≤3+3, and one or two biopsy-positive cores [4]. Clinical stage was based on DRE or imaging, not on biopsy findings. Patients with clinical T2b and T2c disease were excluded, because they represented intermediate- or high-risk disease. The number of cores obtained at the time of diagnostic biopsy was ≥6, regardless of prostate volume.

Figure 1.

Schema of grouping according to the PRIAS criteria.

Radical prostatectomy was performed using an open retropubic technique in all cases. Pathological evaluation of biopsy and surgical specimens was performed at each institution according to the Gleason grading system and pathological stage was based on the 2002 TNM classification system. We defined upstaging as any occurrence of extraprostatic extension or seminal vesicle involvement, and we defined upgrading as GS 4+3 or greater in the surgical specimen. Total tumour volume was measured at Tohoku University and Hirosaki University, determined by computer-assisted image analysis (NIH Image, developed and maintained by the National Institutes of Health, Bethesda, MD, USA) and defined as the sum of the volumes of individual cancer foci [5]. Index tumour volume was not measured. Tumour volume and insignificant cancer were analysed according to two different ‘insignificant cancer’ criteria. One was based on Epstein's criteria, in which insignificant cancer was defined as index tumour volume <0.5 mL with GS ≤6 and pT2 stage [6], and the other was based on the data from the European Randomized Study of Screening for Prostate Cancer (ERSPC), in which insignificant cancer was defined as an index tumour volume <1.3 mL or total tumour volume <2.5 mL with GS ≤6 and pT2 stage [7].

The follow-up schedule after RP involved a PSA assay every 3 months for the first 2 years and every 6 months for the following 3 years, and annually thereafter. The date of disease recurrence or PSA failure was defined as the point when serum PSA level was >0.2 ng/mL, when RP was carried out if the PSA level did not decrease below 0.2 ng/mL after surgery, or when adjuvant therapy was initiated even if PSA did not exceed 0.2 ng/mL.

All statistical analyses were carried out using jmp 9.0.2 software (SAS institute Inc., Cary, NC, USA). Clinicopathological variables were analysed using the chi-squared test. PSA recurrence-free survival was determined using the Kaplan–Meier method. The significance of the clinicopathological variables associated with upgrading, upstaging and PSA failure was assessed using the logistic regression model or the Cox proportional hazards regression model. A P value <0.05 was considered to indicate statistical significance.

Results

Two hundred and twenty-eight men with low-risk disease who underwent RP between 2000 and 2009 were identified. Of these, 17 men were excluded because of lack of prostate volume data. Of a total of 211 men with low-risk disease, 87 satisfied the PRIAS criteria (Group A), and 124 did not (Group B). Ten or more cores were taken on biopsy in 67.8% (59/87) of Group A patients and in 69.4% (86/124) of Group B patients (P = 0.64).

Table 1 shows patient characteristics, pathological outcomes and prognosis after RP in Groups A and B. In terms of pre-surgical variables, PSA, PSA density, and the number of biopsy-positive cores were significantly lower in Group A than in Group B (P < 0.01). For surgical specimens, pT3a and pT3b was present in 10.3% (9/87) and 0% (0/87), respectively, of patients in Group A and in 16.1% (20/124) and 2.4% (3/124), respectively, of patients in Group B (P = 0.08). Positive resection margins were present in 11.5% (10/87) of Group A patients and in 24.1% (31/124) of Group B patients (P = 0.02). GS upgrading to 4+3 or greater was seen in 19.5% (17/87) of Group A patients and in 29.9% (37/124) of Group B patients (P = 0.01). When the patients who underwent biopsy of <10 cores were excluded, upgrading to GS 4+3 or greater occurred in 22.0% of Group A patients and in 31.4% of Group B patients. Furthermore, upstaging to pT3 occurred in 10.2% of Group A patients and in 17.4% of Group B patients, respectively.

Table 1. Clinicopathological characteristics according to the PRIAS criteria.
 PRIAS criteria
Met (Group A)Not met (Group B)P
Patients, n87124 
Median (range) age, years67 (51–76)66 (47–76)0.74
Median (range) PSA, ng/mL5.83 (1.49–9.31)6.80 (4.1–9.9)<0.01
Biopsy-positive cores, n (%)  <0.01
162 (71.3)50 (40.3) 
225 (28.7)23 (18.5) 
3 28 (22.6) 
≥4 23 (18.5) 
PSA density, ng/mL/cc, n (%)  <0.01
 87 (100.0)22 (17.7) 
≥0.2 102 (82.3) 
Clinical T stage, n (%)   
1c66 (75.9)96 (77.4)0.87
2a21 (24.1)28 (22.6) 
Pathological T stage  0.08
278 (89.7)101 (81.5) 
3a9 (10.3)20 (16.1) 
3b0 (0.0)3 (2.4) 
Resection margin  0.02
Negative77 (88.5)93 (75.0) 
Positive10 (11.5)31 (25.0) 
Pathological Gleason score  0.01
6 or less41 (47.1)32 (25.8) 
3+428 (32.2)55 (44.4) 
4+312 (13.8)24 (19.4) 
8 or more5 (5.7)13 (10.5) 
Median (range) follow-up, months50 (2–115)42 (1–127) 
Cancer-specific death, n00 
Overall death, n01 

Tumour volume was measured in 41.4% (36/87) of patients in Group A and in 52.4% of patients (65/124) in Group B. The median (range) tumour volume was 0.81 (0.03–5.09) mL in Group A and 1.40 (0.04–8.21) mL in Group B (P < 0.01). The rate of insignificant cancer in Group A was higher than that in Group B with any criteria; however, even in Group A, the rate of insignificant cancer was only 25.0% with Epstein's criteria, and 30.6% with the ERSPC criteria (Table 2).

Table 2. Tumour volume and insignificant cancer in men with low-risk disease according to the PRIAS criteria.
 Group A (n = 36)Group B (n = 65)P
Median (range) tumour volume, mL0.81 (0.03–5.09)1.40 (0.04–8.21)<0.01
Tumour volume, n (%) <0.5 mL14 (38.9)11 (16.9)0.01
<0.5mL, GS ≤6 and pT29 (25.0)8 (12.3)0.10
<1.3 mL25 (69.4)30 (46.2)0.02
<1.3 mL, GS ≤6 and pT211 (30.6)9 (13.8)0.04
<2.5 mL34 (94.4)51 (78.5)0.04
<2.5 mL, GS ≤6 and pT211 (30.6)10 (15.4)0.07

The median and mean follow-up was 50 and 51 months in Group A, respectively, and 42 and 43 months in Group B, respectively. The 5-year PSA recurrence-free survival rate was 91.2% in Group A and 86.4% in Group B (Fig. 2; P = 0.47; log rank test). Among patients in Group A, the 5-year PSA recurrence-free survival rate with GS 4+3 or greater was 81.9%, and there was no significant difference in 5-year PSA recurrence-free survival when comparing GS 3+3 and GS 3+4 (94.4 and 92.2%, respectively [Fig. 3A]). The 5-year PSA recurrence-free survival rate was significantly lower for pathological T3 than for T2 (94.0 vs 66.7%, P < 0.01 [Fig. 3B]). There was no cancer-specific death in either group, and only one patient who did not meet the PRIAS criteria died of pneumonia at 40 months after RP.

Figure 2.

5-year PSA recurrence-free survival in Groups A and B.

Figure 3.

5-year PSA recurrence-free survival in the group meeting the PRIAS criteria according to A, GS, and B, pathological stage.

Table 3 shows the multivariate analysis of variables that could potentially predict upgrading, upstaging and PSA failure. No preoperative variables could predict upgrading and upstaging; however, PSA density was significantly associated with upstaging (odds ratio [OR] 4.198, P = 0.045). Moreover, the PRIAS criteria set was an independent predictor of decreased upstaging among those with low-risk disease (OR 8.152, P = 0.009).

Table 3. Multivariate analysis of potential predictors of upgrading, upstaging and PSA failure.
 UpgradingUpstagingPSA failure
ORP95% CIORP95% CIORP95% CI
PSA1.1420.1780.908–1.2811.0340.8450.920–1.1021.1100.4290.856–1.438
PSA density (<0.2 vs ≥0.2)1.0540.9380.725–1.4024.1980.0451.535–15.9831.0320.9730.932–1.204
Biopsy positive cores (≤2 vs >3)1.6430.3120.634–2.5251.9520.0730.930–8.6821.1060.8800.933–1.248
Clinical stage (T1c vs T2a)1.0260.9450.903–1.3151.1470.7650.856–1.4191.4090.5300.973–1.919
PRIAS criteria (yes vs no)1.9490.3870.839–3.3568.1520.0091.817–29.0921.3600.7710.904–1.609

Discussion

A strategy of AS can minimize overtreatment; however, a substantial proportion of patients have to be treated after an initial period of observation, and some patients show unfavourable pathological features at RP and so the value of AS remains controversial. Epstein et al. [6] reported that one-third of men who failed AS and underwent RP had extraprostatic extension or seminal vesicle/lymph node involvement. Most progression after AS occurs 1–2 years after diagnosis, suggesting an undersampling of more aggressive tumours rather than progression of indolent tumours [8]. No selection criteria can identify with complete accuracy truly indolent cancer that would benefit from starting AS. Long-term close follow-up of PSA screening and repeat biopsy should be performed to help detect disease progression.

The evaluation of pathological outcomes after RP in men who met the criteria for AS could help refine the selection criteria for AS. Conti et al. [9] reported that 28% of men who met the AS criteria experienced GS upgrading, 21% had extracapsular extension and 11% had seminal vesicle involvement at RP. Louie-Johnsun et al. [10] reported that 23% of 549 patients who received laparoscopic RP with low-risk disease had GS upgrading, 5% had extracapsular extension, and 0.9% had seminal vesicle invasion.

In the present study, the PRIAS criteria were able to select a very low-risk subset, as defined by low PSA level, low PSA density, and a lower number of biopsy positive cores. These characteristics corresponded with lower pathological upgrading and upstaging at RP and less PSA recurrence in the PRIAS group. Upstaging (10.3%) and upgrading (19.5%) rates were similar to those in previous reports, but if GS 3+4 was classified as upgrading, the percentage would increase to >50%, which is much higher than that seen in previous studies. Upgrading to 3+4 is considered to have a lower potential for progression and may not affect oncological outcomes. In fact, 5-year PSA recurrence-free survival rates were nearly identical between pathological GS 3+3 and 3+4 in the PRIAS group (94.4 and 92.2%, respectively), while the 5-year PSA recurrence-free survival rate was 81.9% for GS 4+3 or greater.

Tumour volume and the rate of significant cancer in the PRIAS group were much lower than those in the other group in low-risk disease; however, even in the PRIAS group, the rate of insignificant cancer was only 25.0% using the Epstein criteria and 30.6% with the ERSPC criteria. In both groups, the rate of insignificant cancer was similar when comparing the Epstein criteria with the ERSPC criteria, even despite the use of different tumour volume thresholds. These results suggest that tumour grade or stage was more important than tumour volume for the determination of disease aggressiveness in low-risk prostate cancer. In multivariate analysis, the PRIAS criteria represented an independent factor that predicted decreased upstaging (P = 0.009), but the PRIAS criteria did not predict decreased upgrading or PSA recurrence.

These results indicate that ∼20% of the PRIAS candidates had high grade (GS 4+3 or greater) or locally advanced cancer and that ∼70% had significant cancer. Although it is likely that the PRIAS criteria can select less aggressive and smaller tumours even in low-risk disease, the present data suggest that these criteria are also associated with a significant rate of underestimation at diagnosis.

Stringent criteria might be necessary to decrease the underestimation of unfavourable disease at diagnosis, because PSA doubling time or repeat biopsy during AS cannot always detect patients with such disease. Moreover, frequent repeat biopsy would increase the risk of infectious complications requiring hospitalization [11]. Diffusion-weighted or spectroscopic MRI [12, 13], saturation biopsy [14] or confirmation biopsy just before starting AS may address this problem. Recently, Ploussard et al. [15] reported that the prostate cancer antigen 3 score accurately predicted tumour volume and might be useful in selecting patients with prostate cancer for AS. These additional selection criteria could increase the accuracy of identifying indolent cancers, but would result in a lower number of men eligible for this conservative method.

Minimization of underestimation of unfavourable disease at diagnosis is a primary goal. Several AS study series have shown excellent oncological outcomes (cancer-specific survival rates of 97–100%) with various criteria and a median follow-up time of 22–82 months [16-21]. Soloway et al. [22] reported that 0/157 patients with low-risk disease showed metastasis or cause-specific death with a median follow-up of 45 months. Krakowsky et al. [23] reported that only five (1.1%) of 453 patients died from prostate cancer and that no other patients had clinical metastasis with a median follow-up of 7.5 years. This very low prostate cancer mortality rate would suggest that unfavourable features, such as upgrading or upstaging at RP, may not directly correlate with worsened oncological outcomes. Long-term follow-up over a period of decades is needed to definitively address this issue.

Active surveillance could minimize overtreatment in patients with low-risk prostate cancer, but a better balance between minimizing unexpected progression and avoiding unnecessary physical and psychological damage is needed before this strategy should be widely used. The most important goal is to avoid missing a diagnosis of life-threatening cancer; thus, close follow-up, including frequent PSA assays and repeat biopsy, is important, especially during the first few years of AS.

The present study has several limitations. Pathological examination was performed at different institutions. RP was achieved in a retropubic manner in all cases, but nerve-sparing or lymph node dissection was performed at the discretion of the individual institution. Furthermore, extended core biopsy and tumour volume measurement were not performed in all cases, and the median follow-up time of 44 months was relatively short for determination of outcomes. Nonetheless, these data provide useful information that could help guide the therapeutic approach used for men with low-risk disease.

In conclusion, the PRIAS criteria can be used to identify less aggressive and smaller tumours in men with low-risk prostate cancer; however, ∼20% of patients satisfying these criteria showed unfavourable aggressive pathological features such as GS upgrading (4+3 or more) or upstaging (pT3) and only 30% of patients meeting the criteria had insignificant cancer with tumour volume <2.5 mL, GS ≤6 and stage pT2. These data suggest that it is difficult to diagnose truly indolent cancer with complete accuracy using these criteria. Thus, a first-year protocol re-biopsy is essential to avoid unexpected progression. Additional selection criteria, such as saturation biopsy, 3-tesla MRI or new tumour markers may help increase the accuracy of these criteria, but longer follow-up periods are needed to determine definitively the value of this suggested approach.

Financial Support

Koji Mitsuzuka received funding from the Kurokawa Cancer Research Foundation.

Conflict of Interest

None declared.

Ancillary