- To establish the rate of higher risk criteria in various definitions of an active surveillance population.
biochemical recurrence-free survival
Memorial Sloan Kettering Cancer Center
Prospective Randomized International Active Surveillance
seminal vesical involvment
University of California San Francisco
In the last two decades, the screening of PSA levels has increased the detection of insignificant prostate cancer (PCa) that would have not progressed if left untreated . In this setting, active surveillance (AS) has emerged as a treatment option for patients with localized PCa and could reduce the burden of treatment without missing the opportunity for cure when the disease evolves [2-4]. Cancers that are amenable to AS usually are identified on favourable preoperative parameters and the risk is estimated by integrating Gleason score, pretreatment PSA level, clinical stage, prostate volume and the extent of biopsy involvement with the tumour. Published AS series use different criteria largely based on the experiences and preferences of individual centres. The most common clinical data used to define AS criteria are a Gleason score ≤6, PSA level ≤10 ng/mL and clinical stage T1c disease. The PSA density is noted as inclusion criteria in some studies with different reported thresholds for AS inclusion. Other characteristics to consider include pathological biopsy parameters with a wide variation with respect to AS inclusion criteria. The currently available prospective AS studies agree on the need for a frequent monitoring of PSA levels and repeat biopsy to detect tumours that have progressed or were initially under diagnosed , There is some uncertainty about the inclusion criteria and their reliability in predicting favourable outcome. Studies comparing entry criteria for AS protocols (especially in terms of biopsy parameters) are needed to clarify the best candidates for AS. Recent studies have emphasized the risk of under-diagnosis and adverse pathological findings and, thus, missing the window of curability if AS is preferred [6, 7]. Although only the long-term results of prospective AS studies will answer these questions, the present study aimed to analyze the currently available AS protocols and determine patient selection and pathological outcomes in a radical prostatectomy (RP) series.
After obtaining institutional ethics board approval, we reviewed our database from January 2001 until May 2011 for patients who underwent RP at our institution. All patients underwent clinical evaluation, including DRE, serum PSA levels and TRUS. Preoperative MRI of the prostate was not available for all patients. Only patients who underwent a prostate biopsy scheme at our department were included to limit bias as a result of inter-observer variability in the pathological assessment. In total, 1161 patients were included in the present study. The biopsies were performed in the order: six sextant biopsies (standard 45° angle); three biopsies in each peripheral zone (80° angle); three biopsies in each transition zone; and, finally, three biopsies in the midline peripheral zone [8-10]. All biopsies and radical prostatectomies were performed at the Department of Urology. Pathological evaluation of biopsy cores and prostatectomy specimens was carried out by senior uropathologists at our institution. Patients underwent clinical and biological follow-up at 1 month after surgery then every 3 months for the first year and biannually thereafter. Biochemical recurrence was defined as a PSA level >0.2 ng/mL. Data from clinical evaluation, biopsy, RP specimens and follow-up were recorded in a prospective database.
For the purpose of the present study, we selected the criteria of six prospective AS protocols: the University of Toronto , Royal Marsden [11, 12], John Hopkins [13, 14], University of California San Francisco (UCSF) , Memorial Sloan Kettering Cancer Center (MSKCC) [16, 17] and Prospective Randomized International Active Surveillance (PRIAS) . These criteria were retrospectively applied to our patients in the present study, who were divided into six groups based on their preoperative PSA level, PSA density (PSAD), clinical stage, biopsy Gleason score, number of positive biopsies and percentage positive core involvement when required (Table 1). The pathological data were analyzed for extracapsular extension (ECE), seminal vesical involvment (SVI), upgrading of Gleason score from biopsy to pathology, and unfavourable disease defined as stage pT3 and/or Gleason score >6.
|Protocol||Gleason||PSA level (ng/mL)||PSAD (ng/mL per mL)||Clinical stage||Positive biopsies||Percentage single core involvement|
|University of Toronto ||≤3+3||≤10||–||–||–||–|
|Royal Marsden ||≤3+4||≤15||–||T1/T2a||≤50%||–|
|John Hopkins [13, 14]||≤6||–||≤0.15||T1||≤2||≤50|
|UCSF ||≤6||≤10||–||T1/T2||≤⅓ of biopsies||≤50|
Biochemical recurrence-free survival (bRFS) was also evaluated in the six groups of patients, as well as the risk factors involved.
Statistical analysis was performed comparing rates of upgrading, ECE, SVI and unfavourable disease (Gleason score upgrading >6 and/or T3 disease) for the six groups of patients. John Hopkins' criteria were used as a reference for this comparison. Either Student's t-test or the Mann–Whitney U-test was used for continuous variables and Pearson's chi-squared test was used for categorical variables. bRFS was established using the Kaplan–Meier method. Curves were tested with the log-rank test. P < 0.05 was considered statistically significant. SPSS, version 13.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis.
The mean patient age in the studied cohort was 63.4 years, mean PSA level was 9.65 ng/mL and mean PSAD was 0.21 ng/mL per mL. The mean (median; range) number of positive cores was 5 (3; 1–21). In total, 83.3% of patients had clinical stage T1c, and the biopsy Gleason score was ≤6 in 776 (67%) patients. Of the 1161 patients included in the present study, 726 (62.5%) were eligible for the Royal Marsden criteria, whereas only 191 (16.4%) patients had the John Hopkins AS criteria. Table 2 summarizes the clinical characteristics for each group of patients.
|Variable||All patients||University of Toronto||Royal Marsden||John Hopkins||UCSF||MSKCC||PRIAS|
|Overall, n (%)||1161 (100)||338 (29.1)||726 (62.5)||191 (16.4)||395 (34)||368 (31.7)||306 (26.3)|
|Age (years), mean (median; range)||63.4 (63.7; 42.8–78.2)||63.23 (63.5; 58.7–68.1)||63.46 (64; 58.6–68.7)||63.99 (64.54; 47.63–76.9)||63.1 (63.44; 42.9–76.9)||63.15 (63.5; 58.3–68.2)||63.6 (64; 42.9–77)|
|PSA level (ng/mL), mean (median; range)||9.65 (6.9; 0.35–99)||6.1 (5.8; 4.8–7.6)||6.9 (6.1; 4.9–8.5)||6.27 (5.5; 1.64–19)||6.19 (5.8; 0.35–10)||6.1 (5.8; 4.7–7.5)||5.9 (5.6; 0.35–10)|
|PSAD (ng/mL per mL), mean (median; range)||0.21 (0.15; 0.01–2.86)||0.14 (0.12; 0.08–0.17)||0.15 (0.13; 0.08–0.18)||0.1 (0.1; 0.04–0.15)||0.14 (0.12; 0.01–0.46)||0.13 (0.12; 0.08–0.17)||0.11 (0.11; 0.01–0.19)|
|Number of positive cores, mean (median; range)||5 (3; 1–21)||2 (1; 1–3)||1 (1; 1–5)||1.29 (1; 1–2)||1.91 (1; 1–7)||1 (1; 1–3)||1.33 (1; 1–2)|
|Clinical stage, n (%)|
|T1c||967 (83.3)||287 (85)||639 (88)||191 (100)||342 (86)||314 (85)||270 (88)|
|T2||173 (14.8)||51 (15)||87 (12)||–||54 (14)||54 (15)||36 (22)|
|Biopsy Gleason score, n (%)|
|≤6||776 (67)||338 (100)||601 (83)||191 (100)||396 (100)||368 (100)||306 (100)|
|3+4||238 (20.5)||–||125 (17)||–||–||–||–|
Pathological study of RP specimens showed ECE in 414 (35.7%) patients and SVI in 102 (8.8%) patients. There were 383 (33%) patients who had a Gleason score upgrade >6 and, in total, 483 (41.6%) patients had unfavourable disease (Table 3). In patients who had the criteria for AS protocols, ECE rates varied between 24% (Royal Marsden) and 14.6% (John Hopkins), with a significant difference between the less stringent and most stringent criteria (P = 0.009). SVI rates were between 3% (Royal Marsden) and 1% (John Hopkins; PRIAS) (P = 0.12). Gleason score upgrading >6 varied from 50% (MSKCC; University of Toronto) to 40% (Royal Marsden) (Fig. 1) (not significant). However, when the limit of Gleason score upgrading was defined as 7 (3+4) these rates dropped to 20% (Royal Marsden) and 10.5% (John Hopkins) (P = 0.004) (Fig. 2). Unfavourable disease defined as Gleason score upgrading >6 and/or T3 disease was found in more than 40% of patients in all series (Fig. 1 and Table 3). There was no significant difference except for the Royal Marsden criteria (P = 0.001).
|Variable||All patients, n (%)||University of Toronto, n (%); P||Royal Marsden, n (%); P||John Hopkins, n (%)||UCSF, n (%); P||MSKCC, n (%); P||PRIAS, n (%); P|
|ECE||414 (35.7)||67 (19.8);||170 (24);||28 (14.6)||77 (19.5);||74 (20);||55 (18);|
|P = 0.15||P = 0.009||–||P = 0.16||P = 0.24||P = 0.27|
|SVI||102 (8.8)||5 (1.5);||21 (3);||2 (1)||5 (1.3);||5 (1.3);||3 (1);|
|P = 0.99||P = 0.12||–||P = 0.99||P = 0.99||P = 0.56|
|Pathological Gleason score|
|6||394 (33.9)||159 (47)||306 (42)||106 (55.5)||195 (50)||177 (48)||158 (51.7)|
|7||648 (55.8)||169 (50)||386 (53)||82 (43)||190 (48)||179 (48)||144 (47)|
|3+4||388 (33.4)||125 (37)||268 (37)||64 (34)||139 (35)||134 (36)||111 (36)|
|4+3||261 (22.5)||44 (13)||117 (16)||18 (9)||51 (13)||45 (12)||33 (11)|
|8–10||119 (10.3)||8 (2.4)||28 (3.9)||3 (1.5)||10 (2)||9 (2.5)||4 (1.3)|
|GS upgrading >6||383 (33)||169 (50);||289 (40);||85 (44.5)||190 (48);||188 (51);||144 (47);|
|P = 0.23||P = 0.24||–||P = 0.42||P = 0.15||P = 0.64|
|GS upgrading >3+4||329 (28.3)||52 (15.4);||145 (20);||21 (10.5)||61 (15);||54 (14.5);||37 (12.3);|
|P = 0.18||P = 0.004||–||P = 0.16||P = 0.24||P = 0.77|
|Unfavourable disease||483 (41.6)||174 (51.5);||305 (42);||87 (45.5)||196 (50);||195 (53);||151 (49);|
|P = 0.2||P = 0.001||–||P = 0.37||P = 0.1||P = 0.4|
The cohort had a median (interquartile range) follow-up of 2.11 (0.99–3.93) years. The bRFS at 5 and 10 years was 76.7% and 63.3%, respectively, for the entire cohort. Positive margins (P < 0.001), pT3 tumours (P = 0.006) and unfavourable disease (P < 0.001) were significant predictors of biochemical recurrence (Table 4). There was no significant difference in bRFS between the six groups of patients who were eligible for AS.
|Cohort characteristic||5 years (% bFRS)||10 years (% bFRS)||P|
|Positive surgical margins||54.4||31||<0.001|
In the era of widespread screening of PSA levels, AS has become a viable option for patients with localized PCa. Selecting the right patients for AS is still a dilemma, and varying criteria are used depending upon the AS protocol. Until the long-term results of prospective randomized AS protocols are available, upgrading and upstaging remain an interesting parameter for evaluating AS criteria. For the present study, we selected six AS protocols that cover a wide range of selection criteria (Table 1) to compare patient selection and the risk of missing unfavourable disease in patients undergoing RP at our institution.
We found that 726 (62.5%) patients were eligible for the Royal Marsden criteria. Pathological results showed significantly higher ECE, upstaging (> Gleason 3+4) and unfavourable disease compared to the John Hopkins criteria. This protocol included patients with low and intermediate risk PCa and was designed to be wide in the context of a clinical trial. A recent study showed that delayed RP for intermediate risk PCa increased the risk of biochemical recurrence and positive surgical margins . Furthermore, the early results of prospective AS studies suggest that the inclusion of patients with Gleason 3+4 should be reserved for low volume tumours and in older patients, particularly those with comorbid conditions [2, 20]. More stringent criteria restricted the number of patients included to 306 (26.3%) PRIAS and 191 (16.4%) John Hopkins. The same trend was found by Conti et al. , although John Hopkins was found to be more selective, with only 4% of patients. Another study testing the PRIAS and John Hopkins criteria found inclusion rates of 9.7% and 6.9%, respectively . In a recent study by Iremashvili et al. , patient selection was much lower at 48.6% for PRIAS and 27.9% for John Hopkins . These large differences in published reports reflect the heterogeneity of the populations under study.
Pathological results after RP showed reduced rates of ECE and SVI for patients eligible for AS compared to the entire cohort (Table 3). The PRIAS criteria resulted in 18% ECE and 1% SVI, which is comparable to our previous study regarding PRIAS . With the added selectivity in John Hopkins' criteria that included PSAD <0.15 and percentage core involvement, there was a 15.6% rate of upstaging, which was the lowest observed in the present study (14.6% ECE and 1% SVI). This value is high compared to other studies where rates of 0– 9% were described for the same criteria [6, 7, 23]. This could be explained by the higher rate of upstaging in our cohort 44.5% in the present study compared to previous studies, which ranged from 16.3% to 33% [6, 7, 23]. It is interesting to see that the UCSF criteria had a higher upstaging at 20.8% (19.5% ECE and 1.3% SVI) than would have been expected. This is probably a result of the inclusion of patients with up to seven positive biopsies because our patients in the present study had a total of 21 biopsies and the UCSF inclusion criteria for positive biopsies is less than or equal to one-third (positive biopsies: mean [range] 1.91 [1-7]).
The present study also investigated upgrading of the Gleason score to more than 6. Although the rate was 33% for the entire cohort, this number was higher for patients qualifying for AS and, amongst those patients, the highest rates of upgrading were found with the protocols that did not include biopsy Gleason 7 patients (Fig. 1 and Table 3). However, when upgrading was considered as Gleason >3+4, the rate for patients eligible for AS was lower than that of the cohort and was the lowest at 12.3% (PRIAS) and 10.5% (John Hopkins) for the most stringent AS protocols. Figure 2 illustrates the difference between the two definitions of Gleason score upgrading, and shows that most of the upgrading for PRIAS, UCSF and John Hopkins is to a Gleason score of 7 (3+4) compared to Royal Marsden (P = 0.004). Further analysis of the pathological data confirms a Gleason score ≥8 for these criteria at a rate <2% (Table 3).
Looking at these results, it can be argued that the inclusion in AS protocols of patients with a Gleason score <7 confers a lower risk of missing high-grade disease and that most upgraded patients at repeat biopsy will have an intermediate risk PCa that can be cured.
Despite the differences in selectivity, upstaging and upgrading between the studied AS protocols, the overall risk of missing unfavourable disease (stage ≥ T3 and/or Gleason score >6) was at least 40% for all groups. Several studies have suggested the benefit of early repeat biopsy  or transperineal template biopsy  to reduce the risk of undersampling. Nomograms incorporating MRI and clinical data could allow the better selection of patients eligible for AS .
If the differences in results between the less stringent Royal Marsden and the most stringent John Hopkins' criteria are identified, the four remaining AS protocols had almost similar outcomes, regardless of the disparity in inclusion criteria (Fig. 1 and Table 3). Selection for AS should constitute an acceptable trade-off between rates of eligibility and rates of inclusion of significant/indolent disease, in terms of morbidity and mortality rates. Wrongly included patients who actually harbour aggressive disease will be filtered out during the surveillance protocol. Furthermore, the inclusion criteria can be widened for patients of higher age and comorbidity.
With a short median follow-up of 2.11 years, there was no difference in bFRS amongst the studied populations, although there appears to be a trend for better long-term results for the more selective criteria. However, these findings should be interpreted carefully and a longer follow-up is needed to confirm these results.
We acknowledge that the present study had some limitations, such as a lack of tumour volume measurement that would have allowed better discrimination between AS protocols. Another limitation is the inclusion of patients before the 2005 International Society of Urological Pathology modified Gleason score system, in that such patients may have more aggressive tumour characteristics compared to contemporary patients. This could have increased the risk of ugrading and upstaging. However, all patients had the same extended 21 core biopsy scheme at our institution, thus reducing bias. This biopsy protocol also significantly improved the PCa detection rate and reduced the risk of undersampling .
In conclusion, there is a varying degree of selectivity in currently adopted AS protocols. The inclusion of patients with Gleason 3+4 presents a risk of missing significant disease and should be limited to older patients with comorbidities. Although the added selectivity of John Hopkins criteria with PSAD ≤0.15 provided the lowest rate of ECE and SVI, there was no significant difference in the rate of unfavourable disease with the other Gleason 6 protocols. There is a low probability that the differences in detailed criteria would have an impact on long-term follow-up. Future studies need to focus on novel biomarkers and imaging techniques that can be integrated in the selection process and allow the safer inclusion of patients with insignificant PCa.