Risk-Based Selection for Active Surveillance

Fig. 6.1
Definition of overdiagnosis (From Etzoioni [55]. Used with permission)


Fig. 6.2
Association of different treatment strategies with life expectancy and risk of PCa progression. The downhill slope between AT and AS with increasing life expectancy illustrates that AT is feasible in young men. The downhill slope of WW with increasing life expectancy illustrates that the 10-year cutoff is arbitrary; a man with high risk of PCa progression with a very short life expectancy is still suitable for WW. In contrast, men with more than a 10-year life expectancy at very low risk of PCa progression may also be suitable for WW (Adapted from Bruinsma S and Nieboer D with permission. Figure 14.4 in this book)

Patient Selection for Active Surveillance: Qualitative Considerations

The main goal of AS is to reduce overtreatment in patients with low-risk prostate cancer [6, 7]. Men with an initially low risk of PCa progression are usually considered as candidates for AS, provided they have a reasonable life expectancy, e.g., more than 10 years. They should be distinguished from men diagnosed with a progressive prostate cancer who are more likely to die from PCa and would benefit substantially from immediate active treatment [5, 8]. On the other hand, we should not select men for AS if their lifetime risk of disease progression is very low. In these cases, watchful waiting would be the optimal treatment option. Figure 6.3 shows the timeline between diagnosis of PCa and death due to PCa or due to other causes for three exemplary patients with various absolute risks of PCa progression , based on cancer risk and life expectancy. The most beneficial strategy for the particular patient is colored green: watchful waiting [WW], active surveillance [AS], or active treatment [AT].


Fig. 6.3
Timelines of possible outcomes after watchful waiting (WW) , active surveillance (AS), and active treatment (AT) for three exemplary patients with various risks of PCa progression. The colored boxes display the time between PCa diagnosis and time of death; the colors indicate the most beneficial (green), intermediate (yellow), and most harmful (red) treatment strategy for each specific patient

For example, patient 1A with PCa clinical stage T1cNxMx, Gleason 3 + 3 at diagnosis, and no PCa progression anticipated during his normal life span will die from other causes and not from PCa. This patient is therefore categorized as a patient with a “very low risk” of progression of PCa. WW is the optimal management in this case.

Patient 1B is an 80-year-old man with cardiovascular comorbidity and a life expectancy of less than 10 years. This patient has probably a higher chance of dying from his cardiovascular conditions than from prostate cancer even though he has a relatively high risk of progression. In his case, we expect no benefit from AS, only sustaining the negative parts of AS (e.g., prostate biopsy, risk of infection), so WW is best.

Patient 2 has a similar clinical stage T1cNxMx and Gleason 3 + 3 at diagnosis but is expected to show disease progression. If patient 2 had received WW, this progression would be missed, leading to a suboptimal treatment later on. AS would have been the preferable strategy, as disease progression would be detected on time, with PCa still in its curable stage. If thus patient received AT immediately, he would most likely be cured. The delayed treatment with AS compared to immediate AT implies that the man will sustain a better quality of life for the years before treatment.

For patient 3, diagnosed with PCa T2bNxMx and Gleason 4 + 4, AS was in fact unsuitable, because disease progression would be detected at too late a stage. This would lead to a poor clinical course compared to the situation of this patient receiving immediate AT.

An illustrative article provided an attractive analogy between the progress of disease and the speed of locomotion of turtles, rabbits, and birds [9]. In our examples, patient 1A is the turtle (very slow-growing disease and dies likely from a different cause than his prostate cancer). AT would have been unnecessary. AS would have prevented him undergoing surgical treatment, but he would still have to experience the discomfort of prostate biopsies and regular checkups during AS. Ideally, his prostate cancer would not have been detected, but now that it has been, he should receive nothing more than WW.

The other extreme is the bird; in this case the diagnosis was made too late for treatment, and the man is likely to die from prostate cancer. Finally, the rabbit represents the man with PCa who needs to be diagnosed. His risk of progression is higher than that of the turtle and could be life threatening, while the disease is still curable.

Table 6.1 summarizes the risks and benefits of the different patients on AS. Patients 2 and 3 can be considered as rabbits. Note that AT, e.g., surgery, is only immediately required for an intermediate-risk or higher-risk PCa, with substantial life expectancy (patient 3), where treatment delay is anticipated to be harmful.

Table 6.1
Summary of harms and benefits in three different candidates for active surveillance


WW watchful waiting, AS active surveillance, AT active treatment

Patient Selection for Active Surveillance: Endpoints

Clinically relevant endpoints such as time to metastasis or disease-specific mortality should preferably be the main outcomes when deciding which treatment strategy a patient should receive [10, 11]. These endpoints imply that a long-term follow-up period of at least a decade is necessary due to the slow-growing nature of PCa [12]. A more practical endpoint is “progression of PCa” as a proxy outcome. No uniform definition of disease progression is available [12]. Progression can be defined on repeat biopsy findings, RP, or as treatment-free survival [11]. Epstein defines upgrading of the Gleason score at radical prostatectomy as disease progression [13]. Others use biochemically determined recurrence (PSA rise) or presence of distant metastasis indicated by changes in PSA, DRE, tumor grade, and tumor volume on biopsy findings or even magnetic resonance imaging [8, 13].

We note that tissue sampling errors may become apparent from repeat biopsy, often performed within 1 year after the initially positive biopsy. These are usually considered to be different from true disease progression and are labeled “reclassification” [8, 14, 15].

Comparison of Monitoring to Immediate Treatment

Findings from the recent ProtecT screening study suggest that two-thirds of the patients diagnosed with PCa may be eligible for AS [16, 17]. The ProtecT trial compared radiotherapy, surgery, and active monitoring [18]. Active monitoring is a variant between WW and AS. The study has a less restrictive entry criteria (allows inclusion of patients with GS ≥ 8), and the follow-up scheme is less rigorous compared to contemporary AS protocols. The ProtecT study demonstrated no survival benefit in men for the different treatment options (Table 6.2). However, the results at 10-year follow-up imply less disease progression and PCa metastases after surgery and radiotherapy compared to active monitoring. These results should be interpreted with caution since, as said, the study used an active monitoring protocol instead of a more modern AS protocol. For example the inclusion of patients with GS ≥ 8 in an active monitoring group will result in poorer outcome. Two other major randomized controlled trials have been conducted but compared surgery with WW, rather than AS. The first study is the Prostate Cancer Intervention Versus Observation Trial (PIVOT) : patient enrollment started in 1994, a total of 731 patients were randomized, and 26% of the patients had GS ≥7 on initial biopsy [19]. After 12-year follow-up, the study showed no overall or PCa-specific survival benefit to RP compared to WW. On the contrary, the Scandinavian Prostate Cancer Group Study Number 4 (SPCG-4) had a much longer follow-up of 23 years, randomized 695 patients, and found lower disease-specific mortality for RP compared to WW [20]. Unfortunately, no analysis has yet been reported for the more typical contemporary AS candidates.

Table 6.2
Main outcomes of the ProtecT trial comparing different treatment strategies

Active monitoring

N = 545


N = 554


N = 545

PCa-specific survival – % (95% CI) at 10 years

98.8% (97.4–99.5)

99.0 (97.2–99.6)

99.6 (98.4–99.9)

Incidence of metastatic PCa per 1000 person-year (95% CI)

6.3 (4.5–8.8)

2.4 (1.4–4.2)

3.0 (1.9–4.9)

Traditional Selection for Active Surveillance

Inclusion criteria for AS patients are different between the major international AS cohorts (Table 6.3) [7, 11, 15, 2128]. Furthermore, AS cohorts differ by protocol and in practical implementation [2]. For example, the Gleason grading system changed in 2014 [29]. This means that patients included in 1995 based on the Epstein inclusion criteria could have had a higher risk than those included in the more recent cohorts. Overall, current guidelines recommend patients as being the most suitable for AS if they have pretreatment clinical stage T1(c) or T2a prostate cancer, serum PSA <10 ng/ml, a biopsy Gleason score of 6, a maximum of 2 tumor-positive biopsy core samples, and/or a maximum of 50% of cancer per core [4]. Some guidelines include statements that patients with stage T2b–T2c can also be recommended for AS. The Dutch Urology Association (DUA) guideline even recommends selecting patients with T3 for AS. Age and comorbidity are relevant, because a considerable life expectancy is important for AT, and hence AS, to show any long-term benefit. Finally, some guidelines state that patients’ preference should be considered in order to reduce the dropout rate of AS patients due to anxiety [4].

Table 6.3
Patient- and biopsy-based inclusion criteria in different active surveillance protocols


In green, similarities between the cohorts are shown resulting in a more stringent inclusion of “low risk.” The yellow color shows a wider range of criteria, which allows for a slightly higher-risk inclusion

PSAD prostate-specific antigen density; PRIAS Prostate Cancer Research International Active Surveillance; UCSF University of California, San Francisco; UM University of Miami; UC University of Copenhagen; MSKCC Memorial Sloan Kettering Cancer Center New York

aIncludes Gleason ≤3 + 4 and/or PSA ≤ 20 if life expectancy ≤10 years or present comorbidities

bIncludes patients with Gleason ≤3 + 4 if age > 65 years and PSA <15

cT2a only if PSA ≤10

Risk-Based Selection for Active Surveillance

In this section, we review studies that may improve AS risk-based selection for men with PCa using nomograms or risk calculators predicting PCa progression. Generally, the studies are similar in terms of patient’s inclusion criteria, predictors, and primary outcome (Table 6.4). The primary outcome was progression, defined as upgrading of GS at radical prostatectomy, except for the Canary Prostate Active Surveillance Study which used GS upgrading on a follow-up biopsy [30]. Nomograms were constructed by logistic regression analysis. However, the Johns Hopkins AS study used a Bayesian joint model including all biopsy data during AS to improve prediction of GS ≥ 7 at RP [31]. The studies are described in more detail below.

Table 6.4
Comparison of predictive ability on the risk of prostate cancer for men potentially eligible for active surveillance


No. of men

Inclusion criteria





95% CI




Kattan 2003 [33]


T1c-T2A, PSA ≤ 20, GS ≤6 on Bx, ≤50% positive cores, ≤20 mm total cancer, and at ≥40 mm benign in all cores

Tumor volume < 0.5 cc, confined to prostate and GS ≤ 6 on RP

PSA, TRUS PV, T-stage, GS, and total length of cancer in biopsy cores

0.79 (−)

0.59–0.79 [39]

Steyerberg 2007 [34]


Same criteria as Kattan

Same outcome as Kattan

Same predictors as Kattan

0.76 (0.70–0.82)


Ankerst 2015 [30]


GS ≤6, PSA ≤ 20

GS ≥7 or ≥34% positive biopsy cores on follow-up biopsy

Age, month since last biopsy, latest PSA, %positive cores, prior negative biopsy

0.72 (−)


Johns Hopkins 2016 [31]


T1c-T2a (PSA < 10), PSAD < 0.15 ng/mL, GS ≤6, ≤2 positive biopsy cores, ≤50% positive per core

GS ≥7 on RP

Joint model: PSA, age, PV, biopsy information, RP information

0.74 (0.66–0.81)


Truong 2013 [38]


Gleason ≤6 on biopsy and at least 10 cores

GS ≥ 7 on RP

PSA TRUS density, obesity, no. of positive cores, max. no. of cores involved

0.75 (0.69–0.82)


AUC area under curve of the receiver operator characteristic curve as a measure of the predictive model performance, RP radical prostatectomy, NP not performed

European Randomized Study of Screening for Prostate Cancer

In a sub-study of the European Randomized Study of Screening for Prostate Cancer (ERSPC), 864 patients were evaluated with a mean follow-up of almost 9 years; the patients had undergone radical prostatectomy and 619 patients had a Gleason score of ≤6 and cT1-2 at diagnosis [32]. At radical prostatectomy, 66% of these low-risk patients still had a Gleason pattern ≤3 + 3 and ≤pT2 on final pathology. To distinguish between low-risk PCa (Gleason score ≤ 3 + 3 and ≤pT2 at radical prostatectomy) and significant PCa, a nomogram originally developed by Kattan et al. in 2003 was used. This prediction model was validated and updated by Steyerberg et al. in 2007 for use in a screening setting [33, 34]. Predictors were serum PSA, TRUS prostate volume, clinical stage, prostate biopsy Gleason score, and total amount of cancer and non-cancer tissue in biopsy cores. The performance of the model was moderate, with an area under the ROC curve between 0.60 and 0.69 at external validation [32, 35].

Canary Prostate Active Surveillance Study

The Canary Prostate Active Surveillance Study (PASS) cohort is a prospective and observational AS study. The investigators defined progression as either a Gleason score upgrade from 6 to ≥7 or an increase in percentage of cancer cores positive for cancer from <34% to ≥34% at repeat biopsy. The progression rate ranged from 10% to 30% at sequential biopsies (first, second, third biopsy), and the overall progression rate was 24% after 28 months of follow-up among 905 men [30, 36]. Noteworthy is that 55 (8%) of the 689 men without disease progression opted for active treatment. PSA, percentage of cores positive for cancer on most recent biopsy, and history of at least one prior negative biopsy were associated with progression. Together with age and number of months since last biopsy , the model had an AUC of 0.72 at internal validation [30].

Only gold members can continue reading. Log In or Register to continue

Feb 9, 2018 | Posted by in Uncategorized | Comments Off on Risk-Based Selection for Active Surveillance
Premium Wordpress Themes by UFO Themes