Active Surveillance in 2020: A modied Delphi study to develop a practical guide for selecting patients with prostate cancer for Active Surveillance

Background Active surveillance (AS) is a management option for men diagnosed with lower risk prostate cancer. There is wide variation in all aspects of AS internationally, from patient selection to investigations and follow-up intervals, and a lack of clear evidence on the optimal approach to AS. This study aimed to provide guidance for clinicians from an international panel of prostate cancer experts. Methods A modied Delphi approach was undertaken, utilising two rounds of online questionnaires followed by a face-to-face workshop. Participants indicated their level of agreement with statements relating to patient selection for AS via online questionnaires on a 7-point Likert scale. Factors not achieving agreement were iteratively developed between the two rounds of questionnaires. Draft statements were presented at the face-to-face workshop for discussion and consensus building. Results 12 prostate cancer experts (9 Urologists, 2 academics, 1 radiation oncologist) participated in this study from a range of geographical regions (4 USA, 4 Europe, 4 Australia). Complete agreement on statements presented to the participants was 29.4% after Round One and 69.5% after Round Two. Following robust discussions at the face-to-face workshop, agreement was reached on the remaining statements.


Background
Prostate cancer incidence globally is increasing and prostate cancer speci c mortality is decreasing, in part due to an increase in the diagnosis of low-risk prostate cancer (1). The risk of disease progression in men with localised prostate cancer at diagnosis is mainly determined by the Gleason grade, a histopathological prostate cancer grading system (2). Patients with localised Gleason grade 3+3 prostate cancer have almost no risk of metastatic progression, and the risk increases with higher Gleason grades (3). Active Surveillance (AS) is a treatment option for patients diagnosed with localised prostate cancer with low-risk of disease progression. AS involves proactive, regular monitoring to defer or avoid radical treatments that have signi cant associated morbidity, such as prostatectomy or radiotherapy (4).
Large single centre (5,6) studies, multicentre (7,8) AS cohorts, and a large randomised controlled trial of active monitoring versus radical treatment (9) have shown that the risk of disease progression and prostate cancer mortality is very low in men with low-risk disease after long-term follow-up. These ndings were relatively consistent, despite each study employing different AS eligibility criteria and protocols.
A major challenge for urologists and prostate cancer multidisciplinary teams in identifying appropriate patients for AS is the accurate diagnosis and grading of prostate cancer. The traditional diagnostic pathway using prostate speci c antigen (PSA) testing followed by transrectal ultrasound guided biopsy (TRUS), which published AS studies to date have largely relied upon, is known to over-detect low-grade prostate cancer and under-detect high grade prostate cancer (10). The limitations of the traditional pathway can be seen in the rates of diagnoses in men with previous negative biopsies undergoing repeat TRUS biopsies (11) and the reclassi cation of 25-33% of patients on AS to higher grade disease after initial diagnosis in cohort studies (7,8). Multiparametric magnetic resonance imaging (mpMRI), which is a relatively new imaging modality that can be used in the pre-biopsy setting to detect lesions in the prostate and guide targeted biopsy, is a more accurate test (10,12) whose role in AS is still being de ned (13,14).
Numerous biomarkers and genetic mutations have been identi ed as potentially being useful for prostate cancer risk strati cation and prognostication, but the evidence base remains heterogeneous and weak (15).
Given the challenges in accurately diagnosing and classifying prostate cancer and the range of AS strategies that have been trialled to date, the wide variation in AS that currently exists in clinical practice is probably not surprising. Numerous studies have shown signi cant variation in all aspects of AS, including patient selection, follow-up protocols, and criteria for switching to radical treatments (16)(17)(18)(19).
This variation in AS practice has been shown to exist between institutions, regionally within countries, and in national and international guidelines. Whilst survey data shows that the majority of clinicians perceive AS to be effective (20), the optimal approach to selecting patients for AS is still unknown. This modi ed Delphi study aimed to develop consensus on practical guidance for clinicians in the appropriate selection of men with lower risk prostate cancer to recommend active surveillance.

Methods
This study followed a modi ed Delphi method (21) to achieve consensus amongst a group of international prostate cancer experts (See Figure 1). Participants of the 20 th Asia-Paci c Prostate Cancer Conference were invited to take part in the study via email. Basic demographic information was collected from all participants. The aim of recruitment was to include at least one participant from each continent represented at the conference.
A review of the literature was undertaken to understand the current practice of AS in different healthcare systems, and previous attempts to achieve consensus in aspects of AS. Key references were used to inform the development of the online questionnaires and the consensus statements.
Two rounds of anonymised questionnaires were delivered to participants using a secure online platform between May and July 2019. Participants were sent a link via email to complete the questionnaires, which remained open to responses for two weeks. After each round of questionnaires, the participants received a summary feedback report and were given the opportunity for comment and suggestions to feed into the iterative process of re ning statements that had not yet achieved consensus.
Participants responded to individual statements about factors affecting decision-making to offer AS on a 7-point Likert scale. Agreement for each statement was de ned by a mean/median score of 5.5 or higher on the scale. Disagreement for each statement occurred if > 33% of respondents were divergent in their views. Participants were also asked to rank the relative importance of individual factors in decisionmaking for AS, and build their preferred criteria for AS. Free text comments were also analysed to inform the draft consensus statements for discussion and agreement.
The consensus workshop was held at the 20 th Asia-Paci c Prostate Cancer Conference in Melbourne, Australia. The majority (n = 9) of Delphi study participants were in attendance. Audio-visual recording of the workshop was undertaken to document the discussions of the participants in full. The workshop was opened with a summary of the results of the online questionnaires and the areas where consensus had yet to be achieved. Draft statements were discussed and developed in the workshop, and the nal statements were circulated to all participants following the workshop for agreement.

Results
12 international prostate cancer experts participated in this modi ed Delphi study. Most participants were practicing urologists (9) and most were male (10). There were 4 participants from each of the continents represented at the 20 th Asia-Paci c Prostate Cancer Conference (North America, Europe, and Australia).
One participant did not complete Round Two of the questionnaires, and three participants were not present for the face-to-face workshop. All participants reviewed the nal consensus statements prior to publication. All eligible participants who were invited consented to participate in this study.
The initial questionnaire presented participants with 49 statements regarding individual clinical factors affecting decision-making for prostate cancer. Clear consensus was achieved on 29.4% of the statements. A ranking exercise of the relative importance of individual clinical factors showed that participants felt the Gleason score / Gleason Grade Group provided the most useful data for decisionmaking in relation to offering active surveillance for prostate cancer, however it was emphasised by multiple participants that the decision is multifactorial. Table 1 shows the results of an initial exercise in designing an AS protocol.   Table 2).

Discussion
This modi ed Delphi study sought to address current issues surrounding the consistency and quality of care for men with prostate cancer who could potentially bene t from active surveillance as a treatment modality. The panel considered a range of issues relating to AS, including underlying principles of treatment, diagnostic information, and risk assessment. The agreed consensus statements outline a relatively exible approach to offering AS, re ecting the current uncertainties and evolving evidence around the optimal use and delivery of AS.

Key principles of Active surveillance
There were a number of key principles on which the participants held clear agreement. The panel agreed that a patient with prostate cancer should be in a reasonable state of health in order to bene t from active surveillance, and that their life expectancy, medical co-morbidities, suitability for radical treatment, and treatment preferences should all be taken into consideration for management decision-making. In the context of the known limitations of the current prostate cancer diagnostic patient, with the potential for over-diagnosis and misclassi cation of prostate cancer, a rigorous approach to AS is critical to avoid over-treatment and the associated adverse effects for men. The panel also felt that an optimal AS followup protocol would be able to accurately identify disease progression to inform decisions about switching to radical treatment.

Diagnostic information
The importance of accurate diagnostic information was discussed at length by the panel, underlying the key role that this data plays in identifying men who are potentially appropriate for AS. This is exempli ed in the two cases captured in Table 3. The index of suspicion following a negative mpMRI and the need to consider biopsy in these men with some similar characteristics is clinically different, and underlines the importance of thorough investigation to make a clear diagnosis of prostate cancer. There was strong consensus on the minimum set of diagnostic tests needed to inform the decision about offering AS (see Table 2), however other available tests such as Free:Total PSA ratio and the number of positive cores were more contentious. Multiparametric MRI was felt to be of high importance, and is increasingly being used as a diagnostic test for prostate cancer; however pre-biopsy mpMRI has not yet been recommended in many national level guidelines (17,22). The number of positive cores and positive core length can be in uenced by the location of the biopsy sampling and the total number of cores taken, and the relative importance of these factors was not agreed upon by the panel. Most prostate cancer guidelines do not include the number of positive cores or the percentage of cancer per core in prostate cancer risk de nitions, although many individual centres still include these measures in their AS eligibility criteria (17,18). The panel generally felt that Transrectal Ultrasound-guided biopsy (TRUS) alone was not su ciently accurate to be used as the basis for histopathological assessment of a prostate cancer, and ideally a Transperineal (TP) biopsy approach was followed. This re ects the general trend towards increasing use of TP biopsy due to its lower adverse effect pro le and ability to access all areas of the prostate (23). The need for a con rmatory biopsy in tertiary centre referrals was also debated, with no clear opinion reached; however the importance of concordance between mpMRI and biopsy ndings was felt to be vital in reassuring the clinicians and the patient that an accurate diagnosis has been made.

Prognostication of localised prostate cancer
A clear challenge for clinicians treating men with localised prostate cancer at the present time is accurately assessing the risk of disease progression in men with Gleason score 3 + 4 / Gleason Grade Group 2, and recommending the appropriate treatment. The risks of morbidity from radical treatment versus the risks of disease progression on AS are often nely balanced. This was a particularly contentious issue for the panel, and no consensus was reached. The agreed statements in Table 2 could be considered conservative in their approach, but the majority of panel members were more comfortable with not recommending AS as the optimal treatment option to this patient given there is currently no reliable way of distinguishing lethal from non-lethal prostate cancer for these men.

Strengths and limitations
This modi ed Delphi study was conducted in a methodologically rigorous manner. A diverse panel of international prostate cancer experts was assembled, with an optimal number of participants who engaged throughout the whole process. Agreement was achieved on the consensus statements through robust discussion and iterative work over two rounds of questionnaires, followed by a face-to-face workshop to re ne the statements (24). The panel has developed a set of practical recommendations that take into consideration the latest evidence in the eld. The role of certain tests, including number of positive cores and Free:Total PSA ratio, in patient selection for AS was not agreed by the panel, although this could be considered to be a re ection of the current state of the evidence in these areas (15).

Conclusions
Active surveillance is an appropriate treatment option for men with localised, low-risk prostate cancer. It can be utilised to achieve good outcomes for patients and avoid overtreatment. It is vital that complete and accurate diagnostic information is obtained to ensure that the correct patients are recommended to undergo AS, and a patient's suitability and treatment preferences are factored into the shared decisionmaking process. The ndings of cohort studies of patients receiving AS that are currently on-going will help to ll some of the evidence gaps around patient selection and follow-up protocols that exist today.