Christopher Wallis: Hello, and thank you for joining us for this UroToday Journal Club. Today we're discussing a recent publication from the ARTISTIC Collaborative, entitled "Adjuvant or Early Salvage Radiotherapy for the Treatment of Localized and Locally Advanced Prostate Cancer; a Prospectively Planned Systematic Review and Meta-Analysis of Aggregate Data". I'm Chris Wallis, a Fellow in urologic oncology at Vanderbilt, and with me is Zach Klaassen, Assistant Professor at the Medical College of Georgia. Here, you can see the citation that we are going to be discussing, which was recently published by the ARTISTIC meta-analysis group in the Lancet.

So by way of background, most UroToday viewers will know that there are a plethora of trials looking at the role of adjuvant radiotherapy in prostate cancer. These trials were, for the most part, accrued and reported in the early 2000s. And across the board, we have comparisons of adjuvant radiotherapy with a watchful waiting type approach in the control arm.

And so you can see two publications each from the SWOG trial, the EORTC 22911 trial, and the ARO trial, and more recently published only in 2019, is the FinnProstate group trial. Recently we pooled these results and so this is a recent systematic review and meta-analysis, and what you can see here is that there is a consistent benefit in terms of adjuvant radiotherapy compared to observation in terms of biochemical progression-free survival, but the other more clinically relevant outcomes like metastasis-free survival and overall survival are less clear. On the basis of the data presented, so far today, we have AUA and ASTRO guidelines on adjuvant and salvage radiotherapy. And so the conclusion here is important. The group concluded that physicians should offer adjuvant radiotherapy to patients with adverse pathological findings that are, seminal vesicle invasion, positive surgical margins, or extra prosthetic extension.

However, the utilization of adjuvant radiotherapy is quite poor and certainly is under one-fifth of eligible patients. This is because most physicians take the alternative approach using salvage radiotherapy. This spares many men the toxicity and burden associated with adjuvant radiotherapy while seeking to rescue or salvage those men who have biochemical recurrence following surgery. And so the goal of this ARTISTIC collaborative is to compare outcomes for adjuvant radiotherapy compared to a strategy of planned early salvage radiotherapy following radical prostatectomy. And to do this, they relied on three randomized control trials which were concurrently running, and they are the RADICALS-RT, the RAVES trial, and the GETUG-AFU 17 trial. And there are some differences between these trials, which Dr. Klaassen will discuss in more detail. But the overall study question was the same comparing adjuvant versus early salvage radiotherapy. And so the ARTISTIC collaborative undertook a prospective systematic review and meta-analysis utilizing the FAME approach, and that is the framework for adaptive meta-analysis.

And to do this, they looked for trials with patient-level inclusion criteria including men aged 18 years and older, who had intermediate and high-risk prostate cancer without evidence of distant metastasis and had undergone prior radical prostatectomy. Post-operatively all men had to have an undetectable PSA, and this differs somewhat from the prior trials we alluded to at the beginning. And then they had to have one or more high-risk features for biochemical recurrence, and these include pT3 or 4, Gleason scores of seven or greater, a pre-operative PSA greater than 10 or positive surgical margins. The intervention of relevant trials had to be a randomized comparison of adjuvant or early salvage radiotherapy, and all trials excluded men who had received prior radiotherapy or ADT before randomization.

The outcomes of interest differed for the three included trials, and in RADICALS-RT, this was time free from metastasis, and in GETAG-AFU 17 this was event-free survival, which is predominantly driven by PSA related events, and the RAVES it is the biochemical progression entirely dictated by PSA driven events. However, to allow for pooling across these trials, the authors created an aggregate primary outcome that they deemed "event-free survival," which was the time to the first of biochemical progression defined as PSAs equal or greater than 0.4 following radiotherapy, clinical/radiological progression, initiation of non-trial treatments that was typically androgen deprivation off the trial, death from prostate cancer or any PSA greater than two at any time.

Among the included trials, the authors assessed the risk of bias at the trial level using the Cochrane tool among randomized trials, and the prospective design of this meta-analysis allows for monitoring of the event rate in the control arm to determine when pooled analysis would be adequately powered, and this allows for the earliest meaningful data to be reported. The authors undertook a fixed-effects meta-analysis of hazard ratios and undertook a sensitivity analysis using random-effects models. These standard meta-analytic approaches to assess heterogeneity and plan subgroup analyses according to important potential effect modifiers, including preoperative PSA, Gleason score, seminal vesicle involvement, positive surgical margins, CAPRA-S score category, and the use of ADT.

And now I will pass it over to Dr. Klaassen to talk about the results and implications of this meta-analysis.

Zachary Klaassen: Thanks, Chris. So the authors did a literature search looking at 761 records that were identified. They excluded 754 of these, did seven full-text reviews, and then had three ongoing trials eligible for inclusion in the meta-analysis. And one potential eligible trial was identified in updated searches but was found ineligible because of incorrect treatment comparison, leading to three trials that were included in the meta-analysis. So this is a busy table, it is a full table looking at the trial characteristics of these three trials included. And I'll go through this step-wise so you can see. Over here, RADICALS, GETUG-AFU 17, RAVES, very similar accrual period, sort of the mid to late 2000s to 2015, 2016, slightly different key eligibility criteria in RADICALS, one or more positive margins, T3a, T3b, or T4, or a Gleason 7 to 10, and GETUG-AFU study, pT3a, 3b or T4 positive margins, but also, it included extracapsular extension. In the RAVES trial pT2, T3a, or T3b and either positive margins or extracapsular extension.

So little nuances to the exclusion criteria, certainly the use of hormone therapy varied between these trials. Most notably patients in the RADICALS-RT trial could choose to enter a second randomization, which was then randomized to either no hormones, hormones for six months, or hormones for 24 months, and participants not randomized could receive hormone therapy off protocol. There was no use of hormone therapy in the RAVES study, whilst in the GETUG-AFU study, all participants received hormone therapy alongside radiotherapy, both in the adjuvant or early setting. All three trials had radiotherapy to the prostate bed. The adjuvant radiotherapy timing was less than six months after radical prostatectomy. And you can see here that the radiation schedule was quite similar for RADICALS, 66 Gray in 33 fractions or 52.5 and 20 fractions. And in GETUG study, 66 Gray in 33 fractions and in RAVES, 64 Gray in 32 fractions.

The other important note is the timing of early salvage radiotherapy, which in RADICALS-RT was less than two months of a trigger PSA. GETUG study as soon as possible after PSA relapse and before the PSA reached one nanogram per milliliter. And in the RAVES study, less than four months after a trigger PSA. The primary outcome measure, as Chris mentioned, freedom from distant metastasis for RADICALS-RT, event-free survival for GETUG, and freedom for biochemical progression in RAVES. Interestingly, the RAVES study was a non-inferiority design, whereas the other two were superiority

Once again, another busy slide, but I'm going to walk through this quickly. Risk of bias assessment: you can see at the top, the risk of bias arising from the randomization process was low-risk for all three. The risk of bias due to deviations from the intended intervention was also low-risk for all three. The risk of bias due to missing data, once again, low-risk for all three of these trials. The risk of bias and measurement of the outcome, once again, low-risk for all three. And finally, the risk of bias and selection of reported result, low-risk for all three, leading the authors to an easy overall judgment, of low-risk for all three of these clinical trials.

In terms of the patient characteristics, you can see here, we have the RADICALS-RT on the left, GETUG study in the middle, and the RAVES study on the right. Very comparable in terms of patient age, most of these patients in the mid-sixties. Slightly larger study for the RADICALS-RT, almost 700 patients per arm, compared to smaller sample sizes in the other two trials. A median follow-up was more than five years in all of these studies. Median PSA preoperatively was 7.8 to eight in the RADICALS study, not available in the GETUG study, and 7.4 in the RAVES study. So the majority of these patients were T3a or T3b at about three quarters in RADICALS, almost all of them in GETUG and about three quarters in the RAVES study. The majority of these patients were Gleason seven about three quarters in all three studies. Seminal vesicles were not involved in about 80% in all three studies, extracapsular extension, part of the criteria for GETUG a 100%, and about 60% to 70%, for RADICALS-RT and in RAVES. Once again, you can see the CAPRA risk groups here, the majority of these being intermediate three to five, about half the patients in RADICALS-RT, not calculated in GETUG, and about 60% in the RAVES trial.

So this is the result of the effect of radiotherapy timing on event-free survival. You can see the three studies listed here, all of these 95% confidence intervals crossing the midline at one and not significant. You can see a hazard ratio of 1.10 and a 95% confidence interval of 0.81 to 1.49 at RADICALS because it had the most patients, it had the highest weight when calculating the meta-analysis hazard ratio. The overall hazard ratio, 0.95, with a 95% confidence interval, 0.75 to 1.21, so no difference between favoring adjuvant radiotherapy or early salvage radiotherapy.

The next two slides will show the results of the subgroups on the effect of radiotherapy timing on event-free survival. You can see here that they found no difference in the relative effect of adjuvant and early salvage radiotherapy on the basis of these subgroup characteristics listed on the left. On this first panel, you can see this included a pre-surgical PSA stratified less than or equal to 10. You can see Gleason score stratified by Gleason seven are greater than or equal to eight. And you can see once again, all of these hazard ratios crossing one and thus not contributing to the effect. Similarly, on the second panel, this included on the left seminal vesicle involvement, surgical margins, and CARPA risk group, all non-significant hazard ratios.

So several discussion points from this data presented today, and that adjuvant radiotherapy does not improve PSA driven event-free survival. And so the author's approach, which is why this paper is so significant and published in such a high-impact journal is this FAME approach, reducing the bias, several ways that you can see listed here. So the first metric for reducing bias is a selection of studies based on eligibility criteria, while the trials are ongoing or unreported. So this is not a retrospective look at trials that are published, but this is sort of happening coinciding with the trials. Secondly, harmonizing the event-free survival definition and planning all the analyses before the results were actually known. Once again, able to do this while the trials are going on, and finally working with the trialists to include up-to-date results from all of the patients that were randomized.

Additional benefits of this ARTISTIC collaboration is that this helps the trialist to reassure patients that are in the trial and the funders, that there is value in continuing trials. And so this is important. These are long trials, nearly 10 years. And so to continue to encourage them to keep going with the trial. It also provides opportunities in real-time to resolve these trial issues. Generally, prospective meta-analyses are typically... they use individual patient-level data, except that this may take many years to become available, and so this is a benefit of having the trials alongside the collaboration with the ARTISTIC folks.

Importantly, we are unable to explore the effect of giving hormone therapy alongside radiotherapy. In the GETUG study, as we have mentioned already, there was some concomitant radiotherapy and hormone therapy. In the RAVES, there is only radiotherapy alone. And in the RADICALS-RT trial, the patients were given optional second randomization too long, versus short, versus no hormone therapy. So this is one of maybe the limitations of this study.

So in conclusion, adjuvant radiotherapy following prostatectomy does not improve PSA-driven event-free survival compared to early salvage radiotherapy. Early salvage radiotherapy provides the opportunity to avoid or postpone radiotherapy and associated adverse events for many men with no effect on event-free survival, and this is really avoiding the overtreatment of men that would never benefit from early adjuvant radiotherapy. And finally, 88% of men remain event-free five years after radical prostatectomy. The likelihood of delaying radiotherapy, having deleterious effects on long-term outcomes is low. However, the authors have mentioned as the data matures, they will continue to conduct further analysis.

Thank you very much for your attention to attending this UroToday Journal Club.