Read the Full Video Transcript

Jason Efstathiou: All right, disclosures. All right, let's start with a case. So a 55-year-old man with T3b node-positive Gleason 9 disease has a 7% chance of remaining disease-free at five years post-op according to the Sloan Kettering nomogram. Should this patient be getting adjuvant radiation?

If we turn to the NCCN guidelines, it tells us that patients with T3 positive margins, in general, should be monitored. That is the preferred approach versus treating. If they have lymph nodes, they could be either treated or monitored. And when being monitored, it means early salvage.

So, we know that there's been an increase in surgery for high-risk diseases. It's been discussed nicely. This has led to increased indications for post-op radiation and its use. And when we're talking about adjuvant radiation, how do we define that? We define that as no evidence of disease after surgery. It really means an undetectable PSA. And you're considering radiation for significant risk of future recurrence based on pathology to improve local control and avoid dissemination of disease.

So what is an undetectable PSA? I think in general, it's defined as a PSA less than 0.1 in most contemporary studies. The NCCN says that if there's two or more increases or a PSA over 0.1, that becomes detectable. AUA uses the threshold of 0.2. RADICALS says two consecutive rises with a PSA of 0.1 or higher or three consecutive rises at any level. And then there are studies out there that say the optimal PSA really varies by underlying clinical pathologic risk of recurrence. So for high-risk of recurrence, maybe a single PSA of 0.05 or higher should be the trigger.

And in the world of adjuvant radiation, you've heard, there's lots of trials that have shown benefit over observation for adverse path features. Yet the use, as Alberto showed, of adjuvant radiation these days remains very low, under 10%. And yeah, maybe that's validated by studies like RADICALS, RAVES, the GETUG study.

We all know that when comparing adjuvant to early salvage, there was no difference in progression-free survival or freedom from non-protocol hormone therapy. And the authors said, "Well, therefore you should pursue a program of salvage." But I don't know if you've all seen the curves on the right, which is out in pre-proof mode in their long-term update from the RADICALS trial. And it's really interesting. The primary endpoint was freedom from distant mets. And you can see the values there. It's trending towards significance. Their original primary endpoint was prostate cancer-specific mortality. And look at that. That's significant now in their long-term update. Does that mean we should be treating some patients with adjuvant? And this is a favorable cohort in RADICALS. If you look at their inclusion, they allow just single risk factor like Gleason 7 disease or a PSA just over 10.

But the conclusion was: pursue early salvage. If doing early salvage, we know the lower the PSA, the better when you use radiation. Some have suggested threshold levels of 0.25 or lower as the threshold. But really, let's come back to this question. Shouldn't we be doing adjuvant? What about that long-term update from RADICALS? And RADICALS, in these other studies, were limited by patient inclusion. Less than 20% were Gleason 8 to 10, only 5% were node-positive. And if there's going to be any benefit to adjuvant, it's going to be limited to a subset of high-risk patients. So if you look at this study by Tilki et al., in a large cohort in patients with node-positive or Gleason 8 to 10, T3, T4 disease, there was benefit, there was benefit to adjuvant over early salvage. And if you excluded the node-positive, there was even more benefit. So maybe it should be considered in patients with very high-risk features.

And what about node-positive? Many retrospective studies, Briganti and others, have looked at this and suggest benefit to adding radiation to hormone therapy in node-positive disease. And the Tilki study says that there's increasing benefit the more nodes that were found at surgery and especially if there's four or more nodes.

So can genomic classifiers, as we heard nicely from Dan, help inform this practice? Yes. In the post-prostatectomy setting, there's genomic classifiers like Decipher. In general, if you look at guidelines, they say that you should only use them. As Dan said, if they have the potential to change management, they should not be ordered reflexively. But if you look here at guidelines, they'll say for T3 node-positive PSA undetectable, Decipher should be used to consider adjuvant treatment. And what is that based on? It's based on studies like this. If the Decipher score is 0.4 or higher, maybe there is a benefit to adjuvant over salvage, but not if the Decipher score is low.

I actually really like these risk stratification models that say if you have two or more of these risk factors, T3b, T4, Gleason 8 to 10, node-positive, Decipher score greater than 0.6, there is a benefit to doing adjuvant over early salvage.

Now, if you're doing adjuvant in some of these patients, should you add hormone therapy? We had a trial in the EORTC, but it failed to accrue, unfortunately. So we're extrapolating data from trials in the salvage setting that have shown benefit to adding ADT to radiation. There's an ongoing debate of whether there should be a PSA threshold, though, for the use of that ADT.

If we look specifically at the NRG/RTOG 0534 study, it has shown incremental benefit to adding: A, hormone therapy to prostate bed radiation, and then pelvic nodal radiation to prostate bed radiation. And so if you do it all—hormone therapy and pelvic nodal radiation—it's far better in all subgroups compared to prostate bed radiation alone, including in PSA levels of less than 0.35.

What about the duration, then, if we're using hormone? Well again, we're not so informed in the adjuvant setting, but if we draw from the salvage setting and the RADICALS HD trial, which has been presented at ESMO, not yet published, it was a three-arm randomization to no hormones, short course, long course. And they allow two-way and three-way randomizations. They ended up with two groups: no hormones versus short course. Short course versus long course in the latter, there were higher-risk patients, more of the T3 and Gleason 8 to 10 patients in that short versus long comparison, but still not that very well represented. It's still a minority of patients that have those kinds of higher-risk factors. But in the short versus long, there was a benefit to longer duration over short duration in the RADICALS HD trial in terms of metastasis-free survival.

So what about intensifying ADT? We just heard about, I don't know, is it called ARTA? RP? RC? I don't know. But one of those is the NRG-GU008 trial in node-positive patients. They can have any PSA above zero. That means 0.01 qualifies you. They're randomized to radiation hormone therapy plus minus Apalutamide on study.

So can we use genomic classifiers to inform the use of ADT in this post-op RT setting? Well, yes. In the salvage setting, Decipher has been looked at in RTOG-9601, and in the subset of patients that got early salvage radiation, if you had an intermediate or high-risk Decipher score, there was benefit, but not if you had a low-risk Decipher score.

And so what do the guidelines tell us? They tell us that Decipher can inform the use of ADT with early salvage radiation, though the optimal duration is unknown. And what about AI? We heard about this again from Dan. ArteraAI digital histopathology-based machine learning tool.

What about that in terms of informing ADT use with salvage radiation? This is kind of, not even hot off the press, it's going to be presented at AUA next week. So please don't put this on Twitter or anything. I got permission to show this here. But in a multimodal AI tool, high-risk group, there is benefit to adding ADT to salvage radiation. But if that AI score is low-risk, there is no benefit. Very similar to what we've seen with the Decipher score.

So in conclusion. And Silke and Aurelius said, what do I do? What do I do in practice? I favor early salvage radiation over adjuvant radiation in most patients, but I use a definition of two to three consecutive rises at any level or increases to 0.1 or higher, to trigger early salvage. I consider adjuvant radiation. I do consider adjuvant radiation in otherwise fit, very motivated, very high-risk patients with two or more of the following risk factors, T3b, T4, Gleason 8 to 10, node-positive, Decipher score greater than 0.6.

These patients were underrepresented in RADICALS. That's where you're going to see the benefit. And guess what? That update in RADICALS maybe shows benefit in a more favorable cohort to adjuvant even.

So in high-risk patients, I use a lower threshold to initiate what I call ultra-early salvage or adjuvant plus radiation. You can consider monitoring with an ultra-sensitive PSA and consider the trigger being 0.05 or higher for those high-risk patients. And if you're giving adjuvant radiation, it implies it's high-risk disease. So I do treat the prostate bed and pelvic nodes. I recommend short course hormone therapy generally, maybe long-term depending on the risk factors. I'm extrapolating from studies like 0534, and I may consider genomic classifiers or AI tools to help with that informed decision-making.

Ultimately, our goal is to avoid or delay radiation and its potential toxicity in those whom we can, but don't miss a window of opportunity to cure patients who are guaranteed to recur. Thank you very much.