Christopher Evans: Thank you also Silke and Aurelius for the invitational. I’ve not made it down the Rhine swim yet. I’m hoping to be able to accomplish that. I was asked to talk about PSA recurrence or persistence after prostatectomy. These are my disclosures and I want to thank Darya Toki who was a fellow with me. She’s done a lot of the studies I’m going to show you. I’m going to try to avoid repeating some of the data that was shown in the excellent previous session, so I’ve moved some slides around to try to not duplicate that too much. And Declan Murphy shared some of his PSMA slides with me, as well, so thanks to both of them.
So I’m asked to talk about the definitions of persistence or recurrence and the clinical implications, imaging and then management, and I’m going to narrow this down to a little bit about salvage radiotherapy and the node zero patient or negative patient in salvage lymphadenectomy. It’s interesting how the definitions vary across the different guidelines, and this is very relevant to when you get to vote on this question. It’s important because it’s changing. The AUA has a definitive cut point of remission is discussed as a nadir PSA of less than 0.2 nanogram per milliliter. The NCCN does not have an absolute number, but says that persistence is when the PSA does not fall to undetectable levels, and recurrence is when an undetectable PSA after radical prostatectomy rises with two or more determinations, but again does not list an absolute number.
Let’s first talk about PSA persistence. Between 5 and 20% of men will have a detectable PSA after surgery and it’s most studies defined as 0.1 or greater. Now, it could result from persistent local cancer, preexisting metastasis prior to surgery, or residual benign disease, and that usually occurs at the bladder neck, where a large prostate’s pushing in. In my personal experience, a detectable PSA above 0.1 is rarely benign disease unless a gross amount of specimen has been left behind, which is uncommon. I think the first steps are to validate that PSA and then check it at two to three-month intervals to determine what it’s doing to get a read on the etiology.
This is a paper that looked at persistent PSA in radical prostatectomy for the Martini Clinic, over 11,000 patients of which over a thousand had a persistent PSA above 0.1 after surgery. And what’s quite clear is that if you look at undetectable PSA in orange versus persistent PSA in blue, you can see that the survival probability is worse in the persistent PSA from both metastasis-free survival and cancer-specific survival. If you look at multi-variable analysis of the clinical and pathologic features which drive this outcome, it’s quite clear that higher grade, higher stage positive margins in lymph node-positive disease make for a higher likelihood of persistent disease.
What about salvage radiotherapy in this study in these group of patients? Well, the addition of salvage radiotherapy versus no salvage radiotherapy on cancer-specific survival, 10 years for this entire mixed cohort was significant, 81.6 versus 93.7%. Clearly, the addition of salvage radio therapy as a group helps. And if you look at the pathologic lymph node positive patients, again, the cancer-specific survival at 10 years with salvage radiotherapy was 96.2, much higher than the 55.8% in those who didn’t get it. And again, the addition of ADT to that salvage therapy was variable across this population, but most did get it.
In this paper from Fossati and Briganti, they looked at the impact of early salvage radiotherapy in patients who both had persistent and recurrent PSAs, and this was a multicenter collaborative group, and they identified five subgroups of patients who might benefit, but only three of them truly did benefit. Now, they broke those down into low risk … These are the three groups that would benefit … low risk defined as a patient who had an undetectable PSA after surgery, Gleason score seven less, and they had tumor stage PT3B. Now Alberto, I would argue that low risk in my book would not be a pathologic T3B, but in your definition it was. But I would argue maybe that’s not. They benefited as did the intermediate risk undetectable PSA after surgery, but high Gleason score, and high-risk meaning of PSA persistence after radical prostatectomy, but a lower Gleason score. The two groups who didn’t benefit were the very low risk, defined below is undetectable PSA low Gleason, and low tumor stage or the very high-risk patients.
Now, let’s shift to biochemical recurrence. The patient initially is undetectable and rises. It occurs in up to about 35% after surgery and the natural history is variable. Originally, this definition comes from a couple studies around the year 2000, 2001. This one is from Chris Amling and a group at the Mayo Clinic. And you can see if you look at progression-free probability at 10 years, greater than 0.4 nanograms per milliliter had a 60% likelihood of being progression-free. The AUA took a slightly more conservative cut point of 0.2, but these are the studies upon which this PSA recurrence definition came into being that many of these patients don’t recur.
Charlie Pound’s paper from the Hopkins Series back in 1999 looked at almost 2000 patients who underwent surgery, of which 315 had a PSA recurrence, removing the 11 patients who did get adjuvant or salvage rate ADT. And you can see that the five-year metastatic progression-free survival was pretty good at 64%, but again, this cohort was mixed but heavily low in intermediate-risk patients. If you look more contemporarily here at 14,500 patients done by Tilki, you can see if you stratify them by risk into low risk, at the top, red is intermediate and green is high-risk by the Capra stratification, that the biochemical recurrence-free survival is clearly much worse in the high-risk cohort. But overall, it’s about 20%.
What about the impact on the time of the surgery to the recurrence? Well, clearly patients who recur with their PSA less than 12 months from the time of surgery do worse than patients who recur greater than 12 months. And this is just a straightforward reflection of the likelihood of the stage being metastatic and those with a faster doubling time and earlier recurrence.
This is a systematic review put together by European Urology, and it looks at prognostic value of biochemical recurrence for prostate cancer. About 30% will have a biochemical recurrence after surgery, but only 16% will die from the disease. This is the take-home slide for the voting because when you look at the definition of recurrence of the AUA, the NCCN and the EAU, the EAU is now changing the definition of recurrence based in two groups. Low-risk biochemical recurrence is defined as a PSA doubling time of greater than one year and a Gleason score less than eight. The high-risk group for biochemical recurrence is defined as a PSA doubling time less than or equal to one year or a high Gleason score 8 through 10.
Notice there is no PSA value, absolute cutoff value. It’s included in this and I think this makes more sense, but has it been validated? The group at the Martini Clinic looked at a thousand patients, 500 low-risk, 500 high-risk. The orange is the low risk, the blue is the high risk. You can see the low-risk patients clearly do better, supporting the definition there. And you can see if you apply giving salvage radiotherapy early, which is the ESRT, late, LSRT, or not at all, on the left are the high-risk patients, and you can see here that those who got no salvage radiotherapy did much worse than those who got it late and in the high-risk biochemical recurrence group at EAU, those who got it early did the best. But if you look at the low-risk cohort, really not a lot of difference as to whether they got salvage radiotherapy with regard to metastatic progression or cancer-specific mortality over time.
What about ultra-sensitive PSA? This is used quite variably in the US. We did this systematic review on it, and you can see the algorithm we came up with looks a bit like a neural network. However, I would point out there is some really insightful data and several people in the room have done studies on ultra-sensitive PSA. But ultimately, an ultra-sensitive PSA that goes above 0.05 has a high positive predictive value of 92% to achieve the AUA definition of 0.2 or higher. Also, ultra-sensitive PSAs above 0.03 and rising at beyond six months, there’s a higher likelihood of biochemical recurrence and poor overall survival, but most importantly, this is not included in guideline recommendation for testing, and in large part because a single ultra-sensitive PSA may be a false positive. It should not be used to make any disease treatment determinations.
Imaging– Well, this has been discussed pretty heavily. I would point out again that we have the Axumin scan in the US, and there are specific recommendations for when you should use it. PSMA, not yet approved in the US. That’s why it should be used in clinical trials. I don’t want to repeat a lot of the petting the PSMA scan. However, to avoid the cowbell, I’m going to leave most of this data out because the PSA cowbell might catch me. It’s been discussed.
But I want to just point out this again. This was shown earlier by Declan. This is a nice systematic review which clearly shows the nice sensitivity and specificity of the PSMA PET scan. But relative to this talk, I would ask this question. At a 42% positive scan detection rate, you are in the range where you want to be giving salvage radiation therapy, and are you going to forgo it if you have a negative PSMA PET scan on a patient who has a rising PSA because of that or are you going to say the natural history of the recurrent PSA which is rising would lead you to give adjuvant radiation or are you going to wait for the PET scan to more likely become positive in a realm where we know that late salvage radiation therapy isn’t as effective?
Personally, I’ve cautiously used the PSMA PET scan with less enthusiasm than the previous group because I don’t think that it’s going to drive my clinical decision on whether to use salvage radiotherapy with ADT or not in many of these patients, but that’s why I think the guidelines say it’s weak evidence, weak strength to use it and only if it’ll influence your treatment decision, which is critical.
If you give radiotherapy, there are two forms. You can add by bicalutamide for two years with radiotherapy and that benefits patients who particularly have a PSA above 0.07 or specially above 1.5 at the time of treatment, which was done in the SWOG trial. I do prefer to treat them more likely according to the GETUG-16 trial, which added six months of ADT with radiotherapy in patients with a PSA of 0.2 to 2.0, although when you get up into this higher range, clearly now imaging may be a benefit.
Salvage lymphadenectomy. There’s very poor evidence. The two and five-year biochemical recurrence-free survival, as you can see, two year 23 to 64, 6 to 31. These are all limited by the quality of these studies. But if you look at the best study down here by Fossati, 44% have undetectable PSA at 30 months. Optimal candidates, lower Gleason score, longer time to recurrence, no retroperitoneal nodes uptakes, fewer hotspots, and lower PSA.
In conclusion to this talk, I would say PSA persistence of greater than 0.1 is a poor prognosis, but salvage radiotherapy is beneficial. The guidelines are changing. I like the EAU high-risk guideline without an absolute cutoff. Ultra-sensitive PSA has value, but it’s not in the guidelines. You have to use it cautiously. Again, EAU low-risk has low-risk of progression and may not need salvage therapy. Salvage radiotherapy is important in the high-risk group, and surgical approaches, in my view, are immature and should be considered cautiously. Thank you.