Baseline PSA Level in Midlife and Aggressive Prostate Cancer in African American Men - Mark Preston
August 21, 2019
Mark A. Preston, MD, MPH Assistant Professor of Surgery Urology Brigham and Women's Hospital, Boston, Massachusetts
Alicia Morgans, MD, MPH, is an Associate Professor of Medicine in the Division of Hematology/Oncology at the Northwestern University Feinberg School of Medicine in Chicago, Illinois.
Alicia Morgans: Hi. My name is Alicia Morgans, and I'm a GU medical oncologist at Northwestern University. I'm excited to have here with me today a friend and colleague, Mark Preston, who's an Assistant Professor of Surgery at Harvard Medical School, as well as a urologic oncologist at Brigham and Women's Hospital and the Dana-Farber Cancer Institute, where he has the joint appointment there.
Thank you so much for joining me today, Mark.
Mark Preston: Thank you very much for having me.
Alicia Morgans: Of course. I really wanted to talk with you about a paper that you and the team have recently published, and I think that there are some really clinical pieces that are going to be, I think, critical for the listeners to think about. This paper was on baseline PSA levels in African American men thinking about whether those baseline PSAs could really correlate with aggressive prostate cancer.
And I'd love to hear you tell us about how that study was constructed, and what you've learned in that process.
Mark Preston: I would be happy to. This has been a process over years now. It initially actually started looking primarily at a baseline PSA, and in long-term cohorts. We initially did the study in the Physician’s Health Study. There was a cohort of thousands of primarily Caucasian men that had a very long follow-up. It was initially a randomized control trial that was started in 1982, and there was blood taken and stored.
And that transitioned into this really well-run prospective cohort, with fantastic follow-up of thousands of men, for upwards of 30 years, and with the ability for us to look at lethal prostate cancer as an outcome. We had previously looked at the ability of a baseline PSA during midlife to predict the lethal prostate cancer, with the goal being that we might be able to better risk stratify who or which men would have most to gain from prostate cancer screening while limiting screening in those who were less likely to benefit.
And what we found in that paper was that the midlife PSA predicted very well for lethal prostate cancer with 30 years of follow-up. For example, we found that for men aged 45 to 50, the median PSA was about 0.7 or so. If your PSA was above 1.7, you're actually in the 90th percentile for that age range, and your subsequent risk of lethal prostate cancer was over 10 times greater than if your PSA was below the median.
What else we found was that while that was very predictive, and would allow us to risk stratify future screening, this was primarily in Caucasian men, and we didn't have adequate data for African American men. This is, as you know, a common issue in I'd say almost the majority of randomized trials.
From the screening perspective, and also from treatment, as well, and so what we wanted to do was to see if that similar finding would hold true among African American men.
What we did, there's a cohort-based out of the southern United States called the Southern Community Cohort Study, and so we used that to replicate a similar study design. It was a nested case-control study with 197 prostate cancer cases, matched with just over 500 controls, and with the similar study design we looked at, the baseline PSA was amongst men who were controls and didn't develop prostate cancer, and looked at subsequent aggressive prostate cancer over the following median 10 years of follow-up.
And two of the key findings from this study was that, interestingly, the median PSA values between Caucasian and African American men were actually very similar. If you were from 45 to 50, it was around 0.7, and if you were around 60, it was around 1, which was an interesting finding. If you were to use risk stratified PSA values, the numbers would actually be similar between African American and Caucasian men.
In addition, what we then looked to see was how well that baseline PSA predicted, and we also found that the data actually worked very well. For example, when we compared the odds ratios for aggressive prostate cancer among men who had a PSA above the 90th percentile versus those who had a PSA below the median, the odds ratio was 174 for prediction of aggressive prostate cancer.
And it was so high, basically, because all men had a PSA above the median, and so the prediction was actually very good, and we found that held true when we limited it to men with a PSA below 10 or below 4, and even when we limited it to either fatal or metastatic prostate cancer, it remained as predictive.
Overall, we found that this PSA level during midlife strongly predicted for the future development of aggressive prostate cancer among black men and that targeted screening based on this midlife PSA might identify men at high risk while minimizing screening with those at low risk.
Alicia Morgans: Thank you so much for walking us through that, and I think sometimes these nested case-control studies can be complex in how they're constructed, and how they maintain their statistical rigor given the kind of non-randomized type of a design, but just to break it down for everyone listening, if you are seeing an African American guy in your clinic, what are the numbers that you're thinking about?
In terms of, is there a cutoff or is there a range of numbers where you would say, "This PSA and this man, of this age, is something that's going to make me think twice and maybe do something additional in the workup so that I don't find this patient with an aggressive prostate cancer in a few years."
Is there a PSA that makes you think that, or is this more really just a trend that tells you that these numbers exist in a continuum, and we should just be thinking about them more generally?
Mark Preston: I would say there actually is. I think a lot of the trouble with PSA in terms of overdiagnosis and overtreatment does happen when your PSA values get higher during your 60s or mid-60s when you also have a lot of benign prostate growth or other reasons to be causing a PSA elevation.
A lot of this comes back to the autopsy studies that have been conducted, men who died of other causes, that prostate cancer typically starts in your 30s, 40s, early 50s, but may take decades to show up and present.
And I think that is the reason why the PSA values at that level are actually predictive because there are early stages of prostate cancer present, but without the confounding that you would get from the benign prostate growth at that point. For example, if this was an African American man who was coming to see me and he was 49 years old, and his PSA was 2, for example, he would actually be above the 90th percentile for his age group for that age.
And he would be at significantly increased risk of aggressive prostate cancer. It may not be over the next year, it may not be for years later, but that would be a man that I would consider yearly PSA screening, and if that were to elevate further, do investigations with MRI or biopsy.
Alicia Morgans: Great. Yeah, and thank you for breaking that down. I think ... and perhaps I'm looking at the paper wrong, but I think in this paper, this was published in a really elegantly done paper in European Urology. You were the first author. It was published actually just earlier this year, and we could try to post a citation next to this recording.
In table three in the paper, I think that it gives a nice sort of illustration of patient age, and then the quartiles, and then the median, and all of that, for PSA, and it seems to give some help for people to try to think through an age-adjusted PSA and think about what percentage of these patients may have aggressive cancer.
Like you said, if you're looking at the top 10th percentile for patients who are in the 40 to 49 age group, with a PSA greater than 1.68, you could see that a lot of these patients are going to have an aggressive cancer, so it says actually 100 percent of patients with that high PSA will develop, potentially, an aggressive cancer.
Now, remember; this is not a prospective study, and there's probably few people in the world that we can say, "You have a 100 percent chance of getting aggressive prostate cancer." But this is just helping you think through, I think, as a clinician, when age is this low and your PSA is this high, the risk is higher and you do need to think about extra steps.
Is that table three? Am I interpreting that properly, Mark?
Mark Preston: You are. Yeah. The gist of it is that current guidance would say start PSA screening at age 55, for example. I think there's some value to risk stratifying, and the guidelines even recognize that. They say if you've got a family history, or if you're African American, for example, you could consider starting earlier. There's been some good papers, actually, where they've looked at the value of the baseline PSA, and it seems to actually be a stronger predictor of future aggressive prostate cancer than even family history or races.
What I would envision is a screening strategy, for example, where you would start it at age 45, for example, with a baseline level, where the values should be low [inaudible] screening every year. If your PSA's below the median, you might not need another PSA for five or 10 years, for example, and interestingly enough, this is all dealing with aggressive prostate cancer with a median of nine years of follow-up, but if you look at the prior paper, that was 30 years of follow-up looking at lethal prostate cancer, and the findings actually were very similar for white men.
There's also been a lot of work done in Sweden with Ponto Lili and Andrew Victors who've done studies looking at the Malmo Cohort, which is a similar long term prospective cohort, and actually, the PSA values were quite similar in that group, as well. If you look at the final table, table five, they do look at some of the cohorts where this has been investigated as well.
I think there's a good evidence base, based on existing prospective studies, like these have been prospective cohorts that have been immaculately followed for decades, with these method case-control trials conducted within them.
Alicia Morgans: Absolutely. I think the other point I would make that I think is really unique about this particular trial is that this is one of the first, if not the first, and maybe the only prostate cancer paper that I've seen come out of the Southern Community Cohort Study. I could be wrong; there might be another one that's out there, but you know, I was at Vanderbilt for a number of years and that's an institution that's highly invested in studying and working with the Southern Community Cohort Study, and it took a lot of effort, I think.
I'm sure it did, to get that dataset to have sufficient follow-up to do something like this. The reason that that is such an important cohort, I think, is because it's a community-based cohort based in the south, the southern United States, that is really focused on men and women, but the men are the ones developing prostate cancer, of course, but in a prostate cancer study men who are of potentially lower socioeconomic status who may not have access to academic centers.
And it's really a community-based study where you have access to this population that you would not necessarily otherwise have access to, so I think having results that look similar to these meticulously documented cohort studies from other countries where populations are predominately Caucasian and are not in this situation where there could be access issues for these patients, is, I think, also really striking, and really importantly demonstrates that it's the biology, probably, that's driving this, and these are probably real numbers and real things that we're seeing in these studies.
Though, they're not prospective studies based on endpoints of developing prostate cancer, and monitoring men by PSA, which is potentially where we'd need to go to really solidify a screening procedure that has differential action based on age and PSA level as we're talking about potentially moving to at some point in the future.
Mark Preston: Right. I think that's a nail to challenge, there, is it's really difficult to launch [inaudible] population level PSA median trials, so [inaudible] and the PLCO trial in the USA and those were massive trials, and that's sort of the timeframe that you need to determine, "Okay, who's getting aggressive prostate cancer? Who's actually developing lethal prostate cancer?"
And so, I definitely think that there's value in looking at these long term, prospective cohorts that are well conducted, and then there's certain aspects of modeling studies, as well, that could devise reasonable strategies for smarter PSA screening, as opposed to just disregarding it and saying there's no value. I think there is value to it. It just needs to be done in a smarter way, and I think this is one technique that does provide value.
Alicia Morgans: Mark, the SCCS, that's Southern Community Cohort Study, is really, really an interesting study. It's a fascinating population of people who are treated in community clinics, and the data has been, I think, collected through a standard of care community clinical care, and it's a population that is sometimes treated in ... these are non-academic centers, and so you don't necessarily know if you can trust all the data, or even if a PSA was routinely gathered on these patients.
How were you able to get the PSA data that you needed from this population to actually do the work in this study?
Mark Preston: Yeah. An interesting aspect of this study design was that we selected 1,000 men who were initially, we decided that we were going to study, so these were patients who were enrolled back starting in 2002, and all these patients, when they were enrolled, they did not have cancer at that point in time, but what they did do is that they gave blood samples.
Blood was taken and stored, and when we decided to conduct the study, we were able to go back, select those who had cancer, select all the control populations, and then we went back through their initial blood store to take the blood and measure fresh, basically, in a standardized fashion what the PSA levels were prior to anybody having a diagnosis of cancer, so that it wasn't impacted by screening strategies or anything like this.
This was just what their blood was pre-diagnosis, and that was what served as the baseline measurement, so we took all the control population, the 570 patients who never developed prostate cancer over the course of the study, and the median values were based on that population.
If you didn't develop prostate cancer over the course of the study, the median PSA was 0.7, for example, and I think that's a unique aspect of what we were able to do with this study design, and it's similar to what we'd done in the prior JCO paper, as well.
Alicia Morgans: That's wonderful because it's really such a powerful way to standardize your PSA measurements and really ensure that your data's going to be believable on the backend, so excellent strategy and design, and thank you so much for sharing your work with us.
Thank you for doing the work in the first place, and I'd love to hear your final thoughts, a message to the listeners, on your work on baseline PSA and aggressive prostate cancer in African American men.
Mark Preston: I think what this comes down to is using PSA screening in a smarter fashion. I think starting in the mid 40s has some value to establish what that person's inherent risk is, looking at family history, looking at their baseline PSA, and just being aware of that as an additional data point as you do your shared decision making, and deciding on, "Should we screen or not?" And with the goals of focusing more on the people who are more likely to benefit from screening by avoiding screening in those where overdiagnosis could be avoided.
Alicia Morgans: Absolutely. Well, thank you for getting us one step closer to the smart screening strategies that we are all looking for. Thank you for doing this work.
Mark Preston: Thank you for having me.