The Peter Attia Drive - #388 — Prostate cancer screening： why current PSA guidelines are failing men and how modern tools improve early detection and save lives

00:00:00.000 Hey, everyone. Welcome to The Drive Podcast. I'm your host, Peter Atiyah. This podcast,

00:00:16.540 my website, and my weekly newsletter all focus on the goal of translating the science of longevity

00:00:21.520 into something accessible for everyone. Our goal is to provide the best content in health and

00:00:26.720 wellness. And we've established a great team of analysts to make this happen. It is extremely

00:00:31.660 important to me to provide all of this content without relying on paid ads to do this. Our work

00:00:36.960 is made entirely possible by our members. And in return, we offer exclusive member only content

00:00:42.720 and benefits above and beyond what is available for free. If you want to take your knowledge of

00:00:47.940 this space to the next level, it's our goal to ensure members get back much more than the price

00:00:53.200 of the subscription. If you want to learn more about the benefits of our premium membership,

00:00:58.040 head over to peteratiamd.com forward slash subscribe.

00:01:04.260 Welcome to a special episode of The Drive. In this episode, I'm going to walk through

00:01:09.680 a single topic in depth, and this is prostate cancer screening, something that I'm very

00:01:17.560 passionate about. And because we consider this a really important PSA, no pun intended, the full

00:01:24.160 episode and the detailed show notes for this discussion will be available to everyone,

00:01:29.200 regardless of whether or not you're a premium subscriber. In this episode, I discuss why

00:01:34.340 advanced and metastatic prostate cancer diagnoses are rising despite the availability of screening

00:01:40.400 tools, what PSA actually measures and why it's more useful when interpreted over time

00:01:46.380 than it is in a single one-off number, how MRI, PSA density, PSA velocity, and improved

00:01:53.600 biopsy techniques can reduce unnecessary procedures while improving the detection

00:01:58.400 of aggressive cancers, how active surveillance helps avoid overtreatment in men with low-risk

00:02:03.580 cancers, why the evidence used to argue against PSA screening was deeply flawed,

00:02:09.680 how drugs like finasteride, which are commonly used to prevent hair loss, can suppress PSA and

00:02:16.600 lead to missed warning signs if clinicians and patients are not vigilant, and why I think regular

00:02:22.120 PSA testing is so important if one of our objectives in life would be to eliminate prostate

00:02:29.460 cancer mortality. So without further delay, I hope you enjoy this special episode of The Drive.

00:02:39.680 Well, today I want to dig into something that I feel quite strongly about and something that I

00:02:46.400 think we're getting wrong at the population level in a way that is measurably costing lives.

00:02:52.300 We're going to talk about prostate cancer screening, specifically PSA testing, and why

00:02:58.240 I believe the current guidelines have failed men and obviously what I think a more modern

00:03:04.120 approach to screening should look like in 2026 and beyond. And this is really a discussion that

00:03:10.340 I think is relevant for anybody listening to this as a potential patient, someone who cares about a

00:03:16.100 patient, or whether you're a clinician. I'm going to spend some time towards the end of this

00:03:21.020 discussion on something that I also think is dangerously underappreciated, which is the

00:03:26.560 interaction between a drug called finasteride. This is a drug that literally millions of men

00:03:32.640 are taking for hair loss prevention and how that drug interacts with our ability to interpret PSA

00:03:38.980 correctly. Because if you're on finasteride and your doctor doesn't understand the implications

00:03:44.360 of that and how that impacts your PSA values, you could be sitting on a ticking time bomb without

00:03:50.260 anybody knowing it. Let me set the stage. Prostate cancer is the second leading cause of cancer

00:03:56.760 death in men, with only lung cancer claiming more lives. About one in eight men will receive

00:04:03.800 a prostate cancer diagnosis at some point in their lives, and this year alone, roughly 36,000

00:04:10.440 American men will die from it. So what makes this disease both uniquely hopeful and uniquely

00:04:17.880 frustrating is that when you catch it early, the prognosis is outstanding. So we're talking about

00:04:25.060 a 15-year survival rate approaching 97% from first diagnosis. And we have detection tools

00:04:33.900 that are simple, widely available, and getting more sophisticated every year. And yet here we

00:04:39.160 are watching the rates of advanced stage prostate cancer climb. That's right. So the rates at which

00:04:47.000 advanced prostate cancers are showing up is higher today than it was 15 years ago. Recent data out

00:04:53.200 of both the United States and Canada confirm what many urologists have been warning about for years.

00:04:58.720 We're catching more and more of these cancers only after they have already metastasized or spread.

00:05:06.260 And that's what we call stage four disease. And once you're at stage four, the picture

00:05:10.100 changes dramatically. Five-year survival falls to 38%, and the median time from diagnosis to death

00:05:18.380 is about two and a half years. So the question that I think demands an answer is quite

00:05:24.060 straightforward. How did we get here? We have the screening tools. We have effective therapies for

00:05:30.120 early cancers. Why are outcomes getting worse instead of better? And of course, most critically,

00:05:36.260 then what do we need to do about this? Now, to answer any of that, you need to start with the

00:05:40.260 basics of what the PSA actually is, because it's the linchpin of this entire conversation. And

00:05:46.620 quite frankly, if we had the PSA equivalent for every other cancer, the world could look a heck

00:05:52.720 of a lot different than it does today with respect to cancer mortality. So PSA, which stands for

00:05:57.460 prostate-specific antigen, has been available as a blood test since the early 90s. It's a protein

00:06:04.120 made by the prostate gland. Its biological function is to help keep seminal fluid in a liquid state.

00:06:11.480 So a tiny amount of PSA naturally leaks from the prostate into the bloodstream,

00:06:16.840 and that's what we measure when we draw a person's blood.

00:06:20.780 So here's the kind of important physiologic point.

00:06:23.400 As the prostate grows, which happens in virtually all men as they age,

00:06:28.300 the amount of PSA in the blood tends to rise proportionally.

00:06:32.660 So when the PSA test was first deployed, the logic was simple.

00:06:36.260 if your PSA is unusually high, your prostate is unusually large, and that could mean trouble.

00:06:43.340 An elevated PSA would prompt an ultrasound, an abnormal ultrasound would trigger a biopsy,

00:06:49.120 and cancer on a biopsy would lead to treatment. And the early data on this approach were generally

00:06:55.080 reasonable. Two large trials, followed over 14 and 16 years respectively, showed that routine PSA

00:07:03.080 screening cut prostate cancer mortality by 44 and 64% respectively. That translates to preventing

00:07:13.120 up to 3.1 cases of metastatic, which means fatal, disease for every 1,000 men who went through the

00:07:22.120 screening process. And the test itself costs nothing, a standard blood draw fully covered by

00:07:28.120 insurance plans. So you would naturally ask, why wouldn't we screen everyone? Well, here's where

00:07:35.060 the story takes a turn that I think had a devastating set of consequences. Even though

00:07:40.640 the people who made the decision were operating at least somewhat with reasonable logic at the time.

00:07:47.020 The problem was that the straight line from elevated PSA to biopsy was generating too many

00:07:54.940 false alarms. PSA can be elevated for many reasons that have nothing to do with cancer,

00:08:00.920 including benign enlargement, transient inflammation, such as you might experience

00:08:07.040 during a bout of prostatitis, recent sexual activity, and frankly, even just natural

00:08:12.300 variation. But the biopsies that followed in elevated reading were hardly benign. The standard

00:08:18.780 approach involved punching a needle through the rectal wall to sample prostatic tissue.

00:08:27.320 So now you're creating a contaminated pathway from a bacteria-rich environment into what should be a

00:08:33.220 sterile organ. And even with things like prophylactic antibiotics and preparation,

00:08:38.320 infection rates from these biopsies ran between 5% and 7%, which of course, when you multiply that

00:08:43.860 across the number of people that were having them, is an enormous amount of morbidity.

00:08:48.340 And prostate infections, of course, are not a joke, especially if they reach into the bloodstream.

00:08:53.260 They would frequently land men in the emergency room.

00:08:56.460 They could lead to hospitalizations and take people away from work.

00:08:59.380 So on top of that, there was the overtreatment issue.

00:09:02.620 So back in the 90s and early 2000s, finding cancer on a biopsy almost invariably led to treatment.

00:09:10.600 And that treatment, surgical, carried real consequences.

00:09:14.220 erectile dysfunction, urinary incontinence, bowel problems, especially if radiation was used.

00:09:20.360 The psychological weight of a cancer diagnosis alone is enormous. And what we've come to

00:09:25.740 understand since then is that a significant fraction of the cancers caught during the early

00:09:31.260 PSA era were actually quite indolent. They were never going to progress to the point of threatening

00:09:37.080 someone's life. These men were being subjected to the full burden of cancer treatment for a disease

00:09:43.520 that, if left alone, would never have harmed them. There's a saying in the urology world that says

00:09:50.920 every man will get prostate cancer, but some will die of it. And the idea here is if you live long

00:09:57.760 enough, you're always going to find prostate cancer inside some cells of the prostate. That's

00:10:04.220 not the jugular question. The jugular question when you identify prostate cancer is how many

00:10:10.120 of these cases are cases where, if left untreated, the cancer will spread from this organ to a

00:10:17.820 distant organ, namely the bone, which is the most commonplace of spread. So in response to all of

00:10:23.960 this, the United States Preventative Service Task Force, or USPSTF, made its move. And in 2008,

00:10:31.060 they stopped recommending PSA screening for men over 70. By 2012, they extended that recommendation

00:10:38.160 against screening to men of all ages. The data they leaned most heavily on came from a large

00:10:45.740 American trial called the Prostate Lung Colorectal and Ovarian Cancer Screening Trial, or PLCO.

00:10:54.260 The prostate arm of this study randomized men to either undergo PSA screening or not,

00:10:59.220 and it found essentially no difference in prostate cancer mortality between the two groups.

00:11:05.200 The implication was that screening saved, at best, 1.3 lives per thousand men tested.

00:11:14.280 And so the USPSTF's calculus went something like this.

00:11:18.160 If you're only saving one-tenth of one percent of men while causing serious complications,

00:11:24.280 all the things we talked about, right, infections, incontinence, erectile dysfunction,

00:11:28.360 the stress that's associated with all of these things,

00:11:31.220 In a meaningful proportion of the others, the math simply doesn't add up. The harms of over-diagnosis

00:11:38.620 and over-treatment exceed the benefits. Now, again, I can see how one could arrive at that

00:11:45.160 conclusion. It's a little bit of a utilitarian argument, and there's a logic to it at the

00:11:50.740 population level. But that logic only holds if the underlying data are sound, which we'll discuss,

00:11:57.520 and it's been 15 years now. If the USPSTF got it right, we should see at most a marginal uptick

00:12:06.980 in late-stage disease, maybe a fraction of a percent. So the obvious next question is,

00:12:13.120 what has actually happened over the past 15 years? Well, the data are now in and they paint a very

00:12:19.440 clear and very troubling picture. When you group together all prostate cancers caught at stages

00:12:26.420 one through three, so these are non-metastatic, five-year survival exceeds 99%. But at stage four,

00:12:35.200 which is metastatic disease, tumors that have spread beyond the prostate, survival plummets

00:12:39.940 to 38% with a median survival of about 30 months from diagnosis. So the timing of when you find

00:12:46.180 the cancer is not a minor detail. It's arguably the determining factor in life and death.

00:12:53.540 So what does the recent literature show?

00:12:55.740 Well, data published within the last six months from both the American and Canadian registries,

00:13:01.980 which builds on the earlier work from Ted Schaefer and his colleagues in 2016,

00:13:07.040 is that we are finding more and more prostate cancers only after they've become metastatic.

00:13:12.860 This is exactly what many urologists predicted would happen when we pulled back on screening.

00:13:18.160 But what's even more alarming is that it's not just older men who are affected. The Canadian

00:13:23.960 data show a 3.7 year-over-year increase in stage 4 diagnoses among men under 75 between 2010 and

00:13:33.680 2017. And this is happening despite a declining total number of cancers detected in that same

00:13:40.040 age group. So it's not that more men are getting cancer, it's that we're only finding it once it's

00:13:45.780 too late to cure. The American data tell a similar story. Since the 2012 guideline shift,

00:13:51.720 stage 2 cancer detection has been declining by about 0.1% per year across all ages.

00:13:59.100 Meanwhile, stage 3 has been climbing at 3.3% per year and stage 4 at 6% per year. Those increases

00:14:09.320 in late-stage detection far outpace the modest 0.8% annual increase in total cases.

00:14:17.780 So the interpretation here is not complicated, and it is not ambiguous. We stopped looking for

00:14:22.980 early disease, so we started finding it late. Stage 4 cancer is what drives all prostate cancer

00:14:30.800 mortality, or nearly all of it. And the trajectory we're on is the entirely predictable result of a

00:14:37.960 policy that told doctors and patients that looking wasn't worth the trouble. I think the Canadian

00:14:43.420 researchers summarized this paradox well, and I'm paraphrasing, but the gist was newer treatments

00:14:49.040 have extended survival for men with metastatic disease, but overall survival across the

00:14:55.480 population has actually worsened compared to when screening was recommended, because we now have so

00:15:01.520 many more men presenting with incurable cancer. And it doesn't have to be the story. It just

00:15:08.060 shouldn't be this way. So this is where I want to pivot because I think this is actually the

00:15:13.000 most important part of the conversation and honestly the most hopeful. The USPSTF built

00:15:17.980 its guidance at a time when our tools were far cruder than they are today. The old pipeline,

00:15:24.780 as I kind of talked about a few minutes ago, was essentially PSA goes up, you get an ultrasound,

00:15:28.580 ultrasound looks somewhat concerning. You got a biopsy. If the biopsy finds any type of cancer,

00:15:32.920 you get treated. And there was very little nuance in that algorithm. But the way we deploy PSA in

00:15:39.040 2026, and quite frankly, over the last five years, looks poignantly different. The single biggest

00:15:44.520 conceptual shift is this. PSA is most value as a longitudinal trend, not a single snapshot.

00:15:51.840 Any individual PSA reading is inherently noisy. Day-to-day values can fluctuate by as much as 15%.

00:15:58.580 If a man has recently ejaculated, his PSA can spike by as much as 40%, and the person-to-person variability is staggering.

00:16:06.240 A healthy man in his 60s might have a median PSA of 1.0 to 1.2, but the 95th percentile for that exact same demographic in men with no prostate cancer extends all the way up to 4.9.

00:16:20.100 So in the show notes, we're going to leave a table that gives you all of this information.

00:16:25.560 We break it down by a decade of life.

00:16:27.400 but the point here is that no single PSA value in isolation tells you very much.

00:16:32.440 What does tell you something is the trajectory, and this is what we call PSA velocity, mapping

00:16:38.920 a patient's baseline over time and tracking the rate of change of that PSA value. The prostate

00:16:47.100 naturally enlarges with age, so some gradual PSA increase is expected, but rapid increases are

00:16:54.980 cause for suspecting something pathological. Either the prostate is enlarging abnormally,

00:17:00.460 a higher proportion of PSA is leaking into the blood relative to the amount of prostate,

00:17:05.380 tissue, or both, and disrupted prostate architecture and vascular leakage are the

00:17:10.720 hallmarks of malignant growth. So in general, if we look at increases sustained over an 18-month

00:17:18.120 window and use different cutoffs depending on the baseline value, so if we took a patient with a

00:17:24.320 baseline PSA of less than four nanograms per milliliter, and in our practice, that's the norm.

00:17:29.860 Most of our men are going to walk around with a baseline PSA less than four. A sustained rise

00:17:36.140 by more than about 0.35 nanogram per milliliter per year over 18 months is generally our red flag.

00:17:45.960 In patients with a PSA value that starts above 4, we would let that increase to about a 0.75 nanogram per milliliter per year bump before we would flag that.

00:17:58.820 So let's make this tangible.

00:18:00.160 So let's say I have a 55-year-old patient with a PSA of 1.3.

00:18:04.100 So that's slightly above the median for his age.

00:18:07.300 But I'm not alarmed, especially because, remember, I've got longitudinal data on him, and I realize that this is about where he settles out.

00:18:13.520 So that one time he was 1.6, I didn't overreact because two months later or three months later, he was back to 1.3.

00:18:20.880 Now, if he comes back a year later at 1.7, and then six months after that comes back at 1.9, that's an increase of 0.6 over an 18-month period of time, working out to a velocity of 0.4 per year.

00:18:38.460 So now I'm paying attention. That sustained acceleration is the signal I'm looking for,

00:18:44.060 and it's personalized to him. This is actually what precision medicine is, and incredibly it

00:18:49.500 costs essentially nothing because this test is so cheap, like we're talking dollars, and it shows up

00:18:56.880 in a routine blood test. So this is a great example of precision medicine, and incredibly it costs

00:19:02.960 essentially nothing beyond showing up for a routine blood draw. Again, the blood test itself

00:19:08.720 costs dollars. But even with his velocity flagging a concern, we no longer would just jump to a

00:19:16.920 biopsy here. The next step would now be an MRI, which serves as a much higher resolution test,

00:19:23.580 and it's much less invasive than proceeding to a biopsy. So the MRI gives us yet another metric,

00:19:30.940 which is PSA density, which is simply the PSA concentration in the blood divided by the volume

00:19:36.260 of the prostate provided by imaging. And by the way, you could do that with an ultrasound as well.

00:19:41.740 But with the MRI, you're getting two for one because you're getting a better imaging study,

00:19:46.340 and of course, you're getting volume. So here's the logic. A healthy prostate grows,

00:19:50.640 PSA and volume tend to increase in approximate proportionality. But when you have a malignancy,

00:19:57.360 the tumor can disrupt the normal architecture of the prostate that normally contains PSA

00:20:01.800 to the gland or alter the endothelia running through the prostate. And in this situation,

00:20:07.780 PSA starts escaping into the bloodstream at a rate that outpaces what you'd expect

00:20:12.780 from the size of the gland alone. So if PSA is rising disproportionately to prostate volume,

00:20:19.300 that's a strong indicator that something is wrong. And it gives us that signal before anybody needs

00:20:25.200 to think about a needle. So when you put your PSA velocity from serial blood test together with PSA

00:20:31.180 density from MRI, the discriminating power of PSA screening today is dramatically better than what

00:20:38.800 existed when the USPSTF made its call. Now, if we focus in on MRI for a moment, there's been a really

00:20:45.540 significant development that I think is going to meaningfully change the accessibility equation.

00:20:51.300 Beyond PSA density, the main value of MRI in this, you know, algorithm is that it can identify

00:20:57.280 suspicious lesions or regions in the prostate without invasive probing. It can credibly rule

00:21:03.900 in or rule out the need for a biopsy. And if a biopsy is warranted, it guides the needle to the

00:21:09.460 areas most likely to hand us a diagnosis. The obstacle historically was that the gold standard

00:21:16.300 prostate MRI called a multi-parametric MRI requires a contrast called gadolinium. And that meant higher

00:21:24.500 costs, the need for a medical professional to administer an IV injection, and significantly

00:21:29.000 more time in the scanner. All of this, of course, limited patient throughput and made pre-biopsy

00:21:34.500 MRI impractical in many scenarios. But a trial in 2025 called the PRIME trial changed this. The study

00:21:41.240 directly compared the full multi-parametric protocol, which is a T2-weighted image,

00:21:47.420 a diffusion-weighted image, plus dynamic contrast-enhanced imaging, against a stripped-down

00:21:53.720 contrast-free biparametric version that drops the contrast component altogether. The goal was to

00:22:00.900 establish non-inferiority. In other words, the question asked was, does this stripped-down

00:22:06.400 version perform at least as well as the gold standard. And the results were essentially a

00:22:11.320 dead heat. The biparametric MRI identified clinically significant cancer in 143 out of 490

00:22:19.680 men who had been flagged by PSA screening, whereas the full multiparametric MRI complete with

00:22:25.980 gadolinium caught 145 out of that same group of 490 study participants. The difference between

00:22:35.320 those detection rates was 0.4 percentage points, which was statistically indistinguishable.

00:22:41.700 And the practical implications here are huge, obviously, right? Without contrast,

00:22:45.360 the scan takes half the time, 15 to 20 minutes instead of 30 to 40. You don't need

00:22:49.440 a medical practitioner that's standing there giving you the contrast and overseeing you.

00:22:55.000 And of course, the costs come down, speed goes up, etc. Now, even with all of these improvements

00:23:00.300 in non-invasive screening, some men are still going to need a biopsy. If the MRI flags something

00:23:06.820 suspicious and you need tissue to know what you're dealing with, you've only got one place to get it.

00:23:11.980 So it's worth talking about the progress that's been made there as well, because the biopsy itself

00:23:16.860 has gotten both safer and more accurate. Now I mentioned earlier that the traditional approach

00:23:21.980 was a transrectal approach, and that carries inherent risk of poking bacteria from the rectum

00:23:29.500 into a sterile space. It also has its anatomical limitations. It biases your sampling towards the

00:23:36.620 back of the prostate near the rectal wall, and this can miss the front and base of the gland.

00:23:42.780 The alternative to this is called a transperineal approach, entering through the skin between the

00:23:48.000 anus and the scrotum, and this completely bypasses the rectum, which dramatically reduces infection

00:23:52.980 risk, and it opens up much better access to the anterior and apical portions of the prostate

00:23:59.400 than the transrectal route does, which you tend to miss. So Ted Schaefer, who's been a guest on

00:24:05.020 this podcast many times, is a very close friend of mine personally and is also the urologist that

00:24:10.560 works with many of our patients when they are diagnosed with prostate cancer, was involved in

00:24:16.120 a large multi-center phase three trial with over 1,700 patients that compared cancer detection

00:24:22.520 rates and complication profiles between the two approaches head-to-head. The transperineal biopsy

00:24:29.260 detected more high-grade cancers and fewer low-grade ones, which tells you it's doing a

00:24:34.880 better job of finding tumors that actually matter, generating fewer false negatives from indolent

00:24:40.840 disease and obviously creating less emotional distress. And on the safety front, this is really

00:24:46.620 the headline in my mind, not a single patient who underwent the transperineal biopsy in that study

00:24:52.420 developed an infection, ZIP0. Compare that to the 5-7% infection rate widely reported in the

00:24:59.760 transrectal approach. Importantly, the transperineal method doesn't require prophylactic

00:25:05.120 antibiotics, which is a meaningful benefit both for the patient and for the antibiotic stewardship

00:25:10.680 more broadly as we face an era of antibiotic resistance. Now, adoption isn't universal yet.

00:25:16.760 Only about 37% of American urologists currently perform transperineal biopsies versus a

00:25:22.300 essentially all of them, doing transrectal or being able to. So there's a gap between what

00:25:27.880 the evidence supports and what's happening in practice, but the trajectory is encouraging.

00:25:34.300 All right, so I've been building the case for more screening, for more MRI, and for better biopsies,

00:25:39.160 but I haven't forgotten the other side of the original argument, which is the overtreatment

00:25:43.200 problem. And I really take this seriously because with all of the advances in surgical care,

00:25:49.800 this is not a benign operation. Not every prostate cancer is going to kill you. Our treatments are

00:25:56.280 effective, but they carry side effects that nobody wants to experience if they don't have to.

00:26:01.640 So here's the critical point that I think changes the overtreatment calculus. We no longer treat

00:26:08.600 every cancer we find. The decision tree has gotten much more sophisticated. Now, if you remember the

00:26:16.160 podcasts that we've covered in this topic in the past, we go through something called a Gleason

00:26:21.920 scoring, but I'm just going to kind of reiterate it here a little bit because maybe not everybody's

00:26:26.200 familiar with that previous content. So when a pathologist examines a biopsy specimen, they

00:26:31.980 look at the cells under the microscope and they decide how abnormal the cells look relative to

00:26:39.240 healthy cells. And they use a grading scale called a Gleason score. Now you're getting two numbers in

00:26:46.540 that score because they're looking in two places. So you might get a Gleason 3 plus 3 or 3 plus 4

00:26:53.980 or 4 plus 4. And the highest number there is 5. And the higher the number, the greater the malignant

00:27:02.960 potential of the cell. Now, these numbers get compressed into a grade group, which is a scale

00:27:12.400 that is graded one through five. And again, it's similarly based on the abnormality of these cells.

00:27:19.020 So a grade one or two is a relatively low risk tumor. We know the natural history of grade one

00:27:29.680 and grade two, which are the lower Gleason scores, behave very differently. And these are cancers

00:27:36.000 that may never spread. Now, we don't know that for a fact, but we know that if they're going to

00:27:42.280 spread, they're going to do so by progressing into a higher grade Gleason score. And therefore,

00:27:47.620 for these patients, we can continue to survey them or do other advanced forms of monitoring,

00:27:54.680 which I'll talk about in a second. Conversely, if a patient presents with a higher grade group,

00:28:01.220 like a grade 3, which would be a Gleason 4 plus 3, a grade 4, which would be a Gleason 4 plus 4,

00:28:08.360 or a grade 5, which is anything higher, so a 4 plus 5, 5 plus 4, or any 5-5, these are very

00:28:15.480 aggressive tumors, and treatment is immediately warranted. So going back to what do you do with

00:28:23.800 grade one and two. Well, again, these are patients that have prostate cancer, but our belief is that

00:28:31.780 it is not an imminent threat to their life because it is not clear if these cancers will spread.

00:28:37.860 And if they're going to turn into spreading cancers, they're going to do so by progressing.

00:28:42.260 So we would, in this case, look at other blood markers like tests called the PHI or the 4K score.

00:28:49.660 These are things that we talked about on previous podcasts. We'll link to some of that stuff in the

00:28:53.720 show notes if you want to follow that in more detail. But the point is we now turn into an

00:28:59.620 area called active surveillance. And active surveillance is basically something we're doing

00:29:06.500 to ensure that treatment will be timely once it is required, but no sooner. So the scenario that

00:29:15.620 drove so much of the overtreatment concern 15 years ago, which was you find any cancer, you

00:29:20.620 treated immediately and you deal with the consequences, that's largely been retired from

00:29:24.840 the urologist playbook. The tools and clinical judgments that have evolved today to the point

00:29:30.540 where urologists can meaningfully distinguish between dangerous tumors and ones that can be

00:29:35.400 safely watched is what basically gives us this luxury. And that means that the risk-benefit

00:29:40.100 calculation for screening has shifted substantially in favor of screening. Okay, so I spent a good

00:29:47.360 amount of time walking through the modern sort of screening toolkit, and I've tried to be fair

00:29:54.160 to the original USPSTF reasoning along the way. I think the over-treatment concern was legitimate,

00:30:01.620 and I think the burden of transrectal biopsies was real. And again, I take those objections

00:30:07.600 seriously. But there's something I've sort of held back on until now, and it's a failure in

00:30:12.720 the data that I think is so significant that it essentially undermines the entire foundation of

00:30:18.740 the USPSTF position. Urologists have been pointing this out for years, and I think it's borderline

00:30:25.220 unbelievable that it hasn't led to a formal revision of the guidelines. So I'm talking about

00:30:30.120 that PLCO trial. That's the study that the USPSTF weighted most heavily when it decided that PSA

00:30:37.460 screening wasn't worth doing. Here's what the PLCO study was supposed to show. You take a large

00:30:43.220 group of men, you randomize half to get regular PSA tests, the other half not to get tested,

00:30:47.120 you follow them over a decade, and compare prostate cancer deaths between the two groups.

00:30:51.540 The study found essentially no difference, supporting the conclusion that screening

00:30:55.420 doesn't meaningfully reduce mortality. But here's what went wrong, and I genuinely had trouble

00:31:01.440 believing this the first time I encountered it. The control arm, so this is the group that was

00:31:06.020 not supposed to be screened, was getting screened anyway. The investigators may not have been

00:31:12.840 providing the tests, but when they surveyed the participants, somewhere between 40 and 60 percent

00:31:19.580 of men in the no-screening group reported having had a PSA test within the year.

00:31:26.280 And this wasn't a one-time contamination event. That pattern held every single year of the study.

00:31:32.320 So by the final survey, over 90% of the men in the control arm had undergone at least one PSA test.

00:31:39.360 In some years, more men in the supposed control arm reported screening than in the intervention arm.

00:31:47.160 So we will put the figure from the 2016 letter to the editor in the New England Journal of Medicine in the show notes

00:31:54.060 so you can see this for yourself because it is quite striking.

00:31:57.220 The bottom line is, this was a study comparing PSA screening to PSA screening and then concluding

00:32:03.160 that screening doesn't work when they didn't find a difference. That is the bedrock upon which the

00:32:09.740 USPSTF built a recommendation that has affected tens of millions of men. Now, this data contamination

00:32:17.400 wasn't published widely until 2016, well after the original guideline change. Once it was out there,

00:32:26.160 a group of researchers went back to the PLCO data in 2017 and did something really clever.

00:32:31.980 They mathematically accounted for the frequency of screening within the contaminated control arm,

00:32:38.420 calculated a mean and lead time, essentially how much earlier cancers were caught in men who were

00:32:44.240 more regularly, and then imputed a mortality reduction based on earlier diagnoses. When you

00:32:51.680 run the analysis that way, correcting for the contamination that made the original study

00:32:57.040 essentially uninterpretable, the same PLCO data show a 27 to 32 percent reduction in prostate

00:33:04.920 cancer mortality attributable to regular PSA testing. I just want to sit with that for a

00:33:12.300 second. The same data set, the same patients, analyzed honestly, and it shows that screening

00:33:19.020 prevents roughly a quarter to a third of prostate cancer deaths. And yet the USPSTF has not updated

00:33:26.900 its guidelines since 2018. It still does not recommend PSA testing for men 55 to 69, and

00:33:35.480 still recommends against it for men over 70. I think it's long past time for the USPSTF to

00:33:42.560 revisit this. The same is true of organizations like the Canadian Task Force on Preventive

00:33:47.920 Healthcare and the American Association of Family Physicians. Until they do so, I think we have

00:33:52.640 reason to be confident in taking this into our own hands. So let me try to synthesize everything

00:33:58.700 we've covered because I realize it's been a lot. We started with the observation that pulling back

00:34:04.280 on PSA screening led to a measurable ongoing increase in the incidence of incurable prostate

00:34:11.040 cancers. That's not speculation. It's in the data. It's black and white from multiple countries.

00:34:16.360 We then walk through how PSA testing has evolved. A single PSA value taken once every few years was never a perfect screening tool. It's too noisy, too much individual variation. That part of the old critique was fair.

00:34:30.260 But PSA velocity, PSA density, contrast-free MRI, and transperineal biopsies, along with active surveillance for low-grade disease taken together, represent a fundamentally different screening paradigm than what existed when the USPSTF issued its recommendations.

00:34:50.080 The ability to catch dangerous, i.e. lethal cancers, early while avoiding the harms of

00:34:56.740 over-diagnosis and over-treatment has improved by an order of magnitude. And then we looked at

00:35:02.600 the data that supposedly justified the shift away from screening 15 years ago and found that the

00:35:08.680 cornerstone study was so contaminated as to be essentially meaningless and that a proper analysis

00:35:16.420 of the same data clearly supports screening very strongly. So where does that leave us? Well,

00:35:22.040 prostate cancer, as I said, number two cancer killer of men, and it's not receiving the urgency

00:35:27.740 it deserves. I'll go further. I think it's unconscionable that we are still losing men to

00:35:32.500 a disease that in many cases could have been caught years before it became lethal. And this

00:35:38.380 is something I want people to really internalize because it goes to the biology of the disease

00:35:43.740 itself. I've talked about this many times on previous podcasts, but I don't think I can talk

00:35:48.540 about it enough. Cancer falls into roughly two categories. There are those that progress in a

00:35:54.500 relatively predictable, stepwise fashion, moving through detectable precursor stages before

00:36:02.660 becoming truly dangerous. And then there are those that don't. They can emerge suddenly,

00:36:08.300 metastasize quickly and evade early detection no matter how aggressively you screen. I think that

00:36:15.520 pancreatic cancer is the classic example of the latter, and sadly, many breast cancers also fall

00:36:21.860 into that category. But prostate cancer, along with colorectal cancer, fit squarely in the first

00:36:29.300 group. Prostate cancer tends to follow a very predictable path from early localized disease

00:36:36.020 to advanced metastatic disease. And that biological reality is precisely what makes a screening

00:36:43.540 detectable tool like the PSA so powerful and what makes any deaths, let alone the 36,000 a year in

00:36:52.140 this country alone from prostate cancer, so completely tragic. This isn't a cruelty of

00:36:58.460 biology. It's a failure of our implementation. I've made this comparison before, but I think

00:37:04.120 it's worth repeating. Imagine we had a breast cancer detection tool that cost less than a cup 0.84

00:37:09.820 of coffee, could be incorporated into a routine blood test, and was capable of flagging cancer 0.91

00:37:15.980 years before it spread. In what universe would we not deploy that tool as widely and aggressively

00:37:22.100 as humanly possible? For prostate cancer, that tool does exist. Your primary care doctor can

00:37:28.820 add it to your next blood draw. It might already be there. The barrier isn't technology. It's

00:37:33.260 awareness, it's education, and most of all, it's outdated guidelines. In our practice, we've seen

00:37:39.020 this play out countless times. Well, patients will come to us with PSA data from other providers,

00:37:43.720 sometimes years of it, and when we calculate their velocity, the red flags are right there.

00:37:48.040 We send them for an MRI, depending on what it shows, a biopsy, and we've caught several cancers

00:37:54.780 right at that stage where had we waited another year or two, they likely would have been metastatic,

00:38:00.900 but they were Gleason 3 plus 4s or 4 plus 3s that were right in that sweet spot for treatment.

00:38:06.580 Before I close, I want to just sort of spend some time on a topic that is related to this,

00:38:12.140 but a little bit oblique to it. And I think it's also becoming increasingly prevalent. And it's

00:38:19.020 generally, I think, something we need to be alarmed about and physicians and patients alike

00:38:23.520 need to be aware of, given the current state. And it's the interaction between a class of drugs

00:38:29.480 called 5-alpha reductase inhibitors. The most common of these is finasteride, but there's

00:38:34.980 another one called dutasteride, and how they impact prostate cancer screening. So the use of

00:38:40.600 finasteride has exploded in recent years, largely driven by men using it to prevent or treat hair

00:38:47.000 loss. In fact, we've talked about this drug many times on other podcasts, both in the context of

00:38:51.240 hair loss and testosterone use. We've discussed the potential risks of finasteride and prostate

00:38:57.200 cancer development elsewhere, but I want to have some dedicated sections here where I talk about

00:39:02.100 that. So I want to focus on one specific issue, which is that finasteride, as far as we can tell

00:39:09.140 from the data, does not cause prostate cancer. Okay. But what it absolutely unequivocally does

00:39:16.160 is suppress PSA in a way that can completely mask the present of cancer if the treating physician

00:39:23.140 doesn't know how to adjust for it. Now here's the mechanism. Finasteride blocks the conversion

00:39:28.180 of testosterone into dihydrotestosterone, or DHT, in the prostate. And that reduction in local DHT

00:39:35.700 also dramatically reduces PSA production. As a practical matter, you can expect PSA to drop

00:39:43.600 by about a third within six months of starting finasteride and to be roughly cut in half by a

00:39:50.880 year. That suppressed baseline persists for as long as the patient stays on the drug.

00:39:56.940 So let me state this as clearly as I can. If a man has been on finasteride for a year or longer,

00:40:02.840 his PSA value needs to be at a minimum doubled to be interpreted correctly. And the correction

00:40:10.800 factor we use actually increases with duration of use. In our practice, we multiply by 2.3

00:40:18.460 after 2 to 7 years on the drug and 2.5 beyond 7 years. Now, while the FDA label recommends

00:40:26.980 multiplying by 2, we do this based on data from the Prostate Cancer Prevention Trial performed in

00:40:32.580 the late 90s and early 2000s, as well as a follow-up analysis from 2005 performed by Ruth

00:40:39.840 Etzioni and her colleagues. We've included a table from that in the show notes. For PSA velocity,

00:40:47.860 the practical implication is simpler but equally important. After about 12 months on finasteride,

00:40:53.880 you should not expect your PSA to rise at all. The ongoing suppression from the drug should

00:41:00.120 roughly offset the natural age-related increase in PSA. In fact, in a large study of 19,000 older

00:41:06.880 men, those on finasteride who were cancer-free actually showed an average 2% annual decrease

00:41:13.440 in PSA. So any upward movement in a man on long-term finasteride should be treated as a

00:41:19.780 serious warning sign. And here's where I get concerned. Most general practitioners are not

00:41:25.740 aware of this, it appears. A retrospective analysis of VA patients undergoing PSA screening

00:41:31.360 found that men on finasteride were diagnosed with metastatic cancer at a rate more than double that

00:41:37.680 of men not taking 5-alpha reductase inhibitors, 6.7% versus 2.9%. The doctors were reading the

00:41:47.120 raw PSA numbers at face value, not realizing they needed to double them or more. The two groups had

00:41:53.080 nearly identical PSA levels at the time of cancer diagnosis, about 6.8 versus 6.4, but once you

00:41:59.680 correct for the finasteride-induced suppression, the true comparison is more like 13.5 versus 6.4.

00:42:06.140 Those are wildly different clinical pictures, and the physicians unfortunately missed it.

00:42:11.780 This is in some ways an even more tragic version of the same problem I've been talking about for

00:42:16.600 the past hour, because even in men who were getting screened, the screening was rendered

00:42:21.900 useless because the physician didn't understand how to correct the results.

00:42:26.700 And I think that's especially urgent because many men are getting finasteride prescriptions from

00:42:30.880 online clinics and then not even mentioning it to their primary care doctors because they

00:42:35.620 don't think it matters. So even a physician who does understand the PSA interaction can't account

00:42:41.580 for it if they don't know the patient's taking it. So if you're listening to this and you're

00:42:46.240 taking finasteride, the onus may fall on you. Please be sure that you not only advocate for

00:42:52.780 regular PSA testing, but make sure that your doctors know about every prescription, including

00:42:58.460 and especially those that involve 5-alpha reductase inhibitors, and ensure that whoever

00:43:02.760 is interpreting your PSA understands both the drug and its suppressive effects. Again, finasteride

00:43:08.760 does not appear to increase your actual risk of developing prostate cancer, but it can absolutely

00:43:14.180 prevent your doctors from catching it on time. So I want to end where I think the evidence points,

00:43:20.440 which is a place of cautious but genuine optimism. Between our understanding of PSA velocity and

00:43:27.240 density, again, the increasing accessibility of very high quality MRI, the advent of transperineal

00:43:34.520 biopsy, and the sophistication of active surveillance protocols for low-grade disease,

00:43:40.300 such as those Gleason 3 plus 3s. The field has made enormous progress in minimizing the very

00:43:47.460 real risks of overdiagnosis and overtreatment that motivated the USPSTF's original positioning.

00:43:54.360 And then, of course, against that, we have to weigh the cost of doing nothing, which is a little over 36,000 lives this year in the United States.

00:44:05.120 We are in an extraordinarily fortunate position.

00:44:08.140 We have the tools for early detection that are cheap, simple, and effective.

00:44:12.340 We have treatments that work.

00:44:13.860 We have the ability right now to catch most prostate cancers before they become fatal.

00:44:18.580 we just need to use what we have to advocate for it as patients and to understand it as clinicians.

00:44:25.600 And I genuinely believe that if we do, prostate cancer deaths can be all but eradicated. So it

00:44:32.900 doesn't have to be the second leading cause of cancer death in men. It could not even be in the

00:44:37.880 top 20. So a summary of how we approach PSA testing in our own practice can be found in a

00:44:45.740 newsletter that we published yesterday on the same topic. And we'll also link to it in the show notes

00:44:50.620 where you can also find all of the studies and references from today's episode as well. So I

00:44:56.800 hope you enjoyed this episode on a topic that I'm very passionate about. Thank you for listening to

00:45:04.840 this week's episode of The Drive. Head over to peteratiamd.com forward slash show notes if you

00:45:11.800 want to dig deeper into this episode. You can also find me on YouTube, Instagram, and Twitter,

00:45:17.400 all with the handle Peter AtiyahMD. You can also leave us review on Apple podcasts or whatever

00:45:23.320 podcast player you use. This podcast is for general informational purposes only and does

00:45:29.000 not constitute the practice of medicine, nursing, or other professional healthcare services,

00:45:33.140 including the giving of medical advice. No doctor patient relationship is formed.

00:45:37.920 The use of this information and the materials linked to this podcast is at the user's own risk.

00:45:44.340 The content on this podcast is not intended to be a substitute for professional medical advice, diagnosis, or treatment.

00:45:50.180 Users should not disregard or delay in obtaining medical advice from any medical condition they have, and they should seek the assistance of their healthcare professionals for any such conditions.

00:46:00.300 Finally, I take all conflicts of interest very seriously.

00:46:04.400 For all of my disclosures and the companies I invest in or advise, please visit peteratiamd.com

00:46:10.900 forward slash about where I keep an up-to-date and active list of all disclosures.

The Peter Attia Drive - April 20, 2026

#388 — Prostate cancer screening： why current PSA guidelines are failing men and how modern tools improve early detection and save lives

Episode Stats

Summary

Transcript