PSA Screening Revisited

Last week I wrote about the controversy surrounding PSA screening tests and said I would write this week about the difference between screening and diagnostic tests. In the world of testing, there are screening tests and  diagnostic tests, both very different. As I wrote last week, much has been said today about the evils of the PSA screening. For those who recommend against testing—or even worse, the U.S. Preventative Services Task Force (USPSTF),suggesting physicians avoid bringing up the idea of PSA screening unless the patient asks for it—I have one question: “What are the alternatives?” Are you suggesting that we return to 70 or 100 years ago when the diagnosis of prostate cancer came by having incredible back pain, lung mets, brain mets, and prostate cancer seeping through the skin of the back?

Fifty-one years ago, when I was applying for medical school, the dean had advanced prostate cancer. In those days, treatment was physical castration and Premarin. The physical castration may have had some benefit because the testosterone which drives the prostate cancer was eliminated. But all that the Premarin did was give men enlarged breasts, and there were many men who opted out of that treatment, including my father-in-law.

Before that time, in the late 50s, my dad’s uncle was dying from prostate cancer. We went to visit him twice. He lay in a hospital bed at one end of the living room and was unable to get out of bed. The prostate cancer was oozing through the skin on his back. His wife dutifully changed his bandages, but the room smelled like decaying flesh, which of course permeated the room.

There was another man in the town next to us who had a similar problem. Then in the 70s two of my mother’s cousins, one on each side of her family, developed a prostate cancer which was very difficult to treat. That was again because it was discovered late. In my own case, in March 2006, at the age of 58, I decided to test my PSA for the first time. That was because I had a bad backache and I was reminded of my dad’s uncle. As I explained in my post last week, I was very grateful for catching my prostate cancer early with a PSA screening test.

There are differences between screening tests and diagnostic tests which patients should know about. The PSA is a screen, like a PAP smear or a mammogram. In other words, the screening test is designed to be sensitive, that is, to find out what is being searched for. The PSA screening test will find signs of anyone who might have prostate cancer as well as finding other problems in those who do not have prostate cancer. Finding or identifying all those who have prostate cancer is the purpose of a PSA screening test.

Again, the target group is the group which tests positive on the screening test and identifies those who actually have the disease. These are the true positives, the group the screen is designed to find. The false positive result is the group which tests positive for prostate cancer but does not have cancer. This is the group there is so much discussion about today and the reason there is so much criticism of PSA screening. The implication is that the test is so inaccurate it should not be done. The purpose and the limitations of the PSA screening or any screening test need to be firmly and accurately understood by everybody involved, provider and patient.

The PSA screen is just behaving as a screening test should. It is designed to pick up every man with prostate cancer and to not miss any. That’s the point. The trade-off to pick up every case (to be sensitive), is that it also picks up men who don’t have cancer. The PSA will pick up other things like inflammation. That is not a shortcoming of the test, but rather a misunderstanding of what every screening test does.

The purpose of any screening test, that is finding some false positives (men who test positive but who don’t have prostate cancer) in order to find all true positives (men who test positive and do have prostate cancer) needs to be understood clearly by every provider and that understanding needs to be communicated clearly to every patient. Not understanding the meaning of a screening test is not an option. Neither is not explaining the purpose and the limits of the screening test. The false positives are not a failure of the PSA screening test. False positives are part of any screening test. The screening test picks up other conditions such as inflammation along with the targeted indications of cancer.

What we should really be concerned about are the false negatives, the group which should be identified with prostate cancer but are not by the screening test. A screening test is designed to keep the false negatives to a minimum. And that’s the trade-off. In return for having a small number of false negatives, the screening test will also return a greater number of false positives.

False negatives are not necessarily failures of the PSA screening test or signs of screening inaccuracy. They are instead a problem with prostate cancers which do not make the protein which causes the PSA to rise. False negatives result from the way prostate cancer presents in some patients.

Anybody who is at all versed in screening and diagnostic testing understands these issues revolving around a PSA screening test. When a provider decides to present a patient with the option of a screening test, these possibilities need to be discussed. In today’s office environment of the five minute visit, it is extremely difficult to undertake any meaningful kind of informed consent.

But that’s what we’re talking about—informed consent. In other words, the provider must understand the meaning of screening tests and understand the range of possibilities. These characteristics of a screening procedure need to be discussed with the patient so that they understand all the possibilities. Of course, in the five or ten minute office visit, it may be just simpler to avoid the PSA test and therefore simply avoid a discussion of PSA testing and informed consent.

Let’s get over it! No screening test is going to be 100 percent accurate 100 percent of the time. One hundred percent accuracy has never been the goal of a screening test or for that matter, a diagnostic test. The screening test is sensitive, but not specific. Diagnostic testing (like a biopsy) is specific. That’s the point of following up a screening test with a diagnostic test to rule out the false positives. A diagnostic test should identify only the men who have prostate cancer and no others. Unfortunately, there are false negatives even with biopsies.

Enter active surveillance. One never really knows for sure how that’s going to be carried out. I expect this is the point of Francis Collins’ article, in other words, watching his prostate cancer go from easily treated to difficult to treat. Proponents of active surveillance and the USPSTF indicate that what has happened to Dr. Collins is relatively rare. In my experience this occurrence is not rare at all and constitutes approximately a third of the prostate cancer patients I saw in my practice.

Years ago, when I was a resident learning to do Pap smears and colposcopic examinations and cervical biopsies, I had a patient who had an abnormal Pap smear and normal biopsies. This is the same thing we see with the PSA group that is actively surveilled. Back then as a resident, I went to talk to my professor. He raised the issue that maybe the PAP smear was correct and maybe the biopsies were just missing the pre-cancerous areas? With biopsies, it’s possible that the tissue samples taken miss the cancerous areas. In all the discussion of PSA screening, no one mentions that biopsies can miss the target. As a physician, I would be much more concerned about false negatives than false positives.

As I wrote in my last week’s Substack post, that’s the reason I was actually relieved to hear that I had a Gleason seven in my biopsy specimens with neurovascular space involvement. I was not going to be relegated to the let’s-see-how-bad-this-can-get group.

A few years ago I was doing an ER locums and a 62-year-old man, a recently retired banker, presented with back pain. He had had back pain for some time and had had two CT scans, which indicated nothing according to the reports. He then had an MRI and this indicated lytic (bone eating) lesions in many vertebral bodies. Lytic lesions are compatible with cancer and in this case, prostate cancer.

He was preparing for a very good retirement, but instead ultimately needed to plan for his funeral, which happened within the year. Unfortunately, after the MRI turned up the lytic lesions, the radiologist went back to look at the CT scans. The radiologist could then see the undiagnosed lytic lesions on the CT scans. The point to all of this is that in this day and age, we should not be seeing prostate cancer patients present like they did 70 or 100 years ago.

The moral of the story is to stop blaming the PSA screening test for doing what a screening test is supposed to do. There are no good alternatives at this time, so get over it. Learn how any screening test actually works and explain that to your patients instead of choosing to not provide your patients with the information they need to make an informed decision.