How PSA Came to Be Indispensible
Twenty Years of Contributions from the Baltimore Longitudinal Study of Aging
Bal Carter, PSA pioneer. On his computer screen are classic changes in PSA over
20 years, in men with no prostate problems (yellow), BPH (light purple), localized
cancer (tan), and metastatic cancer (blue). He published these data in a landmark
study 20 years ago. Without the resources of the BLSA, we would still be waiting to
know how PSA changes reflect what's happening in the prostate.
We take it for granted now: PSA, the most versatile weapon in the prostate cancer testing and monitoring arsenal. New uses for PSA show up all the time. It can be parsed - split into "free" and "bound" forms, which tell us different things about the likelihood of BPH and cancer within the prostate. There's PSA velocity - how fast it changes over time - and PSA density, which factors in the weight of a man's prostate. Doctors measure PSA before a man is ever diagnosed with cancer, and for years after he's treated. It's indispensable; one of the "big three" facts (along with the Gleason score and clinical stage) that helps predict how successful treatment will be. Treating prostate cancer today without factoring in PSA is as unthinkable as building anything from Ikea without an Allen wrench.
What most people don't realize is that the vast majority of our understanding of PSA has come from research done over the last 20 years by Hopkins urologist H. Ballentine Carter, M.D., using one remarkable resource - the National Institute on Aging's Baltimore Longitudinal Study of Aging (BLSA). "This work has changed the fi eld," says Patrick C. Walsh, M.D., "When PSA was first discovered, we were facing a thousand unanswered questions, which under normal circumstances would have taken decades to answer. Now, our national guidelines are based on these observations."
What Carter was able to do, using the BLSA data, was like time-lapsed photography. Using decades' worth of blood samples from middle-aged men as they grew older, Carter and colleagues watched what happened to PSA (a protein made by the prostate) over time, in men who developed prostate cancer and men who did not. In honor of Carter's 20 years of life saving contributions to our understanding of how to use PSA, we recently asked him how it all started:
Almost a Blank Slate
When you arrived at Hopkins, knowledge of PSA was really still in its infancy. Nobody understood how to use it to diagnose prostate disease.
I remember talking to Pat Walsh (then director of the Brady) about it. The general consensus was that PSA would not be useful for diagnosis because it was elevated by both BPH and prostate cancer. But Dr. Walsh asked me an interesting question: "Do you think PSA rises faster in men with prostate cancer than in those without it?"
How would you go about answering this question?
That's the problem. It would take two to three decades to follow the PSA levels in a group of men, and then see whether or not they developed prostate cancer. But this undertaking seemed worthwhile, given our lack of knowledge about PSA. A few weeks later, Dr. Walsh asked me if I had heard of the Baltimore Longitudinal Study of Aging. I had not. (The BLSA is America's longestrunning scientific study of human aging. It was begun in 1958 by the National Institute on Aging, part of the National Institutes of Health as a way to study what happens to our bodies as we get older.) Dr. Walsh believed that the BLSA had a frozen serum bank, and suggested that I might want to find out if PSA could be measured in those samples. If it could, that would allow us to find out quickly what changes occur in PSA with the development of prostate disease and with age.
What a stroke of luck, that you had this conversation at the right time, and you were at the right place to pursue this.
Absolutely. Serendipity and an unanswered question of importance brought me together with investigators at the BLSA, just 15 minutes away. It turned out that the BLSA had been storing blood in freezers since 1958 at two-year intervals, and that men usually entered the study in mid-life and were followed to followed for decades until death. This stored blood could provide us a picture of PSA as men aged - some developing and some dodging prostate disease.
One of the great discoveries to come out of this work was PSA velocity. How did this come about?
I had the unique opportunity to work with investigators at the BLSA, James Fozard, Jeff Metter, Jay Pearson, Larry Brant, Reuben Andres. We found that in men without prostate disease, average PSA levels remained around 1 ng/ml. For those who developed prostate enlargement, PSA levels started out around 1, but increased to around 3 over the next two decades.
Treating prostate cancer today
without factoring in PSA is as
unthinkable as building anything
from Ikea without an Allen wrench.
Much of that is due to Bal Carter’s
longtime work with the BLSA.
But for the men who developed prostate cancer, the picture was very different: The rise in PSA was much faster for men with prostate cancer compared to men without. Even more exciting: Five years before the diagnosis of prostate cancer was made, the rate of rise in PSA could reliably distinguish men with and without prostate cancer. We coined the term PSA velocity to describe the rate of rise in PSA , and suggested that among men with PSA levels between 4 and 10, a PSA velocity greater than 0.75 ng/ml per year could predict the presence of prostate cancer. Today, the National Comprehensive Cancer Network recommends that physicians use PSA velocity as an indicator of the possible presence of prostate cancer.
PSA Velocity and Lethal Prostate Cancer
PSA velocity gave us a five-year head start on diagnosing prostate cancer, but you were able to refine your results and even predict years ahead of time which men were likely to have more aggressive disease.
By 2006, PSA had become a routine part of clinical practice, and many investigators were concerned about over-detection and over-treatment of cancers picked up through PSA testing - especially when prostate biopsies were performed at PSA levels below 4. On the other hand, we knew that some men with low PSA levels had life-threatening cancers. We wanted to know whether PSA velocity could predict whether a man would develop lethal prostate cancer decades later. If that were true, then PSA velocity could be used to stratify men with low PSA into two categories: Those who should undergo a prostate biopsy, and those who could wait longer to see what happens to their PSA over time. We found that PSA velocity 10 to 15 years before the diagnosis of prostate cancer, when most men had PSA levels below 4, was closely associated with the risk of death from prostate cancer. We concluded that a PSA velocity above 0.35 ng/ml per year could help identify men who might otherwise be overlooked based on a PSA level alone. The NCCN now recommends the use of PSA velocity as one indicator of prostate cancer risk in men with low PSA.
Your work has also set the standards for when PSA screening should begin. In the early 1990s, the thinking was that all men should begin screening with a PSA test and digital rectal exam at age 50, and then do it again every year. That has changed.
In reading about screening for breast and cervical cancer, I learned that investigators were using results from prior tests to predict whether or not cancer would be diagnosed later on. So I began to wonder, do all men really need a yearly PSA test? If a man maintained a low PSA level year after year, could he afford to be screened less frequently? Would it be reasonable to perform a baseline PSA test at age 40, and then let the result tell us how frequently to repeat it? Once again, the BLSA helped answer these questions. Looking at the PSA levels of men in the BLSA, my colleagues and I discovered that if a man has a PSA level below 2, it was very unlikely that over the next two years it would increase to 3 or 4 ng/ml, which would trigger a prostate biopsy. But if the PSA was above 2, it was not uncommon two years later for a man to reach a level that would trigger a prostate biopsy. Combining these findings with surgical data from Johns Hopkins allowed us to recommend that missing a curable prostate cancer would be unlikely if men with a PSA below 2 were to undergo testing every other year.
How did you zero in on the idea of a baseline at age 40?
There were reasons - including evidence that younger men were more likely to have curable cancers - to believe that PSA testing in men younger than 50 would save lives. To explore this, I worked with two epidemiologists, Harry Guess and Kevin Ross. We built a model of prostate cancer development in a simulated population of men. We then tested different screening strategies for preventing prostate cancer death, and found that PSA testing beginning at age 40, again at age 45, then every other year beginning at age 50, not only reduced prostate cancer deaths, but also needed fewer resources (PSA tests and prostate biopsies) to diagnose a prostate cancer.
" It turned out that a single PSA
level taken in mid-life was a better
predictor of future risk than a
family history of prostate cancer."
It turned out that a single PSA level taken in mid-life was a better predictor of future risk than a family history of prostate cancer. For men in their forties with a PSA level above 0.6 ng/ml, and for men in their fi fties above 0.7 ng/ml, there was a three- to four-times greater risk of being diagnosed with prostate cancer over the next two decades than for men with lower PSA levels. Results from Europe and Scandinavia are confirming this concept of targeted screening, rather than the "one size fits all" approach.
Screening and Older Men
You have been concerned over the years about older men, diagnosed with small cancers, who might be receiving unnecessary treatment.
Yes, but also the other side of the coin, the older men with prostate cancer that could cause harm without treatment - cancers that would be missed if screening were discontinued in all older men. Some guidelines have recommended that men over 75 stop getting their PSA tested. Edward Schaeffer on our faculty, working with our colleagues at the BLSA, wondered if there was a PSA level that would help identify which older men could safely discontinue screening. Schaeffer and colleagues found that no men in the BLSA who had a PSA below 3 in their mid-seventies went on to develop a lethal prostate cancer. Since two out of three men at age 75 have PSA levels below 3, for the first time it was possible to present an alternative recommendation. Instead of telling all older men, "you don't need a PSA test because you are too old," it was possible to say, "based on your low PSA, you are at minimal risk of dying of prostate cancer, and you don't need to worry about PSA tests anymore." I have found that most men are relieved to hear that news.