Professor Emeritus, Dept. of Physics, University of Pennsylvania
March, 2001
The PSA score is used by urologists to search for the possibility that men are developing cancer of the prostate. If the score rises suddenly it is customary for the urologist to suggest that a biopsy be performed. While this is a simple procedure it is an invasive procedure and should be avoided if it is not necessary.
My urologist recently said that I should undergo a biopsy because my PSA had risen abruptly to 11.7 from about 5.5, its value in the previous year.
This suggestion seemed reasonable to me at first. But then I remembered that in 1996 I had gone to my internist because of a sudden stomach disorder and taken a new PSA when he noticed I had not had one for almost a year. It had risen to 45 which looked like a measurement error. To verify this I decided to repeat the measurement, following the score for several weeks as it fell back to 5.2.
The hitherto unpublished logarithmic plot of these data are shown in Fig. 1, showing the exponential shape of the falloff.
Since I had not been told that such fluctuations could occur and because the data verified my hypothesis that the PSA should fall back to normal exponentially with time. I wrote up this result (1) for my web page and later published a small "Letter to the Editor" in the Scientific American (2). A few days before the biopsy was to take place, I decided to see if the latest 11.7 reading might not be another upward fluctuation. My new PSA now showed a drop down to 6.2!
I repeated it a few days later to make sure it was not an error and to see if the results from two hospitals agreed. It was 5.7 which agreed within error with the 6.2 result. A third measurement weeks later gave the same score, 5.7. The 11.7 reading was taken around the time that I had suffered a small kidney infection.
Fig. 2 shows a plot of my PSA scores since 1992, showing the high spike in 1996 and the new spike in 1999.
So now, six years later, my last readings, as can be seen from Figure 2, were about the same as my average over the previous six years, 5.4 +/- 0.6. Therefore it did not seem that a biopsy was in order.
An ultrasound study found no evidence for concern. I present this material to suggest that it would be most valuable to study existing hospital records to search for other examples of sudden PSA increases, correlated with other infections or ailments, to determine which ones may distort the score. What is needed is a secure computer search of hospital records. Of course it is known that prostate infections, labeled ``prostatitis", can cause elevations. In that case one should see whether such an infection really is in place.
And if that is not the case one should not assume that the upward fluctuation is a necessarily a cause for concern. Other agents in the blood, appearing because of an unrelated infection, could interfere with the PSA measurement and be responsible for the sudden rise. (3)
Now that free PSA rather than just total PSA measurements are available, the correlations for both tests could be studied in future years' data.
Conclusion: If hospital record studies were to show that these observations are confirmed for many patients, urologists should be informed of this correlation possibility. They might then choose non-invasive procedures like ultrasound and MRI before deciding on a biopsy based on a rise in PSA score alone or will uniformly arrange for repeated PSA measurements to rule out the fluctuation.