Posted 9/23/2023 12:52 AM (GMT 0)
Here is my understand of ultrasensitive PSA testing.
There is uncertainty in the right-most digit of a lab test. There are several reasons:
Extrinsic factors:
2) Tests often don't always return the exact same value when repeated run on the same drawn sample: statistically a Standard Deviation is calculated: plus or minus some value. If you apply 2 S.D. around the mean value, you usually cover about 98% of all lab values. For instance, an independent lab calculated that the ROCHE assay that Labcorp uses for its 3-decimal ultrasensitive PSA test is ±0.001 when testing a prepared exact concentration of 0.01. That means that if I get back a result of, say, 0.013, I can presume that if the test were run 20 times on the one sample I gave, almost all the result should fall in 0.011 to 0.015 range. However, I should not expect to get 20 results that all show 0.013.
(If I remember the S.D. for PSA concentrations around 0.1 was a bit higher, which means even more variation in results!)
3) Sample/reagent prep, equipment calibration, etc. Point (1) holds for ideal lab conditions, which aren't the norm, so any deviation from ideal in sample preparation or equipment calibration can affect test results. Humans and equipment aren't perfect. Sometime you even get an "outlier" -- a test result that seems wrong and needs to be repeated. And BTW any unexpected rise in PSA merits a retest, as most uros will do, whether immediately or after an interval.
4) Very low test values that are close to a test's lower limit of detection run into the Limit of Blank (LoB) issue. It may seem strange, but many tests that are run on a solution with 0 amount of substance will not return zero! For example, when Labcorp's uPSA test is run on blood with no PSA, the result could be something like 0.001 to 0.005 (if I remember correctly). This is called the Limit of Blank (since your testing a blank--a solution where the concentration of analyte is known to be zero). The lowest value Labcorp reports for this test is 0.006, which is just above the LoB and I believe very low test results could be somewhat inaccurate because of the uncertainty/contribution of the LoB. If the equipment shows a lower value, the lab reports it's "undetectable", the Lower Limit of Detection with a < sign. In this case you'd get back <0.006 from Labcorp and not, say 0,005, because that 0.005 isn't reliable and may not be PSA at all.
Intrinsic/Phsyiological Factors:
We probably don't maintain a constant concentration of PSA in our blood. I don't know if it fluctuates hour to hour or day. I believe when PSA was first discovered, they did daily measurements on men, but I forget the daily variation-- it may have been 10 or 20%. Keep in mind that was for PSA for men with healthy prostates, and that 1-4 (approx) range is a far cry from the extremely tiny amounts of PSA we're talking about after prostate removal.
Perhaps there is a tiny fluctuation due to one's state of hydration. Perhaps one's state of health (e.g. infection, inflammation, etc.) can affect the minute amounts of PSA produced by non-prostatic tissue, such as the urethra and Cowper's glands.
The bottom line is that I suggest you follow your PSA with a test that has one digit more than the digit you are monitoring. For example, if you round a 3-decimal test to 2 decimal places, you get a more reliable result that if you use a 2-decimal test and take it on face value. Very often the exact value is of lesser importance: what counts is whether your PSA is essentially stable or increasing, and if increasing, the rate of rise (or doubling time).
When you have a single significant digit in a result, for example 0.02, you can't put too much faith in the accuracy of that digit. 0.02 could reflect an actual PSA of anything from 0.015 to 0.024 from rounding alone. If your actual value the day of the test is, say, 0.024, variation in the test assay itself--point (1)-- could very easily lead to a lab slip with 0.03: you may worry about "jumping" from 0.02 to 0.03, but your PSA may very well be stable.
I've been following my PSA with Labcorp's uPSA and even with 3-decimal results, it has taken quite a number of data points to establish both a reliable value and a trend. The absolute PSA value is often less importance than knowing whether one's PSA is stable or rising, and if rising, the rate of rise (or doubling time). As mentioned in a post above, doubling times are not reliable or important based on extremely low values.
___________
A reminder: when reporting decimal numbers below 1 (e.g. uPSA results), always put a leading zero in front of the decimal place. If you write, say, .02 we have no way of knowing if (1) you forgot the leading zero and meant to type 0.02 or (2) you did type the leading zero, but misplaced the decimal point and meant to type 0.2. Thanks ;)
Djin