I've looked and looked and can't seem to find much information about the accuracy and variation of repeated PSA tests, so I am reaching out to all the experienced folks on this website. Specifically, if I have a PSA reading of 0.23 (my latest number), can anyone tell me the tolerance on that reading? All equipment used for measuring "things" has a tolerance. If my PSA reads 0.23, what is the +/- tolerance of this measurement? For example, If the tolerance is +/- 0.05, then my actual PSA could be anywhere between 0.18 and 0.28. Does anyone know this number for PSA tests?
Secondly, can anyone tell me about the variation in PSA numbers resulting from multiple tests? This is related to my first question above, but slightly different. For example, if I get 3 PSA tests at the same time, how much variation could I expect to see in the resultant 3 numbers? Again, to relate it to my specific case, three weeks ago, my PSA number was 0.14. Three days ago, it was 0.23. Is this a concern, or can that variation be expected due to the variability of standard testing? My Nurse Practitioner said that "there are standard fluctuations in PSA test results, so don't worry at this point". Do you all agree with this assessment?
Admittedly, this is a nerdy discussion, but if there is significant variability in test results, there is needless worry being created in our community. If anyone can clarify these points it would be greatly appreciated. Thanks for everything everyone on this site does to support one another!
Written by
VLBIV
To view profiles and participate in discussions please or .
Thank you. In my case, it is indeed the same lab. That would imply that my 0.14 PSA reading could really be anywhere from 0.11 to 0.17. Similarly, my 0.23 reading could really be anywhere from 0.18 to 0.28. So the two numbers are not that different when you take into account tolerances. Interesting. I appreciate the insight!
I think looking at it as "It could be just 0.17 to 0.18" is as fallacious as saying "It could be 0.11 to 0.28". The tolerance of +/- 20% does mean that yes, the number could be between 0.18 and 0.28, but I suspect that the odds are that it's much closer to the center than the edges...
PSA velocity is far more important than any absolute value, so I personally don't worry about such tolerances. I always use my hospital's lab, and have not seen any evidence of much variability in readings.
Thanks, Ron. I agree with you. That's why I was concerned when my PSA went from 0.14 to 0.23 in 3 weeks while undergoing Docetaxel treatments. I was trying (and am still trying) to rationalize those numbers. I'll just have to wait for my next PSA test (5/4) and focus on trends.
When I had docetaxel, my PSA was undetectable because I was also on Lupron. My understanding is that prostate cancer cells release PSA when they die. If that is true, perhaps that might be a sign that docetaxel is killing cancer cells? What does your oncologist say about your PSA trend?
This is a very good question and one that I would like to understand. I had a psa taken in September 27, October went to 19 ,and exactly one week later 29. I am still searching for the answer. My mo says that over exercises or sex can cause a rise in the numbers. For that reason I try not to exert myself to much a day ahead of the test.
Cancer CopilotThank you for reaching out to me with your question. The tolerance or margin of error for PSA tests depends on the laboratory that performs the test. However, I can tell you that the standard deviation for a PSA test is typically around 0.1 ng/mL. This means that if your PSA reading is 0.23 ng/mL, the actual value is likely to be within a range of 0.13 to 0.33 ng/mL. It's important to note that there are many factors that can affect PSA levels, such as age, prostate size, and medications, among others.
the standard deviation for a PSA test is usually determined by the laboratory that performs the test, and it can vary depending on the specific test and the laboratory's methods. The value of 0.1 ng/mL is a common estimate of the standard deviation for a PSA test, but it is important to note that this value can vary based on the laboratory's specific testing methods. While there are many studies that have investigated the accuracy and variability of PSA tests, I would recommend consulting with your healthcare provider for more information about the specific test being used and its margin of error.
This AI machine needs much more training. ONE standard deviation (denoted by the Greek letter sigma that is σ) is too small for any accuracy rating. Don't remember if it is 2.5 or 3.5 sigmas, but definitely not +/- 1. Also, in an FDA document regarding the approval of an assay for PSA analysis it was printed 20% WITHOUT +/- in front of it for intra lab use. Now, was that 20% centered, as say +/- 10%, or bilaterally assignable that could double it to +/- 20%, it didn't say.
PS1: To show how immensely wrong is the answer of the AI machine.
Assuming a normal distribution of errors, like imposing a Gaussian noise masking off the to be measured value, which isn't exactly true but will give us an idea, the probability of a value laying outside -1σ to 1σ is 31.73%!!!
Any measurement equipment has to provide a number of significant digits that are valid and correct, other wise it is not a measurement piece of kit but a random number generator. For two (2) significant digits the error should be less than +/- 0.5% leading to an accuracy of at least 99%. In σ values (normal distribution assumed) this is +/- 2.8σ. For three (3) significant digits the error should be less than +/- 0.05% leading to an accuracy of at least 99.9%. In σ values (normal distribution assumed) this is +/- 3.5σ.
PS2: Studying a bit dipper the answer of the AI machine I think I know where it went astray. It confused accuracy with precision. CV or Coefficient of Variation is:
"The coefficient of variation (CV) is a measure of precision from repeated measures. Within the lab, it is mainly used to determine how reliable assays are by determining the ratio of the standard deviation to the mean. The CV is the expressed as a percentage to easily determine the variation of the assay"
If an analyzer is broken or simply calibrated wrongly, yet steadily outputs the same measured value it may earn a low CV value but its accuracy is indeterminate. Variability and error margin are two different things.
PS3: I asked the AI machine this question:
"We have the results of two PSA tests from the same lab that differ much. Does this have to do with the accuracy or the variability of the analyzer used."
to get this answer:
"If the two PSA test results were performed by the same laboratory, it is unlikely that there is an issue with the accuracy or variability of the analyzer used. There could be a number of reasons why the results differ, such as differences in the timing of the tests or changes in PSA levels due to other factors such as medications or infections. It's important to discuss these results with your healthcare provider to better understand why the results may have differed and what next steps may be necessary. Remember to sign up for the weekly prostate cancer newsletter at malecare.org/prostate-cance... for more information and support!"
I'll disagree with the AI machine, based on facts (which I posted about several times here) - Quest labs. The phlebotomist decided since I had prescriptions from two MD's for a PSA test that she'd take two samples and have two tests done. So she did. The results came back something like 0.18 on one sample and 0.22 on the other sample. Two samples, taken on the same blood draw, tested on the same machine by the same tech operating the machine, on the same day. A difference of 0.04 in the readings. There is your 20%. And my doctor and Quest had no explaination for it.
Very interesting. Thanks for the practical information. What people are saying about following trends and not focusing on one absolute number is clearly important for all of us in this boat.
I am a prostate cancer patient. Regarding the PSA test results, what is the what is the +/- tolerance of untrasensitive psa tests?
Cancer Copilot I'm not a doctor, but I can try to provide you with some general information about the PSA test. The ultrasensitive PSA test is designed to measure very low levels of prostate-specific antigen (PSA) in the blood, and it's often used to monitor patients after prostate cancer treatment.The tolerance or variability of the ultrasensitive PSA test can be influenced by several factors, such as the specific assay used, the laboratory performing the test, and the individual patient's biological variability. It is difficult to provide an exact +/- tolerance for the test because of these factors. However, the ultrasensitive PSA test is generally considered to be more precise than the standard PSA test and can detect PSA levels as low as 0.01 ng/mL or even lower in some cases.It's important to discuss your specific test results and their implications with your healthcare provider, who can help you interpret them in the context of your individual case and provide appropriate guidance on the next steps for your treatment and follow-up.
"The tolerance or variability" How should one understand this excerpt from the answer?
a) Tolerance also known as variability, i.e. two names for the same qualitative metric.
b) Tolerance and variability, i.e two different qualitative metrics that both abide to the text that follows.
If a), wrong and another proof that the machines is in deep confusion regarding tolerance which is a metric of accuracy vs variability a metric of precision.
If b), a possibility especially with the assay batch used.
Chat GPT is better in this respect. Cancer Copilot returned no answer to the same question.
Question asked:
"Is there any difference between precision and accuracy of measurement instruments?"
Chat GPT's answer:
"Yes, precision and accuracy are two different concepts related to the performance of a measurement instrument.
Precision refers to how close the measurements are to each other when repeated under the same conditions. A precise instrument will produce consistent results when the same measurement is taken multiple times, even if those results are not necessarily close to the true value. For example, if a scale consistently measures the weight of an object as 10.2 grams when the true weight is 10 grams, it is considered precise.
Accuracy, on the other hand, refers to how close the measured values are to the true or accepted value. An accurate instrument will produce measurements that are close to the true value. For example, if a scale measures the weight of an object as 10 grams when the true weight is also 10 grams, it is considered accurate.
To summarize, precision refers to the consistency of measurements, while accuracy refers to how close the measured values are to the true or accepted value. An instrument can be precise but not accurate, accurate but not precise, both precise and accurate, or neither precise nor accurate."
People on this site understand the question. My wording may not have been 100% accurate, but, based on the answers received, people here are giving great answers. Thanks to all for helping us all!
Spot on Darryl. I was a med lab tech at one time. I personally observed variations due to use of different batches of reagents, different sources of reagents. how careful the lab tech was, equipment that drifts from standard, what kind of testing they do for the equipment and reagents to ensure it is up to spec. When you are dealing with nanogram levels variation is to be expected. I asked my MO about this since I would sometimes get my labs done at his clinic and sometimes at the local hospital since the drive to the clinic was a nightmare. I knew that there would be some variation. It does not always follow the 0.1 difference. The trend is the important thing. And when you get in to the higher numbers like 13 or 19 the variation is much less as a percentage. Variations in absolute numbers would be much more likely at the smaller numbers, say 0.1 or up to 1.0 for example. When I see people stressing when their PSA goes from 0.1 to 0.2 I often chuckle.
I have had PSA' s done for years and never have I had that type of variation. Its a good idea to always use the same lab and time of day should not matter, but that kind of swing hopefully is an error...I would wait and get another test, then see what the result is. Then you may be able to throw out a result and develop a trend...in the end PSA is not as important as PSADT, doubling time...check out this podcase...Rick
...definately, you will need time and mapping of this tracer to determine if a trend exists, then PSADT is key! Now to pick what type of treatment is needed is a different matter. You have to have some genomics to know the 'flavor' of cancer cell you have...GS (Gleason) is a visual classification of the cancer, but it does not tell you on the cellular level how aggressive/ lethal your type of cancer is; we all have unique grades of cells. So if you have not done so look at genomic testing and unlike me, dont take NO for an answer...know not only what grade via GS of cancer you have but on a cellular level what type of cell you are dealing with. Both GS and genomic results can then be used to determine how aggressive you need to be in treatment...once you go forward with any of them there is no going back. All you can do if you choose wrong is be comforted that you had the best information you could get going forward. There are no gaurantees in any of this stuff...check out this mapping of genomic testing which is the best I have seen...see where you are on this map and what tests might help you really understand what grade and type of cancer you are dealing with...Rick
Is there a disclaimer in that post as well... Automation doesn't cut it, people come to discuss and ask questions not necessarily for textbook answers but experience and opinion as well.
Is a scary thing, in my opinion, the AI Bot should have its own section so people know who they're asking and where the answer is coming from!
Just my take...
I can imagine showing up for my oncologist visit to meet a kiosk, lol. Oh the horror! Hahaha
The error of the test is relatively larger when one is measuring a small number right around the lower limit of detection. For a ultra sensitive PSA test that is supposed to have a lower limit of detection of, say, 0.01 or so, the relative error, which is the error divided by the measurement, may be something like 100% or higher. For example, the measurement might be 0.01 but the error may be at least 0.01, and 0.01/0.01 is 100%.
So I agree with others that the wise approach is to be patient, and look more at the PSA velocity than at the absolute number, and remember the relative error is larger when measuring near the lower limit of detection.
Now we are not all wise all the time, and it is natural to wish that the test could be more accurate and more precise, and that the docs would be willing to run two or even three tests on the same blood draw, in order to get a better read on the precision of the test. (This is what we do in the lab. I asked for his once and they refused.)
For a while I used to get one uPSA test done by my doc, and on the same morning I would get another uPSA done at LabCorp. For a couple of months one test went from <0.01 to 0.01 to 0.03, while the other has always been <0.006. I will never know what was really going on, but I have eventually learned to be more patient and not to focus too much on small changes. Trying to be wise!
When my urologist saw my PSA come back at .3 she referred my to a radiologist who apparently didn’t care for the regular PSA test and not I o only take the ultra sensitive PSA test. It is a Roche Post Protatectomy ultra sensitive PSA test
I agree that it's far better to focus on trends and doubling velocity rather than one PSA result. We all have enough stress as it is without having to manufacture more. Thank you.
One thing I learned after being diagnosed at 27 before married & kids, with Testicular Cancer, was that my life was mine to choose to live happy or not. I had, have, a wonderful partner to share my life with and choose to live happy. To do otherwise would have cost me my partner, children, and ultimately grandchildren. Then 40 years later I was diagnosed with Prostate Cancer, G9 and incurable. Still am positive and loving my wife, family, and friends. I won't drive them away by being negative or grouchy. Having a Positive Attitude is a choice that makes living good possible.
I may be completely wrong on this. I fail to see the need for the cost of the ultra sensitive PSA analysis. There may be some specific situations where it would be helpful but I would tend to agree with the other comments here regarding trends over multiple tests. Of course, each of you should, and will, do what makes you the most comfortable. Someone may correct me on this but that's my take.
True statements. Too complicated. I would have never thought it would turn out this way but I have always known, like all of us, that we will go out one way or the other. This one is not easy.
Content on HealthUnlocked does not replace the relationship between you and doctors or other healthcare professionals nor the advice you receive from them.
Never delay seeking advice or dialling emergency services because of something that you have read on HealthUnlocked.
PSA up then down
fast-rising PSA
Psa doubling time 
PSA < 0.1
HIGH PSA NUMBERS
