We asked 43,000 men to find out.
The most frequent question we are asked by people who have been to a GP and got their Prostate-Specific Antigen (PSA) blood test done is: “Is my PSA normal?”. Often men do not get a sufficient explanation as to why they are or why they are not going on for further testing, and so they turn to us.
This question is tough to answer in general. In fact, we built a company around answering this question. Many factors make up someone’s risk of developing clinically significant prostate cancer. Our doctors and Artificial Intelligence (AI) consider many more factors than merely one PSA value. Thus, this post does not constitute medical advice, but the numbers in our database tell a story.
What makes this tough to answer is — what is normal? What is normal for him is not the same for you. What is normal for someone with a family history of prostate cancer is not the same for someone without a family history. Before we can say someone’s result is healthy or not normal, we need to know if it is normal, for them. The question now changes from: is my PSA normal to what a normal PSA is for me?
What is a normal PSA … for me?
To answer this question, we need to understand two things:
Firstly, knowing the age, family history status (one or more direct blood relative with a positive diagnosis), and the results of a recent PSA test, what is the patient’s risk?
Secondly, knowing the patient’s risk, is this an acceptable risk level?
Before discussing the findings, we would like to clarify:
- There are several reasons why someone’s PSA might be higher or lower than a certain level. Some are benign reasons; some are serious reasons. To uncover the truth requires proper testing protocols such as ours at Maxwell Plus.
- A single PSA test is rarely enough to form a conclusive opinion; therefore, our members have blood tests and MRIs before discussing biopsies. Our view is that if a result places someone in a higher risk category, we recommend further testing to provide more information. The next step should rarely be a biopsy if your PSA is elevated.
What is the patient’s risk?
To begin with, we need to define risk. In this situation, we define risk as the chance of a positive diagnosis of cancer in the next 12 months. Refer to the section below Notes on Data for more information
To provide an overview of risk, we grouped our patients into 5-year age categories and PSAs into 1.5 ug/ml categories, giving us a “cell” for each combination. The PSA value is the value that the man had within 12 months of a biopsy. More information in the section Notes on Data. Table 1 shows a breakdown of how many people fit into each of these cells.
|PSA (ng/mL)||0 – 1.49||1.5 – 2.99||3 – 4.49||4.5 – 5.99||6 – 7.49||7.5 – 8.99||9 or greater|
|Age||55 – 59||3,237||1,437||645||343||165||79||122|
|60 – 64||6,183||3,006||1,583||1,040||580||251||370|
|65 – 69||5,261||3,177||1,632||1,230||637||329||526|
|70 – 74||3,065||2,087||1,113||900||548||334||479|
|75 – 79||1,031||663||330||272||196||131||239|
Table 1: Distribution of patients. Total Cohort is equal to 43,221
Our first step in the investigation was to look at the proportion of men whose biopsy returned positive results. By determining the percentage of men in this cell with a positive biopsy, we can begin to assess each cell’s risk.
The number of men used in this category does not match the total men is because we removed cells that had less than 75 people in it.
Table 2 below shows these results. What is shows is as men get older, and their PSA rises, their chance of being diagnosed with clinically significant prostate cancer in the next 12 months increases. Nothing new.
Proportion that received a positive diagnosis within 12 months
|PSA (ng/mL)||0 – 1.49||1.5 – 2.99||3 – 4.49||4.5 – 5.99||6 – 7.49||7.5 – 8.99||9 or greater|
|Age||55 – 59||0.5%||2.3%||7.8%||22.2%||21.2%||30.4%||43.4%|
|60 – 64||1.3%||4.0%||9.3%||24.1%||26.0%||22.7%||38.6%|
|65 – 69||2.3%||5.3%||14.0%||33.0%||36.6%||35.6%||50.4%|
|70 – 74||3.5%||6.8%||17.5%||34.8%||40.7%||41.9%||56.6%|
|75 – 79||4.4%||9.0%||20.6%||44.5%||49.0%||51.1%||70.3%|
Table 2: Distribution of patients with a positive diagnosis.
The exciting part of this analysis is not the specific value of the cell but rather how much a man’s risk changes with slight changes to either their age or PSA levels.
For seemingly small changes in age and PSA levels — the risk of being diagnosed with clinically significant prostate cancer in the next 12 months can increase rapidly.
Here is an example of changes in risk as PSA and age raise. Take three different men, two that are 62 years old, one that is 67 years old. All with no Family History.
Let’s call them Fred, John, and Bob.
Three similar men, now let’s compare their risk of being diagnosed with prostate cancer in the next 3 months. By look at their risk according to the value of their cell, we can see the have markedly different risks.
Figure 1: Comparisons in risk for three similar men, with small changes to their age and PSA levels.
What is astounding here is that John’s risk is more than 2x higher than Fred’s. On top of this, Bob’s risk is nearly double John’s again. Small changes in age and PSA levels, large increases in risk.
What about Family History?
What if these three men were exactly the same, but now they all had a family history? Once again, we see quite a significant increase in risk.
Figure 2: Changes to the risk of men when family history is present.
Having a family history of prostate cancer compounds the risk even more, placing individuals at an even higher risk.
It is already well understood how risk increases with age, family history, and rising PSA levels. However, the amount the risk changes is an astonishing fact to learn for many of our patients. For seemingly innocuous changes in PSA levels over a few years, an individual’s risk of developing cancer can increase 2-4-fold.
Is this an acceptable risk level?
So, what is an acceptable risk for an individual? Is 2% ok but 3% is not? Does a 55-year-old with a PSA of 3.001 have a 3.4x increased risk than a 55-year-old with a PSA of 2.9999? Now the question of what is normal becomes very tough.
More in-depth analysis is required to understand what the patient’s risk truly is, and to know where it could be going.
One way to do this is by taking smaller segments, say 0.1ng/ml rather than 1.0ng/ml. Or using single years instead of using five years. These smaller cells form a far much more accurate picture of an individuals risk.
Now, what happens if we added in prior testing results as well to see what the rate of change of the PSA is? Or we added in additional blood markers that are available such as the Free to Total Ratio of the PSA test? Or what if we distinguished the difference between the number of family members diagnosed?
This analysis is what our Artificial Intelligence does best. It considers many more factors and at a much more segmented level. This analysis provides a more granular view of the risk of the patient than Table 2.
Having a more granular view also allows us to understand what could happen with that risk if things change. For example, we can ask questions such as “what if the PSA continued this trend over the next 12 months?” and determine whether the potential increase in risk outweighs the burden of another blood test.
Taking all the information available, our AI determines a risk score for the patient and provides this to our clinicians. Our clinicians take this result and use it to help decide “Should I send this patient for a follow-up test, or should I see them again in a year?”.
On top of this analysis, we have many advanced, non-invasive tests available to help us decide whether a patient should need further tests such as a Prostate Health Index blood test or a Multi-Parametric MRI. These provide even more high-quality information that can significantly help in making sure only those patients who need to get a biopsy receive one. All our diagnosed patients receive at least three blood tests and an MRI before a discussion about a biopsy occurs.
Understanding whether a patient’s risk is acceptable requires a much more detailed analysis and understanding of other factors. This work is our focus and what our AI and doctors do every day.
Is my PSA normal is a question that asks for a yes/no answer to a problem that is influenced by many factors. What we have shown here is by changing the question to “is my PSA normal for me?” allows for a range of essential elements to be considered by the clinician. These factors are individual to the person, and only this alone can determine whether the risk is acceptable for the patient.
If you would like to discuss your results with our doctors – please register here.
A quick note on our data
Our analysis combined our data sets, which include the PCPT and PCLO datasets, as well as several smaller datasets that we have collected. Each patient has their last PSA test within 12 months of a confirmatory biopsy. The biopsy determined the presence of significant cancer. This combined data set is unique because it is a population risk study, as every man received a biopsy. It is not a study of high-risk men or men who already have prostate cancer, which allows us to have a much more applicable view of risk for the average man. It also provides a realistic view of population-level risk, not just the population that decided to get tested.
n = 43,321
Positive cancer diagnosis = 4,793
Age range = 28 to 94, mean = 65.3, median =65.2
PSA range = 0 – 2100, mean – 3.45, median = 2.0
Eckersberger E, Finkelstein J, Sadri H, et al. Screening for Prostate Cancer: A Review of the ERSPC and PLCO Trials. Rev Urol. 2009;11(3):127-133.
Ian M. Thompson, Jr., M.D., Phyllis J. Goodman et al. Long-Term Survival of Participants in the Prostate Cancer Prevention Trial. N Engl J Med 2013; 369:603-610
Ian M. Thompson, M.D., Donna K. Pauler, Ph.D. et al. Prevalence of Prostate Cancer among Men with a Prostate-Specific Antigen Level ≤4.0 ng per Milliliter. N Engl J Med 2004; 350:2239-2246