Next chapter in the Labcorp <0.006 vs <0.014 saga

Many of us (me included) have found that undetectable results of <0.006 (and low readings a bit higher) are now being reported by Labcorp as <0.014. I got no reply when I asked Labcorp about this through their patient portal. This is from a ROCHE spec sheet (updated 28.10.2020) for their PSA assay, which Labcorp uses:

Detection Limits
LoB 0.006 ng/mL, LoD 0.010 ng/mL, LoQ 0.014 ng/mL (cobas e 411 analyzer, cobas e 601, cobas e 602 module)
LoB 0.006 ng/mL, LoD 0.014 ng/mL, LoQ 0.030 ng/mL (cobas e 801 module)

Measuring Range
0.006 - 100 ng/mL (cobas e 411 analyzer, cobas e 601, cobas e 602 modules)
0.006 - 100 ng/mL (cobas e 801 module)

[Emphasis mine].

Forum brother TJ123 posted to a PSA thread last month:

TJ123 said...
Djin,
All my Labcorp PSA tests were taken at the very same clinic and the samples sent to the very same regional lab for analysis. So nothing changed except for the lower limit on my very last Labcorp test in May.

I called the Labcorp regional lab for an explanation. I was told the manufacturer changed the lower limit for their test.

Since I got no answer from Labcorp despite repeated attempts, I decided to email ROCHE, who replied yesterday:

ROCHE said...

The lower measuring range claim for the total PSA assay is 0.014 ng/mL defined by the Limit of Detection for all instruments running the test. The Limit of Blank is 0.006, which can be measured by the cobas e 411, e 601 and e 602. The Limit of Blank cannot [be] measured by the cobas e 801. It may be helpful to contact the laboratory to determine the instrument and convention used for reporting at the lower measuring limit.

[Emphasis mine]

It turns out that when you repeatedly run blanks with this test's reagents and no PSA at all, you don't get back results of nothing or zero, but rather some number that can be as high as 0.006. This is defined as the Limit of Blank (LoB). Clearly, you can't detect/report actual patient values in this range. The lowest value above the LoB you can reliably distinguish from the LoB is called the Limit of Detection (LoD).

For definitions and excellent explanations of LoB, LoD, and LoQ (Limit of Quantitation), see the text and figure in

Limit of Blank, Limit of Detection and Limit of Quantitation (2008, Full Text)

I had thought that the change to <0.014 was because uPSA tests were being run on less sensitive instruments, but I now think that is not the case; rather, someone, either at Labcorp or ROCHE, probably questioned using <0.006 and Labcorp realized (or was told) they should be using the assay's LoD (or LoQ), not the LoB, for reporting "undetectable" results. (I found a much older ROCHE spec sheet for this assay, and the numbers were the same, so nothing changed with this assay at ROCHE's end as far as I can determine.) Note that <0.014 used to be Labcorp's undetectable for uPSA until a few years ago, when they lowered it to <0.006.

Djin

mattam said...
So, is it really ever worthwhile to run a PSA assay below a sensitivity of 0.1? Maybe for pulling the SRT trigger, but that's even debatable.

The sensitive tests do a dammed good job of generating anxiety. I've considered asking my MO to go to a 0.1 cutoff.

From my string of Labcorp's uPSA results and those of others, I believe their resilts are reasonably precise and reliable. As I have mentioned, I am the only patient in my uro's practice using (at my request) this test. For his other patients my uro uses an in-office Labcorp instument that detects to 0.05. He told me he gets plenty of lead time with it before a patient reaches actionable BCR.

I think it all depends on (1) how likely BCR is for the patient and how much lead time they would want to plan any treatment. (2) For patients with intermediate- or high-risk PCa electing surgery, an accurate post-op nadir can be a prognosticator of BCR risk, again, if you are interested in this. And (3) we can differ in our comfort level regarding PSA testing and anxiety. A more precise test can be useful for those who barely escaped needing adjuvant therapy, but who want to pull the trigger on salvage therapy well below 0.2. Lastly, (4) for high-risk men or those with adverse findings post-op, <0.1 may not be sensitive enough to reveal persistent PSA. A post op PSA of, say, 0.09 may not meet a technical definition of persistent PSA, but I would be worried.

In my case, since I know my PSA nadir was very low and has remained so for 3 years, I can't see why my PSA would start to go up other than because of recurrence; so I want to know of a rise that goes beyond a slight fluctuation. On the other hand, a G6 guy with no adverse finding probably has little interest or reason not to be happy with a continuous string of <0.1.

Djin

Thanks for getting to the bottom of it, Djin.

I couldn't handle not getting a uPSA test, but I'm blessed with no anxiety before a blood draw. A big PSA jump could change that.

The more info, the better, allowing for an accurate trend line, and more time to choose a course of action. Like you, I'm the only patient in my docs practice that gets a uPSA.

....and so it goes.

Thanks for the detective work!

I wonder if these guys realize the angst they cause?l

DjinTonic said...
I'm not sure I did get to the bottom of it, Gary, but for now at least, my theory seems to fit the information we have.

Maybe not, but closer than anyone else has gotten. Your research ability is remarkable, Jared. Thanks for all you do.

Gary

Halbert, for some men <0.1 isn't sensitive enough. See also my post above.

Some Forum brothers have indeed pulled the trigger on SRT at values below 0.1. In order to know when to make this call with some confidence, you need to know (1) at what value your PSA rise began; (2) how long after surgery did the rise start; and (3) the velocity, or doubling time, of the rise -- although doubling times at very low values may be of questionable value. That information, together with other factors, e.g.,pathology and G score, goes into the decision pot.

In my case, after a RP for G9 (4+5), it was essential to know the complete pathology picture as well as my PSA status. Many G8-10 men will need adjuvant therapy after surgery. Fortunately, I had no adverse findings. However, a further condition for avoiding ART is that there be no persistent PSA. Even though many studies use 0.1 and above to mark persistent PSA after surgery and my first PSA at 5 weeks post was <0.1 on a standard PSA test, my uro uses a <0.05 test after that (I requested we use the outside Labcore uPSA test). My first uPSA test at 3 months was 0.010. Had it been, say, 0.08, that would not bode well at all. Nor would a rapid rise from, say, <0.05 to 0.08. I'm sure my uro would have referred me to an RO in either scenario.

It is true that post-op PSA dynamic have to be looked at from the risk level of the patient. Other risk groups have the luxury of waiting until 0.2 and much high to begin SRT.

The post-treatment PSA dynamics are, as we all know, very different for RP vs RT as primary treatment. A number of studies have shown that for RP men, their post-op PSA nadir is predictive of BCR risk. It's good to have one's nadir below 0.010 or (fortunately, mine was 0.007, but we now know that is flirting with the unreliable end of the uPSA assay and the change to <0.014 will take that knowledge away). Other studies have shown that being at 0.03 or below at 3 years is also a very good sign of remaining BCR-free.

I would venture that the many uros follow their high-risk patients with at least a 2-decimal test. Labcorp's uPSA is, as we know, 3-decimals, and Quest's is 2. Both companies tout their tests as being for post-RP men.

Back to the precision question. I do think that, in light of ROCHE's data, a LoD of 0.014 is very reasonable and is well below the 0.1 LoD of the standard PSA test.

I've posted my post-op PSA history before, but, if we assume for the moment that my PSA is fairly stable, it appears that the Labcorp test appears to be reliable, based also on the values reported by others. Also, the absolute value is less important than the trend, within some limits of course -- just like a bathroom scale that may not be very accurate, but is precise enough that you know reliably when your weight goes up or down.

(3 m ) 0.010
(6 m ) 0.009
(9 m ) 0.007 (nadir)
(1 yr.) 0.018 (huh?)
(13 m) 0.013 (retest)
(15 m) 0.012
(18 m) 0.015
(21 m) 0.015
(2 yr. ) 0.016
(28 m) 0.015
(34 m) <0.014

These values are well below my uro's office PSA test LoD of 0.05, but this uPSA test gives me more time to plan salvage treatment if my PSA starts to rise above these fairly small fluctuations. Hopefully, advanced PET scans will continue to improve in lesion detection at increasing lower PSA levels. That, coupled with uPSA testing may prove to be very important for some of us.

Djin

My uro said that uPSA remains controversial even today. In threads about BCR I sometimes have pointed out that when institutions were decuding on an offical BCR definition, JH proposed 0.4 before the current 0.2 was agreed upon because some men do plateau at these levels. With the advent if uPSA tests, the very low PSA values proposed in research papers are usually done so for their predictive use, i.e., who is more or less likely to encounter BCR in the next few years, and not to redefine BCR itself.

For men with high G disease, post-op PSA can be the decider when it comes to adjuvant treatment recommendation. The best argument I could probably make for 3 decimal places is that some men and their docs are following the 2nd decimal place, and as we know there is always uncertainty about the right-most decimal. So a 3-decimal test rounded to 2 places might be the best of both worlds. If I'm remembering correctly, our Pratoman opted for very early SRT at 0.08. Again, it is not necessarily any one reading, but it also depends where your nadir was. A initial 0.08 reading will probably be watched carefully; a 0.08 rising from, say, an initial 0.02 may be reason to set up an RO appointment.

Djin