PSA projection model?

[Chugach]

is there a mathematical model that is used to estimate trajectory of PSA rise? Certainly doubling rate is a consideration easy enough to calculate, but at what point do these rates become non-linear?

My last 3-yrs on Keytruda have been great - life pretty much back to normal. Yr 1undetectable PSA; Yr 2-3 slow creep from 0 to 1

I’m starting Yr 4 and trying to estimate how long I have before this reaches ~10 and then 20 at that point lutetium is the next (last?) recommended arrow in the quiver.

How much longer can I expect to ride the Keytruda Train?

[Tall_Allen]

PSADT is non-linear. You need at least 3 values over 0.1 to fit the curve. But the error of estimate increases with time. It's pretty useless for finding the kind of projection you want.

[Chugach in reply to Tall_Allen]

Thanks Allen - I’ll give it a shot. I know it’s non-linear but at low values over 2 years mine has appeared overall linear. It just seems I’m about at the inflection point where it starts to switch to exponential.. (or at least that’s my fear). As always thank you!!

[dac500]

I have been pretty successful in predicting in what narrow range my next PSA reading would be during the last four years on iADT for my metastatic prostate cancer. But recently things changed. I was expecting my PSA on January 23 would be in the range 0.15 to 0.20. But it was actually 0.32 slightly up from 0.29 three months ago with castrate level testosterone. Honestly speaking it makes little sense trying to fit PSA to a mathematical formula.

[Poowater]

The normal doubling time is once a year. Mine was doubling every 2 weeks ie aggressive cancer. Now I'm on 2nd treatment Pluvicto. That has brought psa down by 59% in a month.

[cesces]

I recollect there was some radiation oncologist who had a PSA algorithm he used for another purpose.

He would use it to calculate the doubling rate.

People take PSA tests whenever. So it is hard to calculate a doubling rate.

I guess this , or a similar, algorithm could do what you want.

Though as Tall Allen says, it will be of limited utility.

Still it might be comforting.

Actually I bet there are a number of people here who could whip up something like on a shareable Google spreadsheet in minutes.

[Boywonder56 in reply to cesces]

Your in a room full of castrated men....we cant " whip" up anything. BW

[Justfor_]

If your detectable (undetectable don't count) time series comprises 10 or more 2 decimal places samples, I could crunch the numbers for you. Yet, I can tell you that the confidence level of such a projection into the future will be +/- some years, if that finally serves any purpose for you.

[kreg001]

You may find this interesting. I’d suggest looking for papers that cite it as well.

ncbi.nlm.nih.gov/pmc/articl...

[Chugach in reply to kreg001]

this is great- thanks!

[StayingOptimistic]

it’s never linear and no one can predict. This disease has a mind of its own.

[MateoBeach]

An inflection to steeper rising will alert you that resistance to the current regimen, including Keytruda, is emerging and it’s time to reevaluate (scans, liquid biopsy, and ?). There are more arrows in the quiver than just one. BAT combined with Olaparib for one example. Look through available trials and emerging research. That is what we do here. We try to stay a step or two ahead.

[Chugach in reply to MateoBeach]

thanks Mateo!! I’ve been through a lot (including a highT clinical trial) and all the normal SOC, but I appreciate the reminder to never quit looking for arrows!

[Stoneartist]

I use a simple PSA vs time plot with the PSA plotted on a log scale in excel (see my profile). It works for me and I have been predicting my next value with great success - except for one outburst of Castrate resistance which showed up like a beacon on the curve. Next PSA test in 10 days - expecting around 0.07 - but there will be a error range due to that I am getting near the sensitivity of the test

[NecessarilySo]

When a cell divides, it doubles. The number of cells, x, equals 2n, where n is the number of divisions.: X=2n. that is not linear, but charts to a curve, parabolic. 1, 2, 4, 8, 16, 32, 64, ...etc. Cancer beats us because it forgets how to die like normal cells. It can be slowed down, and it can be killed selectively.

[TomMaloney]

No model possible.PSA goes down, until it goes up.

[Boywonder56 in reply to TomMaloney]

"See jr. he thinks of the simple things first".........name that movie..