Results from today's Med Onc visit (3/13/17)

Thanks in advance to all of you putting up with my musings. I always appreciate your thoughts!

So, today's PSA was 0.7, and testosterone was an amazing 544! Basically, a mixed outcome. It's not 0.5, which I would have seen as well-controlled end state from RT. It's up, so maybe something's still active? But I'm lovin' that testosterone number.

While the PSA's still up from the previous reading, it doesn't mean we need to do anything any time soon. We're going to continue watching for another 3 months, and check the PSA then. My MO isn't concerned at this level, wanting to watch the trend longer before we do anything else, like HT.

This increase made me feel a little deflated, until I decided to so some of my usual engineering-style coping and play with the numbers. You can probably hear my wife's eyes rolling, but this works for me! I even have a couple of graphs to look at, if anyone is interested. Offhand, I can only think of a couple guys who will be, but here it is anyway. :P

Anyone up for a little math? Come on, stay with me here! Is PSA growth geometric, or is it logarithmic? Most natural processes are one or the other. (We'll leave the Fibonacci series out for now, since nowhere do I see a curve fitting process to that one!).

Geometric (2nd order fit):
With 2 readings one can only get a line. With a third point one can calculate a 2nd order (parabolic) geometric curve fit. A linear projection from the 1st two yields 0.77 for today. So, if today's reading had been above 0.77, then it's a positive growth curve (accelerating, positive 2nd derivative). Since it was below 0.77, then it's a negative growth curve (decelerating, negative 2nd derivative). This curve peaks at around 0.89 late this year, and should just hover around that value for a long time. (yay!) (The projection continues down eventually, but I don't think reality would do that. More likely to just level out.)

Logarithmic (natural log fit):
The three points with a logarithmic curve fit give a rising projection in this curve, very nearly linear in appearance. That doesn't do anything exciting for a while either.

Online doubling time calculator:
I don't really get why this is so much more aggressive. I put the same three numbers into one, and it comes back with a 4 month doubling time! This seems to greatly disagree with either the geometric or natural log fit. This one worries me if it's realistic, but I just don't know at this point.

"Doubling Time" PSA Calculator

(Is anyone still here? I can't help myself, you know...)

Bottom line:
My MO is cool. He says, "Look, we're treating a disease, not a number. We need to see what it's really doing before taking any actions.". I like that. Not to be cavalier, not to ignore, not to delay. Rather, a form, it seems, of Active Surveillance, just in a new setting. The number isn't the point. He said for example, if the PSA rose to 10 over 10 years, it would be indolent and not worth worrying about. It's not the number, it's the pattern that indicates what's happening. He said he's seen many G9s and 10s that never progress out of control. Maybe just being consoling, but I get it.

We just have to wait anyway. I'm feeling pretty good tonight, after drilling through all of this. Waiting for this visit was honestly a bit nerve wracking, but for some reason this has actually calmed me down.

Pratoman said...
Redwing, all in all you got a reprieve. So it's good news. And it made you feel good. Isn't that all that really counts. (am I finally learning what "being in the moment" means? Doubtful, but I guess stranger things have happened.)

So regardless of what your engineering background no logarithmic charts tell you, I say, as the song goes, don't worry, be happy.

Deal with it in 3 months. Until then, just say Yay!!!
Looking forward to celebrating your good fortune in Baton Rouge.

Pratoman, you couldn't say it better! It does feel like a reprieve. It's going to be ok leaving it alone this time. See you in just a few weeks now!

Nice read first thing this morning. Celebrate. Got to be wonderful to be off the HT and have testosterone back.

thanks for posting - and the good-natured digressions -

Wilderness

Hello Redwing,

I believe Snuffy says PSA is done logarithmic. Good graphs but I was a BBA major. You know that PSA rises after ADT wears off and also typically after RT. Then, it may fall. I like your MO also. He knows his stuff. You did a good job selecting him. Your PSA is still low so glad that was calming for you. I know that anxious feeling every time I have to have one done....as we all do.

You had excellent treatment and will be fine!!

Robert
NC

Well Buddy I think this is a great approach. We will have lots to talk about in a few weeks at GFMPH.

See you then. I wish my T was 544!

Cajun Jeff

Thanks, Jack.

I played around a little more with this, and found how the online thing is calculating it, at least I'm quite close. It's a log10 fit.

Take the PSA data, then log10(PSA), plot and get a linear fit. Then, convert using that line fit equation to predict PSA by converting back (10^(log10(PSA fit)).

From that, there's another equation involving logs you can use to extract a doubling time. Doing that, my doubling times come out really close to the online tool. I got 3.7 months this way.

These projections indicate June's PSA should be 1.4, and hitting 2.0 by August 2017 (!) if I use all 3 data points. Or, June's PSA could be 1.0, and hitting 2.0 by January 2018 if I only use the 2 latest. Ain't math fun?

Well, beans! I like the limiting polynomial version better.

As one who has only vaguely understood what has been discussed here so far in this thread (as an undergrad I got only as far as College Algebra, an effort at that, on my way to a Humanities BA), I just wanted to say that I admire the effort here to put some mathematical meaning into these numbers, with whatever tools are available.

I know we have some really smart mathematician/engineer types here on the forum, and, who knows, maybe you who fit this bill may have the mathematical skills to make a significant discovery. (Maybe only a few medical researchers have the mathematical skills you guys have, and so are less likely to see something that some of you guys might see).

Good on you math types for trying! Who knows, maybe one of you will actually note something really important!

Redwing, Our positions are very similar. After PSA emerged once again in January '15, I have seen a steady climb to .57 as of last week. I had surgery in 2009 -- that failed in about 18 months. Did SRT + HT in winter of 2011-12 at Sloan-Kettering in New York City. After that PSA stayed undetectable until January '15. My MO is most concerned with doubling time from the point I breached the .20 barrier. He wants to see a doubling time above 10 months; less than that, and he would consider HT. My doubling time as of last week is about 15 months. In several discussions we reviewed the SE of HT, and in my MO's view these are not inconsiderable. I also told him that my preference is to "feel like me" for as long as possible. He was willing to listen to that, and I am appreciative. Meet with my MO in a few weeks to review where things stand. For now I am on a three month PSA cycle.

NewspaperLover
Age 73
Gleason 8
Surgery upstate NY 2009.
After surgery failed, moved management of my case to Memorial Sloan-Kettering in NY City.
HT + SRT 2011-12.
PSA undetectable until Jan '25. PSA .08.
PSA Jan '16 .26.
PSA March '17 .57.
No current treatment.

Paxton said...
Redwing - Does PSADT carry the same meaning after RT as it would before? I remember reading somewhere that it does not, especially at low PSA levels.

Paxton, you can read my comments above about using the calculation tools at low levels, (single significant figure), but several years after RT, and following the end of HT and 3 years of undetectable, PSA measurement is THE meaningful means of monitoring status.

Post Edited (JackH) : 3/14/2017 11:57:55 AM (GMT-6)

For what it is worth, simple doubling times are calculated using the natural logarithm (Log to the base 'e', or LN on excel spreadsheets). To understand how things grow / change in nature, google 'growth natural logarithm e' or something similar. The wikipedia page 'Doubling Time' is a good start to understand the principle involved - it specifically refers to tumour growth.

A simple formula of DT between two PSA readings only is:

DT = Time lapse * LN(2) / LN(end PSA point / start PSA point).

For example, if you have two readings that are say 4.5 months apart, the start PSA was 0.3, and the end PSA was 0.5:

DT = 4.5 * LN(2) / LN(0.5/0.3)
= 4.5 * 0.693 / LN(1.667)
= 4.5 * 0.693 / 0.511
= 4.5 * 1.357
= 6.1 months.

You can use any time unit (days / weeks, etc) so long as you are consistent. Here I used months for simplicity. Remember, LN is the excel function =LN(x)

<D

Nice picture showing the doubling time as a straight line illustrating the formula

http://www.intechopen.com/source/html/41807/media/image5.png

(can't make link clickable for some reson)

Redwing - you are right, because we multiply / divide the logs only, we can use either LN or LOG.

However, the very basis of the number 'e' (or Euler's constant as it is sometimes called) is continuously compounding growth, which is what tumour growth is - continuously. I find that knowing something about that added a bit to my understanding of this.

Thanks for the interest, I guess we will leave the math here, otherwise we may find ourselves in a fairly lonely dialogue here.

Best, <D

At the risk of overstaying my welcome on this, just one more graph. This rather summarizes the various inputs. I tweaked the last two actual readings by only 0.1 each (within PSA test repeatability), and they now fall on a very straight line. I'm even including a "<0.1" as a 0.1 from May. I hadn't included that one before because it was unclear how to do so, but amazingly it now drops right smack on the same straight line.

Natural Log Projection PSA vs. Time

The implication is since stopping HT in March 2016, and testosterone recovering to 132 by May, that the PSA was already starting the trend it's still on. Wow. The absolute log domain linearity is surprising. Saturation theory, saying once the androgen receptors on the cells are saturated with testosterone the cells can be fully active, seems to be supported by this. I see no evidence of prostatitis-induced change, no zig-zagging or variability. These data plot a nearly perfect natural log growth trend, a steady relentless trend.

While I don't much like what it implies, I find the underlying math quite beautiful. Nature is amazing.

I'm going to write my prediction on a 3x5 card for my next MO visit, and hand it to him if I'm right. Maybe it will be way off, but I bet it's going to be reasonably close.