HealingWell
Search Close Search
☰

MSKCC Nomogram - calling all math experts

Support Forums
>
Prostate Cancer
✚ New Topic ✚ Reply
12
❬ ❬ Previous Thread |Next Thread ❭ ❭
Posted 1/31/2015 2:11 AM (GMT 0)
The MSKCC post-radical prostatectomy nomogram predicts probability of remaining recurrence-free at 7 and 10 years after surgery. I assumed that the probability would be higher at 10 years. That was not the case when I put my numbers in the nomogram. My probability was 98% at 7 years and 93% at 10 years.
My input information was:

Hormone Therapy No
Radiation Therapy No
Pre-Treatment PSA 4.4 ng/mL
Age at Surgery 70 years
Months Disease-Free 72 months
Primary Gleason at Surgery Pattern 3
Secondary Gleason at Surgery Pattern 4
Gleason score at surgery 7
Surgical Margins Positive No
Extra Capsular Extension No
Seminal Vesicle Involvement No
Lymph Node Involvement No

How could the recurrence-free probability be lower at 10 years? Seems to me that the longer you are recurrence-free, the higher the probability you will remain recurrence-free.

Post Edited (Tex-Lee) : 1/31/2015 3:03:12 AM (GMT-7)

Posted 1/31/2015 2:22 AM (GMT 0)
We all have a futher chance of recurrance for every year following surgery and there are fellows (although rarer) that even have this occur at 10 or even more years. So that probability must be simply added to the "odds" and that is why you see your score decreasing to 93% at 10 years. I think that your odds of PCa being the cause of your death based on the nomogram are incredibly much lower and try to simply focus on that.

Bill K. from NJ
Posted 1/31/2015 10:06 AM (GMT 0)
MSKCC = Memorial Sloan Kettering Cancer Center
Posted 1/31/2015 11:40 AM (GMT 0)
They compute the probabilities by following a set of patients over time. So at 7 years 98% are still OK but by 10 years only 93% are.
Posted 1/31/2015 6:05 PM (GMT 0)
Tex-Lee-
I think that what you seek are "conditional probabilities." That is, if you are recurrence-free at (1-9) years, then what is the probability of your being recurrence-free at 10 years too? You can see this effect by changing the your answer to "how many months have you been without detectable cancer...?" The more months you put in, the higher the prediction of being recurrence-free at 10 years.

Keep in mind that most nomograms are accurate about 60-70% of the time.

- Allen
Posted 1/31/2015 6:57 PM (GMT 0)
You asked.
"My probability was 98% at 7 years and 93% at 10 years." "How could the recurrence-free probability be lower at 10 years? Seems to me that the longer you are recurrence-free, the higher the probability you will remain recurrence-free."

I'll give you an intuitive answer to help you picture it.
Read the first sentence: "My probability was 98% at 7 years and 93% at 10 years." That's a glass half full statement. Let's reword it as glass half empty: "My probability of recurrence is 2% at 7 years and 7% at 10 years."
Can you see how the numbers must increase as time goes on? There is no way the number for a longer time can be smaller than the shorter time. The longer the period, the more chance of recurrence. In fact the numbers indicate 5% of guys will have a recurrence between 7 and 10 years out. Look how many guys have a rising PSA after 5 years of zeros. Get it?
This is why I haven't changed my name from Worried Guy.

Jeff
Posted 1/31/2015 7:04 PM (GMT 0)
To put it in simplest terms, I guess you could say, that the nomogram is saying that if you take 100 guys with the same pathology as you (and me) who have been BCR free at 5 years (your input), 2 of them will have a BCR within 2 years, and AN ADDITIONAL 5 of them will have a BCR over the three years following that.

Post Edited (Pratoman) : 2/1/2015 9:05:03 AM (GMT-7)

Posted 1/31/2015 7:17 PM (GMT 0)
Prato from MSK page:

"Our pre-radical prostatectomy and post-radical prostatectomy nomograms use dynamic statistical formulas* that draw on data from more than 10,000 prostate cancer patients treated at MSK."

....so these 10,000 dudes were likely mostly gleason 6 and 3+4 Sevens, as surgery would be mostly recommended for them as opposed to us 8-10's. I would therefore surmise that the Nomograms are slightly skewd to you guys. Just a thought.

Cheers, Doug
Posted 2/1/2015 6:37 AM (GMT 0)
Got me, I always accepted it in the sense that we intuitively want to live as long as we want, so I figure Ok real good at five, a little less good at ten and so on. Your post intrigued me , be cause you seem to be right.{ that it would increase} I just got the flash for the answer{ so disregard my first sentence} and it may sound silly, but the longer you live your chance of having a recurrence, a heart attack, a divorce, getting run over by a bus , struck by lightening{I think u get the point} is increased simply be cause you…. you are living longer at the succeeding end point. Im sure there is an equation that could confirm that . Whadda ya think? idea side note to myself. Lord thanks for my gifts. now what the fck am I supposed to do with them, other than annoy folks
Posted 2/1/2015 11:29 AM (GMT 0)
My interpretation is:

at 7 years you have a 98% recurrence-free possibility, which is a 2% chance of recurrence (or 2% chance of shyt coming your way)
at 10 years you have a 93% recurrence-free possibility, which is a 7% chance of recurrence

.
Posted 2/1/2015 3:15 PM (GMT 0)
Tall Allen has the right answer. You are thinking about conditional probability.

Here's an analogy.\

A new-born may have a life expectancy of 80 years.

Suppose this new-born is still alive at age 70. His life expectancy may now be 17 years. Many did not make it to 70. If he got there he has a better chance of exceeding age 80.

Mel
Posted 2/1/2015 3:19 PM (GMT 0)
Mel, but how accurate can these Nomograms be? Pretty rough I would think.

Cheers, Doug
Posted 2/1/2015 3:57 PM (GMT 0)
You may be thinking conditional probability but the nomograms are not, at least not as you first gave the numbers, 98% at 7 years and 93% at 10 because the man who did not make it to 7 will not make it to 10 so the number cannot go up. You can make it give you conditional probability by playing with the years disease free as Allen said but otherwise it is not conditional probability.
Posted 2/1/2015 3:58 PM (GMT 0)
about as accurate as life expectancy figures. Fine for the aggregate. Your mileage WILL vary

Mel
Posted 2/1/2015 4:04 PM (GMT 0)

PSA3DOT7 said...
My interpretation is:

at 7 years you have a 98% recurrence-free possibility, which is a 2% chance of recurrence (or 2% chance of shyt coming your way)
at 10 years you have a 93% recurrence-free possibility, which is a 7% chance of recurrence

.


PSASOT, I think that is the wrong way to express it. The Nomograms are not saying that An individual's chance of recurrence goes up at 10 years. It says that more individuals will experience recurrence when you look at 10 years than when you look only at 7 years. That was the point of my post above, (I had the math wrong but I fixed it)
Posted 2/1/2015 4:10 PM (GMT 0)
J

Shaba.Doug said...
Prato from MSK page:

"Our pre-radical prostatectomy and post-radical prostatectomy nomograms use dynamic statistical formulas* that draw on data from more than 10,000 prostate cancer patients treated at MSK."

....so these 10,000 dudes were likely mostly gleason 6 and 3+4 Sevens, as surgery would be mostly recommended for them as opposed to us 8-10's. I would therefore surmise that the Nomograms are slightly skewd to you guys. Just a thought.

Cheers, Doug

Doug, I'm not sure that's correct, because the nomogram asks you to inout your GLeason score, along with primary and secondary grades, so if you inout for example g8 (4+4), it will isolate the MSKCC EXPERIENCE OF THOSE MEN ONLY AND GIVE THE % BCR free of men with those stats.
What is probably true, is that the size of the cohort (now I sound like tall Allen yeah) that produces the result, is larger for G6 and G7(3+4) than for the g8-10 results.

Not sure if that makes sense but it's how I see it.

Post Edited (Pratoman) : 2/1/2015 9:19:23 AM (GMT-7)

Posted 2/1/2015 4:17 PM (GMT 0)
Just as a point of clarification, these numbers we are talking about refer to rate of No BCR. The same nomogram, using the data that Tex-Lee posted in the original post, shows a Prostate Cancer Specific Survival Rate at 15 years of 99%.

I thought it would be worth pointing this out for the benefit of any newbies that might read this thread.
Posted 2/1/2015 4:19 PM (GMT 0)
I agree Prato! Out of the 10,000 there were relatively fewer 8-10's therefore a smaller pool therefore making the results of more prognostic value (or accurate) for a G6 than say a G10. It's just that they have never been close for me.

Cheers, Doug
Posted 2/1/2015 4:20 PM (GMT 0)
Got it, that is true, about the accuracy, Doug
Posted 2/1/2015 4:28 PM (GMT 0)
I don't think that is what Tex is thinking, but only he knows that. I think he stated well what he was thinking, so I am standing on that. Im surprised no one sees the simplicity of my analysis. I think a lot of folks do, but, wisely, do not want to encourage me. I thank them for that. I also believe some will throw out a thesis that shows a bias. Logic is not a popularity contest, as it skews the results and can betray the genitors own self…sometimes and simply for instant, short-lived gratification. I am very confident that my analysis is correct…but ya know…so what? Its common for folks to replace a question asked, with a spin that they can answer. but fortunately simple logic and mathematics have a life of their own. Me… I scream to be proven wrong… all searchers of truth{operative word, truth} do... they are wired that way. Now the question of good vis a vie truth is really the most puzzling conundrum, in my view. ex; When some one says the truth hurts ,after clobbering some ones souls with it, its not good, if the intent was to hurt and not enlighten. Its a fine line and we are ALL aware of when we do that. Go Hawks!!! PS surfs huge, not for me, so off to the church of the almighty body.
Posted 2/1/2015 4:42 PM (GMT 0)
Logo, as what always happens on thought provoking threads, a tangent appears followed by discussion, which you are fully aware of. Makes for interesting reading and more participation, I would imagine. I think most of us saw the simplicity of your of your analysis and agreed by being silent, at least I did.

Cheers, Doug
Posted 2/1/2015 4:43 PM (GMT 0)
Logo-translator needed!!!

Mel
Posted 2/1/2015 6:09 PM (GMT 0)
Yes, I think I see it now. The closer I get to 10 yrs after surgery, the higher probability I will be BCR-free at 10 years.

6yrs - 93% probability of BCR-free @ 10 yrs
7yrs - 95%
8yrs - 97%
9yrs - 98%

I was surprised to see that each year after surgery (0-9), the probability of being BCR-free @ 10 yrs increases by about 2% - from 82% to 98%. I assumed that recurrences would be skewed toward the early years after surgery. Math or Medical experts comments?

The MSKCC nomogram gives little comfort to the worriers (count me as one) since it shows that, immediately after surgery, men age 50-70 with pre-surgery PSA's 4-6, G3+4=7 and favorable pathology, have a 18% probability of recurrence within 10 years.

PSA anxiety anyone?

Post Edited (Tex-Lee) : 2/1/2015 11:17:44 AM (GMT-7)

Posted 2/1/2015 6:18 PM (GMT 0)

Shaba.Doug said...
Mel, but how accurate can these Nomograms be? Pretty rough I would think.

Cheers, Doug

The problem I have with these calculations is that they are using data pre-Xtandi, pre-Zytiga, pre-Xofigo, pre-Provenge, etc and pre drugs that are in the pipeline that will become available in next 7-10 years, rather than the MSK patients that have not had these drugs since the new treatments have not been arouind 7 years or 10 years.

So it's OK to use them as a point of comparative analysis, but actual results should be quite a bit better prospectively.

LupronJim
Posted 2/1/2015 10:30 PM (GMT 0)

Pratoman said...

PSA3DOT7 said...
My interpretation is:

at 7 years you have a 98% recurrence-free possibility, which is a 2% chance of recurrence (or 2% chance of shyt coming your way)
at 10 years you have a 93% recurrence-free possibility, which is a 7% chance of recurrence

.


PSASOT, I think that is the wrong way to express it. The Nomograms are not saying that An individual's chance of recurrence goes up at 10 years. It says that more individuals will experience recurrence when you look at 10 years than when you look only at 7 years. That was the point of my post above, (I had the math wrong but I fixed it)

A quote within a quote. If you read this correctly this is what it says ... an individuals change (probability) will go UP. Using your example, at 7 years the change is 2% and at 10 years it is 7% (its going up).

They've changed the Nomo output; but given my stats with my G(5+4) and everything "contained" within the prostate, the prognosis still looks pretty awful at the 10 year mark.

2 year 95%: This probability means that for every 100 patients like ME, 95 will remain disease-free 2 years after surgery, and 5 will have a recurrence within 2 years.

10 year 60%: This probability means that for every 100 patients like ME, 60 will remain disease-free 10 years after surgery, and 40 will have a recurrence within 10 years.
I wonder who the other 59 disease-free survivors will be at the 10 year mark

Yep, I suddenly feel upbeat! Yeah!
.
✚ New Topic ✚ Reply
12