Memorial Sloan Katering Prostate Cancer Risk Calculation Nomograms

Posted by soli @soli, Sep 21 10:31am

Memorial Sloan Katering has excellent pre- and post-radical prostatectomy nomograms which can be used by patients before or after their surgical treatment for prostate cancer to enter their data and calculate risks. For example, the post-surgery nomogram predicts the probability of remaining cancer recurrence-free at two, five, seven, and ten years following surgery.

To get your particular result, you enter your age, pre-treatment PSA, biopsy Gleason score, clinical tumor state (e.g. T3b in my case), pathological results (organ-confined or not, extracapsular extension, lymph node involvement, seminal vescible invasion, number of biopsy positive cores etc.) and voila, the system generates the % probability for your 15 yr prostate-cancer specific survival, and progression-free probability after surgery in 5 yrs, 10 yrs etc.

Since my surgeon did not do extensive lymph node dissection, I did one calcuation assuming that lymph nodes were NOT involved, and I did another one assuming lymph nodes WERE involved, and the resutls - as you would expect - were radically different.

I believe this is a very valubale tool to assess your risk profile based on your specific disease characteristics.

Interested in more discussions like this? Go to the Prostate Cancer Support Group.

zmarkv | @zmarkv | Sep 24 5:20am

In reply to @soli "Good to know. Thanks" + (show)

Which AI did you use?

manutebol | @manutebol | Sep 25 5:18am

In reply to @zmarkv "Which AI did you use?" + (show)

@zmarkv

I ran it through ChatGPT, Grok, and Perplexity. ChatGPT and Grok provided almost identical odds. Perplexity a little worse.

My query was constructed like so:

What is my prostate cancer biochemical recurrence probability if I was stage pT2c after surgery, with Gleason 4+3 with 60% pattern 4, equivocal 2mm anterior margin, 10% of the pattern 4 was cribiform, pre-op PSA 34, and Prolaris score 4.0? Note that I have had a PSA of < 0.01 since my surgery two years ago.

jeff Marchi | @jeffmarc | Sep 25 10:18am

In reply to @manutebol "@zmarkv I ran it through ChatGPT, Grok, and Perplexity. ChatGPT and Grok provided almost identical odds...." + (show)

@manutebol
So what did each AI tell you your reoccurrence chance was?

I suspect you never got a decipher score, is that true?

soli | @soli | Oct 1 3:23pm

In reply to @manutebol "I would also add that I’ve played around using AI with respect to my recurrence odds...." + (show)

@manutebol

Learning from your experience, I entered my data into ChatGPT and this is the estimated probability of biochemical recurrence for my spefiic case according to ChatGPT.

While it’s difficult to provide an exact probability without access to a comprehensive model, we can refer to the commonly used Partin tables (which combine pathological stage, Gleason score, and PSA to estimate recurrence risk) and nomograms that incorporate more detailed information (such as tumor size, lymphovascular invasion, etc.).

Given the factors you provided, your overall risk of biochemical recurrence after surgery seems to fall into the high-risk category. Here’s an approximate breakdown of factors:

pT3b stage (seminal vesicle involvement): Increases risk significantly.
Gleason score 3+4 (30% pattern 4): This would likely put you in the intermediate-to-high risk category.
Pre-op PSA 14.5: A PSA level this high before surgery also suggests a higher likelihood of recurrence.
Bilateral seminal vesicle involvement, lymphovascular invasion, perineal invasion: These are all signs of a more aggressive disease, further increasing the risk.

Given all of these, it’s likely that your probability of biochemical recurrence (defined as a detectable PSA level after surgery) within 5-10 years could be in the range of 50% to 70% based on typical recurrence data for high-risk cases. Of course, this is a rough estimate, and using a formal nomogram (like the Kattan nomogram or Partin tables) or other prediction tools would give a more tailored risk profile.