Nomogram Predictive Tools--Are they accurate?

[vincent33]

I'm curious if anyone has used the tools on this site. I have plugged in my information and I'm getting much better numbers than given me by my oncologist. If anyone would like to try it out and let me know how it compares with what your oncologist told you, I would appreciate it.

https://www.mskcc.org/nomograms/colorec ... robability

[zx10guy]

I've used them in the past when I was first diagnosed. It really depends. Each nomogram's results are based on the collection/pool of data fed into it. So you would need to examine how large of a sampling the tool has access to and if the data has been updated recently.

Two others that are out there are one from Mayo Clinic and Adjuvant Online. Adjuvant Online is a tool geared towards doctors and it'll say you have to be a doctor to register for access. I still registered using made up info where necessary and got an account. I know my first oncologist used the Adjuvant Online tool.

[Rock_Robster]

There’s two main issues to consider really:

1) you need to fit the criteria for the tools; specially MSK’s nomograms only apply for stages I-III patients with complete resections. They don’t cover stage IV or unresectable cases.

2) the data that feeds these is pretty old. MSK’s uses a large study spanning 1994-2005, so around 15-25 years old. I would hope we’ve improved in treatment and prognoses during that time.

Rob

[Beckster]

A 2017 nomogram is available at. https://qcancer.org/colorectal-survival/

[Rock_Robster]

A 2017 nomogram is available at. https://qcancer.org/colorectal-survival/

I’m struggling with this calculator - it doesn’t ask about N- or M-score at all. Do we assume this is only for stage I-II disease then?

Thanks,
Rob

[claudine]

I’m struggling with this calculator - it doesn’t ask about N- or M-score at all. Do we assume this is only for stage I-II disease then?

The drop down menu for tumor stage goes form 1 through 4 - isn’t that equivalent to cancer stage I-IV?

[Rock_Robster]

I’m struggling with this calculator - it doesn’t ask about N- or M-score at all. Do we assume this is only for stage I-II disease then?

The drop down menu for tumor stage goes form 1 through 4 - isn’t that equivalent to cancer stage I-IV?

I think that’s T-stage (ie the stage of the primary tumour, T1-4), not disease stage (ie stage I to IV). It doesn’t ask about lymphatic spread or metastases which are some of the most important prognostic factors, so makes me think it doesn’t cover these. Also the results look very optimistic if it were to be stage IV disease.

[vincent33]

I would appreciate if we could go back to the original post and the nomogram tool linked in that post. This tool is far superior to the one later linked. I'm hoping some will consider putting in their own stats and letting me know if the results are consistent with what they were told by their oncologists. Thanks.

[claudine]

Ha ok, thanks Rob! It’s true that the input is very minimal. Regardless, what I found interesting is that the output stated « Of those who don't survive, 39 are likely to die due to their colorectal cancer and 23 due to other causes.«  That surprised me, I'd have expected a much lower value than 23%, but I guess there are many comorbidities associated with cancer, like this friend of my parents who had lung cancer but had a heart attack during surgery...

[vincent33]

Can someone link to "adjuvent online"? I can't find it. Thanks.

[Beckster]

Can someone link to "adjuvent online"? I can't find it. Thanks.

Adjuvent online is still upgrading...however try this. https://oncoassist.com/adjuvant-tools/ You have to create an account so I used "other" in the required fields and it accepted the information. I was never able to find this before....The results from my information showed a difference from taking chemo to not. The old MSK showed a lower survival if I took chemo, but this one shows a benefit, which correlates to the current research.

[boxhill]

Whether these things are accurate probably depends on what therapies you can use effectively. I very much doubt that they include many cases of Stage 4 people who are excellent responders to immunotherapy, for example.

[JJH]

I am posting the information below so that we may more fully understand the colon cancer, Stage I to Stage III nomograms available on the MSK website:

There are two colon cancer nomograms listed the MSK website, and I will call them MSK-nomogram and SEER-nomogram, respectively:

• MSK-nomogram
  https://www.mskcc.org/nomograms/colorectal/disease_free_probability
  This nomogram is based on a sample of 1,320 Stage I, II, III colon cancer patients, all from MSK, who had all cancer removed by surgery and were observed from 1990 though 2000. During this period, the only available adjuvant chemo regimen was raw 5-FU monotherapy. The outcome variables for this nomogram were "5-year and 10-year Disease-Free Probability Following Surgery" i.e., the probability of being disease free at the end of the 5-year or 10-year observation periods.
  .
• SEER-nomogram,
  https://www.mskcc.org/nomograms/colorectal/overall_survival_probability
  This nomogram is based on a sample of 128,853 Stage I, II, III colon cancer patients, from various hospitals nationwide, who had all cancer removed by surgery and were observed from 1994 until 2005. During this period several different adjuvant chemo regimens became available as they were approved by FDA, including: 5-FU monotherapy, 5-FU/Leucovorin, Irinotecan, FOLFIRI, FOLFOX, XELOX/CAPEOX, Erbitux, FOLFOX+Avastin. The outcome variable for this nomogram was "5-year Overall Survival Following Surgery"

FDA Approv.
========
1962-06-01 Adrucil (5-FU)
<<<< 2000 Cut-off date for MSK nomogram: " Disease-Free Probability Following Surgery" (1,320 patients) >>>>

2000-05-01 Camptosar (Irinotecan Hydrochloride)
2001-08-01 Xeloda (Capecitabine)
2002-02-12 Leucovorin (Folinic acid)
2002-08-09 Eloxatin (Oxaliplatin)
2004-02-12 Erbitux (Cetuximab)
2004-02-26 Avastin (Bevacizumab)
<<<< 2005 Cut-off date for SEER nomogram: "Overall Survival Probability Following Surgery " (128,853 patients) >>>>

2006-09-27 Vectibix (Panitumumab)
2012-08-03 Zaltrap (Ziv-Aflibercept)
2012-09-27 Stivarga (Regorafenib)
2015-04-24 Cyramza (Ramucirumab)
2015-08-22 Lonsurf (Trifluridine and Tipiracil Hydrochloride)
2017-05-23 Keytruda (Pembrolizumab)
2017-08-01 Opdivo (Nivolumab)
2018-07-10 Yervoy (Ipilimumab)


There are a number of questions that could be posed concerning these two nomograms. For example:

DATA TOO OLD?
Old, out-of-date chemo regimens?
Old surgical techniques? no laparoscopy; no robotic?
Old technology? old staplers; old suture technology?
Old radiation protocols?

SAMPLING FRAME ADEQUATE?
- Biased sample?
- Small sample?

SCOPE TOO NARROW?
- Location of tumor: Only colon tumors included; rectal tumors not included
- Staging: Stage I, II or III only; no Stage IV/mCRC
- Resectability: Only R0 (totally resectable) cases included; no partial resections included

METHODOLOGY?
-Cox proportional hazards model appropriate? Was "proportional hazards assumption" met?

PREDICTOR VARIABLES APPROPRIATELY CHOSEN?
- Too few biomarkers? CEA only; no CA-19-9; no KRAS/NRAS/BRAF bio-markers; no MSI status
- No chemo regimen details for SEER; No info on time spent on chemo.
- No details on type of surgery.
- No details on possible co-morbidities (e.g., diabetes; heart condition, etc.)
- No details on possible additional risk factors (e.g., smoking; obesity, etc.)
- Overlapping or intercorrelated predictors? (absence of orthogonality/independence of predictors, etc.)

OUTCOME VARIABLES APPROPRIATELY CHOSEN?
- etc.

Any thoughts?

[claudine]

Nothing for stage IV I take it...