How Much Stock Should I Put in R-IPI

As my positives dwindle, the one thing I still hang on to is the revised IPI. I have one or two points (depending on how the bone marrow comes back). I still show a good prognosis which I am holding on to very tightly.

As far as I can tell, this is the best predictor of outcome.


P.S. this is a grreat group and really helps keep me spirits up...I really appreciate it.

Statistically speaking, the IPI has been the best known predictor of outcome, but as you note there is a revised IPI out, which, unlike the previous one, didn't take into account rituximab. So we always kind of knew that the figures on the old IPI were lower than reality.

that said, in order for the R-IPI to be applicable, you kind of have to fall into the specific patient population studied. I know that it seems to be touted as being for 'non-Hodgkin's lymphoma' but in reality, it often refers to DLBCL, and the treatment of R-CHOP. Since you apply to both (I think), then it's reasonable for you to put a fair amount of stock into the figures. Even if your bone marrow comes back positive, you'll still be considered in a low-risk group, although I don't know the five year %s offhand.


I forgot that when I first got my Itouch I downloaded an app entitled 'HemeCalc' featuring a bunch of related calculators relating to blood disorders, and I put your stats in (assumed your LDH was normal) and you despite the extent of your disease, the RIPI still favors you, with a 'cure' rate of around 80%. Prior to rituximab, that was closer to 50%.


Thanks Ross-

My ldh is in the normal range and I still haven't heard about the bone marrow. Even if it comes back positive I would have two points (stage 4 & possibly two extranodal) I still am in the "good prognosis". That's what's weird either one point or two keeps you at the same place.

I also heard an interesting thing about overall survival rates. According to a clinician I know, the os factors many things, but he said many people struggle with the treatment and it can cause vital issues. They include heart disease, diabetes and other health issues not factored into os. Is that your understanding, too?

The more you learn about biostatistics, the less faith you'll have in the medical community at large. They are almost deliberately designed to allow for bias in whatever direction most favors one's desired results. As for OS, it does have a standard definition, but it normally has a narrow application field. Typically it means the length of time one survives beginning at the time of diagnosis and ending, of course, with death from the designated disease-- unless it states otherwise, in which case those people are generally excluded from the results.

i'm not really familiar with diabetes as a common consequence of cancer treatment, but my unfamiliarity means nothing. complications from cancer treatment however, especially secondary cancers, cardiopulmonary toxicity, and other organ or systemic problems are not uncommon.

The thing is, if you are treated for a cancer and you make it past the magical five year line without relapse, you fall into the cured category. in the language of biostatistics, which is used everywhere in cancer, five year survival essentially means a cure. and this stands to some reason-- with each year being cancer-free, your odds of relapsing go down. Still, should you relapse in year six, or should you develop a secondary cancer as a consequence of the radiation etc ten or twenty years down the road, you're still going to be counted among the 'cured' for your original cancer-- and this includes whether you developed heart disease from the treatment.

biostats are rife with bias. detection bias is very common; let's say your cancer was discovered early, in stage I, but despite treatment you die from it in 3 years. then there's someone whose same cancer type was discovered in stage IV and who died after just 6 months. both survival times could be worked into 'mean' overall survival stat, which is 21 months-- a stat that's mathematically accurate, but true for nobody.

that example is too small to really push home the point, but the things to remember are that
1. biostats have their limits, and in the end, you're not a stat, you're a unique individual
2. most cancer treatments have little regard for the future. most oncologists, i would argue, have little regard for you down the road. they want you cancer free now. if you get cancer again, they'll do what they can at that time. if you develop heart or lung problems, you'll see a cardiologist, not an oncologist. the only cancer I know of where this principle is beginning to change is in Hodgkin's, where the 'cure' has become frequent enough for medicine to start looking into ways of making that 'cure' less toxic.

so to answer your initial question about OS, yes, that's my understanding too.


Ross, this is an awesome post, thank you for your dedication to this group.

Good health,


Ross -

Sorry for all the questions, but the "biostat" you are talking about is in relation to the "OS", correct? It's not the R-IPI, right?


Sorry Cowboy, I somehow missed this. And to answer your question, yes the OS is what I'm referring to.

Thanks Ross -

Btw, I got the results of the bone marrow biopsy and it was negative. If I'm doing the math correctly, I have 1 point on the ipi - for stage 4.

I have 1 extranodal location due to bone spots. Less than two does not get a point. I asked the ong and he said the bone counts as 1 even though I have five different spots. Ross, is that your understanding, too?

This doesn't change the prognosis, but it is one less thing to keep me up at night.

Ross, thanks as always for your thoughts and time.

First off, dynamite news about the bone marrow, that's fucking great news.

Second, right-- one spot, five spots, still considered a single extranodal site.

Third, your math is correct, you now score a 1 on the RIPI, which you owe to the staging. And no, it doesn't change the percentages or the prognosis (as per the index), but yeah it's one less thing to keep you up at night, and it's a BIG one less thing.


Ross -

I hate to keep harping on this, but I have four somewhat related questions dealing with predicted outcome:

1 - I'm confused why someone with two risk factors (on the R-IPI) is the same as one - or three risk factors to five? It would seem as there would be some change in predicted outcome. Is this because they changed the number of categories to three (Very Good, Good, Poor?

2 - I've read that "bulky disease" impacts outcome, but it is not factored into the index. Do you know why?

3 - The number of extranodal sites are factored into the index, but where they are located are not. I would assume there is a difference if you have the disease in your central nervous system/brain rather than the spleen, etc.

4 - Can you list in order, from most important to least important, the R-IPI risk factors? Are they equal in importance?

Thanks in advance.

You're not harping on anything, don't sweat it. i just hope I can handle these. I've backed up my answers as best I can with material from the original paper that established the RIPI, and from source material from the NCI and the AJCC, among others. I'm answering 2-4 first, and 1 last because I'm including a chart in that answer and for formatting reasons it works better to include this lower down.

2. To my knowledge, "bulky disease", or tumor bulk, is not factored into the outcome of the RIPI for at least one reason--- because its actual effect on outcome has only been reported thus far in preliminary studies. In other words, it may or may not be factored into a future version of the IPI but so far the data hasn't been sufficiently convincing. truth be told, you can put good money on the notion that the R-IPI will undergo another R, and probably another R after that.

Tumor bulk isn't alone: the same holds true or used to hold true for beta-2 microglobulin levels (a protein on the surface of lymphocytes and other cells that can be detected in blood and urine; a high level could indicate cancer) and the S-phase fraction (the S-phase is the phase of the cell cycle when DNA gets duplicated so that when the cell splits, each new cell will have a full compliment of genetic material. The fraction is a measurement of how many cells in a tumor are in this phase, and it helps give an understanding of a tumor's aggressiveness).

So far I don't know of any molecular markers that have been proven to be reliable as prognostic factors in DLBCL, but the key thing to remember about the R-IPI is that it is only applicable if you are treated with R-CHOP and radiation. In other words it has limitations. Not only that but the study that established it was only meant to assess the validity of the IPI in light of the arrival of Rituximab. The study didn't try to seek out any new prognostic markers. Rather, it recommended that future research should try to make the IPI even more predictive, and no doubt it will.

3. I agree with you that where the sites are probably does have some impact, but the R-IPI found that that didn't matter in terms of the measurements it used. That said, after re-reading a short version of the original paper published in Blood, it doesn't appear that they took what the actual extranodal sites were into account. I might be wrong about that. But even if it did, their results indicate that the fewer the sites, the better the prognosis.

On that note, when staging NHLs, the extranodal site IS important, since there are a few sites that, if they're involved, then by convention you automatically earn the stage IV dx-- bone marrow, liver, and two others that slip my mind are the sites, suggesting that these sites are more serious than others.

4. The five risk factors were identified as independent prognostic indicators, so they can't really be compared to one another and rated according to the R-IPI. So I can't list the order, and I don't think they're meant to be listed, but I'm sure that people could make the case for any of them being more important than any other, but not in reference to the R-IPI.

Now, as one example, you can always argue that the younger you are, the better you will (probably) deal with chemo, which might make the age factor more important, but then again, an unhealthy-as-hell 45 year old is probably going to be worse off than a mentally tough 65 year old who runs marathons. Health status- beyond the cancer-- means a lot. So does how a person deals with treatment. But these things aren't part of the R-IPI, nor can they be measured with any degree of objectivity. The only thing that comes close is the ECOG performance status-- which also happens to be the least objective indicator in the R-IPI.

1. On this question, I think you have it backwards; they changed the number of categories to three BECAUSE the data showed that people with 0, 1-2, and 3-5 risk factors had very similar outcomes (not the exact same outcomes, but close enough), as illustrated by this figure (from the original paper, which I have no business reproducing).


If you have any follow-up questions, or disputes with my responses, or I'm way off on something, let's hear it. Don't hesitate.


I guess my only follow up is on extranodal. Does bone get an automatic stage iv or does it depend how far the lesions are from the node?

So I think you already answered this yourself in the next comment? Bone involvement doesn’t score an automatic stage IV; however, part of stage IV is defined as 'diffuse or disseminated involvement of one or more extralymphatic organs, with or without associated lymph node involvement."

‘Diffuse’ means widely spread out in the body, not localized. Where are your bone lesions?

Ross -

I found this about automatic stage iv -

Stage IV includes any involvement of the liver or bone marrow, lungs (other than by direct extension from another site), or cerebrospinal fluid.

Does this sound right?


That looks right although technically it’s not the lungs but the pleura.

Hey Ross -

Did you see my recent post? I would really appreciate your thoughts.