Oncotype DX Roll Call!

My OncoDX score was 18.  Although my onc very much wanted me to do 10 rounds of TC, I said no.  When the rad onc said I was too young (44) for Mammosite radiation, I joined a clinical trial, randomized in, and had that.  I am now suffering from awful SEs of Tamox and had a trip to the ER just to get some pain meds to get me through until I could see my doc this afternoon.  Sigh.

Moody--

I posted a reply to mbradford on the "please help!" thread, citing an article I recently found on http://cancerfighter.wordpress.com/2008/8/29/chemotherapy-criticised-by-australian-oncologists

I'm not sure that's the complete URL, but you can find the article by entering the headine, "Chemotherapy criticised by Australian oncologists," in your browser. It emphasizes the same point about the failure to distinguish between absolute and relative benefit.

I was considering posting on it, but thought that might get me suspected of being an anti-chemotherapy mole. As Sgt. Friday used to say,"I'm just interested in the facts, Ma'am." 

Seabee,

You asked how positive my Her-2 was.  It was 3++.

Sorry, everybody--I've been on vacation for a couple of weeks and I haven't been doing any homework assignments.

I want to go back to the question of whether we are misinterpreting our Oncotype scores.  On page 3 of this thread, mbradford12000 was quoted as having said this:

"If you look on the second page of the report, you will see TWO graphs.  It is the second graph, the one on the right that shows absolute benefit,  that is the one that shows whether or not chemo will change your chance of recurrence.  Although the graph on the left is not clearly labeled this way, according to my oncologist's explanation, that graph only shows relative benefit.  Here's the difference.

The graph on the left looks ONLY at the people who have had a recurrence.  So, let me use my case as an example.  In my case, with a score of 28, I know from page 1 that 18 out of 100 people with breast cancer like mine get a recurrence.  According to the left hand graph on the second page, recurrence was decreased by about 6 percent by having chemo plus tamoxifen. 

But here is the crucial piece of information.  This means 6 percent of the 18 who got a recurrence, not 6 percent of the original 100. ... 

That is what the graph on the right on page 2, absolute benefit, clarifies."

Okay, first of all, I'm not absolutely certain which graphs mbradford is referring to. The graphs I'm posting below are from page 2 of the "sample" Oncotype report on the Genomic Health website.  BTW, both the graphs are from this article:  S. Paik et al.  "Gene expression and benefit of chemotherapy in women with node-negative, estrogen receptor-positive breast cancer."  Journal of Clinical Oncology Vol. 24 No. 23 Aug. 10, 2006; pp. 1-12.

This is the figure on the left-hand side of page 2 of the report.  (Please correct me if this isn't the graph in your report.):

This graph does not show "relative benefit."  It is a graph of the 10-year absolute risk of recurrence plotted against the Oncotype recurrence score.  The red (dark pink?) solid line is the recurrence risk in women who got Tamoxifen but did not get chemo.  The blue solid line is the recurrence risk in women who got both chemo and Tamoxifen.  The benefit of chemo is the difference--the vertical gap--between those two lines, for any given Oncotype score.

The dashed lines in the graph are the 95% confidence intervals for the respective pink and blue lines.  I'm not very good at explaining statistics.  Basically, the confidence intervals tell you that the red and blue lines are probably not correct--there is a 95% likelihood that the actual, true recurrence risks lie somewhere between the dashed lines; they are not necessarily on the solid lines.

There are some unusual things about the graph.  One is that the recurrence risk--on the vertical axis--is expressed as a decimal fraction instead of the percentages we usually see.  So, if page 1 of the report said your risk of recurrence was 17%, it would appear as 0.17 on the vertical axis on this graph.  A recurrence risk of 40% would be 0.4.

Another thing that makes this graph interesting is that the horizontal axis shows Oncotype Recurrence Scores as a continuous variable, instead of grouped into the 3 categories of low, medium, and high.  The researchers knew that all the scores within a category did not carry the same risk.  Scores of 18 and 30 are both in the "intermediate risk" category, but the actual recurrence risks for those scores are quite different:  18 is an 11% risk; 30 is a 20% risk.  So, plotting the scores on a continuous scale is more helpful.

The bottom line with this graph is, the closer together the red and blue lines are, the less benefit there is likely to be from chemo.  When the solid lines overlap, there is no benefit.  The confidence intervals show how much uncertainty there is in these measurements... and how much more we need to know. 

This next pair of graphs, below, is from the same article.  The top graph ("A") is not in the Oncotype report.  It shows the relative benefit of chemo, for women whose scores are in the 3 categories.  The relative benefit is another way of expressing a change in the risk of recurrence.  This time, the researchers are displaying recurrence risk on the horizontal axis as a relative value.  I'm pretty sure--but not positive--that the value is a ratio of the recurrence risk with chemo, divided by the recurrence risk without chemo.  So, if the recurrence risk with chemo was 15% and the recurrence risk without chemo was 20%, the relative benefit of chemo in this graph would be 15/20 or 0.75.  If the risk with chemo was 10% and the risk without chemo was 30%, the relative benefit of chemo would be 10/30 = 0.33.  The value "1.0" means there was no difference with chemo.  A value > 1.0 means the recurrence risk was actually greater with chemo than without chemo.  Look at the huge error bars, though--there is enormous variation in the results for the low and intermediate categories.

The bottom graph ("B") is the right-hand graph on page 2 of the Oncotype report.  It shows the effect chemo has on the absolute risk of recurrence.  On the horizontal axis, "DRF" = distant recurrence-free.  The horizontal axis is the additional percentage of women who were free of distant recurrence after 10 years--that is, the difference between the percentage that were recurrence-free after chemo and tamoxifen and the percentage that were recurrence-free after only tamoxifen.

As mbradford said, only women in the "high risk" category showed a dramatic improvement in absolute risk of recurrence, with chemo (i.e., benefit from chemo).  Women in the "low risk" category showed no improvement at all in their risk.  It looks like women in the "intermediate risk" group also had no benefit from chemo, but we know that the benefit in that group varies, depending on whether the Oncotype score is at the low or high end of that range.  (See the large error bars.)  Also, as mbradford said, the TAILORx trial will hopefully clarify the benefit of chemo in that group.

I am sorry this is so long and wordy.  Please post or PM if you are still confused.

otter 

Hi Otter, welcome back, and thanks for that very detailed and instructive look at those graphs. You are indeed right in assuming that the first graph, and graph B, are the ones on p. 2 of the Oncotype report.

The second one is modified slightly, in that it is rotated to a vertical view and the decimals are expressed as percentages, but it is otherwise the same.

I think the point that mbradford is getting at is not that the graphs are misleading, but that they can be interpreted in a misleading way that makes the improvement seem greater than it is.  Take for example the first graph.  Someone with a score of 30 has a recurrence risk of 20%.  Tamoxifen (on the average) reduces this to a 15% rate of recurrence. One might say that in relative terms Tam has reduced the risk by 25%, whereas absolute terms it has only reduced the rate by 5%.

Chemo, according to the first chart, drops the rate of recurrence down to 7%, so one might say in relative terms that chemo has reduced the risk by another 40%, whereas in absolute terms it has only reduced the rate by an additional 8%.  However, a 13% overall absolute reduction in risk is pretty good. It just doesn't sound as good as a risk reduction of 65%.

If your oncotype score is thirty you should definitely do chemo according to this graph. The confusion of relative and absolute risk becomes more critial for lower scores.

A score of 22, for example, has a recurrence risk of 12%. Tam reduces this to 10% for an absolute gain of 2%. However, since 2 is 17% of 12, one could say thatTam had reduced the risk by 17%. Chemo further reduces the rate of recurrence to 7%, for an absolute gain of 3%. However, in relative terms one could say that chemo has reduced the risk an additional 25%. This sounds a lot better than a total absolute gain of 5%, which has to be weighed against the risks of side effects.

The Australian article I referenced takes an extreme view, but it does make a point about understanding the difference betweeen relative and absolute benefit.

Alo123, my Oncotype score was also 26, and I had 4 rounds of Taxotere/Cytoxan (3 wks apart).  I finished last June.  It's bad--after all, it's CHEMO ... but it wasn't as bad as I feared it would be.

Seabee, thanks for the added discussion of absolute vs. relative benefit.  Sometimes it's very difficult to tell which is being used when the results of a study are discussed, especially if it's in a news article. All the headlines and shouting about a "40% improvement" can end up looking pretty pale when you see the actual data and realize they're talking about a relative benefit.

Fortunately, my onco was pretty clear.  She said with an Oncotype score of 26, I had a 17% chance of a distant recurrence even after tamoxifen.  (The recurrence risks reported by Genomic Health for their Oncotype scores are for women who've had 5 years of tamoxifen.)  She told me chemo would decrease that risk by about one-third, bringing it down to around 12%.  Later I rechecked that "one-third" figure, and found it was pretty close to the chemo benefit calculated for me by the Adjuvant!Online calculator.

otter 

My oncotype was 19, with a 12% chance of recurrence with 5 years of tamoxifen. Both of the oncologists I consulted said that adding chemo would probably only bring it down to 10% at best. They recommended against chemo, saying that if you only are going to get a 2% benefit, your chances of having some kind of serious problems from the chemo are close to 2%, and in their eyes that suggests that it isn't worth doing. 

I think I will do the chemo!  As much as I know I willnot enjoy it.....I'm not up for the gamble.  If I can reduce my risk by any percent....isn't it worth a few months of hell?!  I don't know, but I'm gonna do it.

Otter, it seems from reading your posts that you are very well informed.  Thanks for your input....

Kamico3!  I am exactly what you are. 19Oncotype with 5 years of tamoxifen, chemo would be of no benefit, as far as my oncologist sees it. I had a mastectomy with an expander in Sept. and I will have my exchange on the 20th!!Yeah! My oncologist feels that my cancer will not reoccur, of course there are no' garuntees 'in life. Thanks for your input it nice to have reinforcement,for my decision!  God Luck! Dawn

It is always nice to hear of someone who is in the same or a very similar situation and who got the same advice! I guess the bottom line is that different doctors have different takes on things and we all also have to examine what risks we feel we face and which risks we can tolerate better than others. I was absolutely paralyzed in terms of making a decision re chemo or no chemo, so the fact that my doctors at Dana Farber were so definite in recommending against it was a relief to me, as it gave me something to use as my decision point, so to speak. That doesn't mean, though, that I don't read these posts and freak out when I see that other people in my situation went ahead and had the chemo. I second guess myself every day, actually. If my cancer recurs, will I kick myself for not having had the chemo? You betcha! Will I then end up being able to rationalize my decision by saying , well, I probably would have had a recurrence even if I had had the chemo? Probably! Am I sure I made the right decision? No, not really. I hope so.

Kamico3!Don't second guess yourself, you made the right decision based on the facts that you had at the time. I am a nurse and have had some experience with chemo, and I have seen reoccrrance even with chemo. Once you are a member of this club it never leaves you, unfortunately. I used to work at BI that is afilliated with The Farber, and it is a wonderful place,if your MDs are steering against chemo,I'd follow their advice.  My MDs are from Dartmouth-Hitchcock,and are suggesting the same thing, so based on their opions we both have made the right decision. I  like you was unsure,but after seeing first hand and figuring %, I opted to have no chemo as the oncologist advised.Did you have a mastectomy or lumpectomy? Which ever way you went was right too! this is so daunting, that you cn't eat yourself up all of the time,I find these decisions have been the hardest ones I've ever made.  Keep the Faith! Dawn

Moody--like Otter I can't tell where some of those figures come from, but if you'll look at the first graph that Otter reproduced for us, you'll see that your survival/risk split may be correct. Your DX score of 36 indicates a 25% risk of recurrence with surgery only According to the graph, Tam reduces this to about 18%, for an absolute gain of 7. Tamox  plus chemo further reduces the risk to 8%, for an absolute gain of 10. This would give you the 92/8 split you note for "combined therapy," and your oncologist seems to be suggesting that chemo plus Herceptin will work as well as Tamox plus chemo, thus giving you the same favorable odds.  This is probably based on some of the research on Herceptin, which I'm not very familiar with since I'm HER2-

I can't do anything with the  percentages of "benefit," but your odds of recurrence-free survival seem excellent.