Recurrence rates vs. survival rates

redsox · September 2010

redsox · September 2010

A clinical trial may report a significant difference in rates of recurrence but no statistically significant difference in survival. If what we care about most is whether or not the disease is going to kill us, should we conclude that the treatment in question is not worth it?

It depends... on the study design, the length of time patients were followed, the expected and actual rates of recurrence and death.

The protocol for a clinical trial specifies the primary endpoint and also multiple secondary endpoints. The statistical design of the study includes a goal for the number of patients to be accrued. This number is derived from calculations designed to assure that the study has sufficient statistical power to detect a difference of a certain amount with a specified degree of certainty. These power calculations which lead to the target number of patients are based on the primary endpoint. Although there are secondary endpoints, the power calculations are not based on those and the study is generally underpowered (does not have enough cases to expect to detect differences at a reasonable level) for these endpoints.

For the trial to find differences there have to be events. If the endpoint is survival the event is death; if it is recurrence the event is a recurrence. If there are few events a statistical difference will not be detectable unless the sample size is large and the time period of follow-up is long. None of the clinical trials for DCIS had survival as the primary endpoint, although it is usually a secondary endpoint and reported as such. These trials did not have sufficient statistical power to expect to find a difference in survival. If they report no significant difference in survival, all you can conclude is that the study was not designed with enough power to find such a difference.

The primary endpoint of most studies for DCIS or early stage IDC is usually recurrence. Most reports will also emphasize the percentage of recurrences that are DCIS vs. invasive breast cancer. Although we do care about recurrence in and of itself, we care particularly about IDC recurrences because they may lead to advanced disease and death. If the study shows a difference in recurrence rates with some % being IDC it is reasonable to assume that for some % of those women in the group with higher recurrence rates the disease will progress and they will die from the disease. If you had long enough follow-up a difference in survival would eventually show up but the difference may be small for DCIS. We each have the right to decide whether or not we are willing to take the risk.

Beesie · September 2010

Interesting.

What we know is that the overall long-term survival rate for DCIS is about 98%. That seems to be the average of all the studies I've read.... a few studies have the number at as low as 95% but most say 98% - 99%.

What we also know is that women who have pure DCIS have a 100% survival rate, if they don't have a recurrence. By definition, pure DCIS cannot move outside of the breast and breast cancer in the breast is never fatal.

The logical conclusion therefore is that only women who have pure DCIS who then have a recurrence will succumb to breast cancer. Or in other words, the 2% of women with DCIS who eventually do die of the disease must have had a recurrence. And the recurrence must have been IDC, not DCIS.

Based on this, the way I look at it is that the objective of DCIS treatment is to minimize the risk of recurrence to an 'acceptable' level. Recurrence risk will never be 0% - that's impossible, no matter what surgery or treatment we have. So there will always be some level of risk. But now we get to the tricky part. What is an 'acceptable' level of recurrence risk? And what is acceptable as a treatment to get to the lowest possible risk level? This depends entirely on the individual.

Some of us are more risk averse than others. A 10% recurrence risk might be just fine for some women but other women might not be comfortable with a recurrence risk at that level.

It's also true that some of us are willing to do things to reduce our risk that others wouldn't consider. Every treatment comes with it's own risks and side effects, so we have to balance the risks/side effects of the treatments with the benefits we get in terms of BC recurrence risk reduction. Some women would remove both their breasts to avoid taking Tamoxifen. Other women would gladly take Tamoxifen if it means that they can keep their breasts.

The last - and possibly most important factor - that plays into recurrence risk decisions is the diagnosis itself. Someone who had a small amount of low grade DCIS will likely have a low recurrence risk (4% - 5%) after a lumpectomy alone. For these individuals, with such a low risk, the benefit from other treatments is low. Because of this, some women in this position choose to take no other treatments. Other women choose to have mastectomies or to have radiation and take Tamoxifen; these two actions can reduce their risk to about 1% - 2%. For the first group of women, these treatments are 'over-treatment' but to the second group of women, these treatments are 'necessary'.

Then there are women who have large amounts of high grade DCIS. After a lumpectomy alone, their recurrence risk could be as high as 40% - 50%. For these women, some additional treatment is necessary. But what treatment? Individual preferences and individual risk profiles really come into play here. Some women will choose to have a mastectomy, to get their risk to 1% - 2%. Other women will consider that to be too drastic and unnecessary, and will instead have radiation and take Tamoxifen, and will live with a recurrence risk of 10% - 12%.

It's all in how we look at recurrence risk, and how we assess it as compared to other risks and side effects (the risks from radiation, a life without breasts, etc.). In the end though, none of us can avoid have some level of recurrence risk.

redsox · September 2010

I agree with you that the goal for anybody with DCIS is to identify the recurrence risk and the level she is willing to accept and to choose treatments that achieve that level.

I posted this topic because I have seen some posters on other threads argue that a study that shows no difference in survival means that the treatment in question is worthless even if there is a difference in recurrence rates. Clinically and statistically that is not a valid argument. With DCIS there is a second chance to treat a recurrence and have very good odds of success, but recurrence is still a real threat for more advanced disease and for another round of treatment. While I don't want to ignore the small probability that the cancer might be advanced at the time of recurrence, for most women who have good follow-up testing the recurrence will be DCIS or early stage. The main question is, "How much am I willing to risk having to go through this again and what am I willing to do to avoid that?"

I would modify somewhat your examples that are based on the size and grade of the tumor. The most important factor determining whether lumpectomy alone is sufficient is the width of the margins and the ability to achieve a truly wide excision. Size is clearly an important factor but grade is less clear (Some studies - presented at the NIH Consensus Conference on DCIS a year ago -- show recurrence rates by grade significantly different at 5 years only to come together at 10 years, indicating that low grade has a longer average time to recurrence but a similar rate of recurrence compared to high grade.). Other factors such as location of the tumor, whether there are signs of residual disease even with clear margins, whether there were symptoms such as discharge that might indicate potential dissemination of cancer cells, what else is going on in the breast (and still more factors) may affect the recurrence risk. I had a small amount of low grade DCIS and after extensive research and discussions I was quite sure my recurrence risk was unacceptably high with only a lumpectomy giving the results mine achieved. (The surgeon did a great job but wide margins were not possible.) The devil is in the details.

JanetinVirginia · September 2010

This is so helpful to read. I picked up literally an armful of material from the breast cancer center, but the content is so "dumbed down" they were pretty worthless. I think a lot of doctors do this too. Where do you find detailed info about trials, survival/recurrence and treatments?

I have areas of IDC and some areas of DCIS. Having mastectomy 9/28 as a first step. Already had an excisional biopsy (pretty much same as lumpectomy) and margins weren't clear. I had read survival same (lump/rads vs. mastec) but recurrence little lower w/mastec. And no guarantee a 2nd lumpec. would be clear and might end up with mastec anyway. Wanted to avoid radiation if I could. But I'm already looking ahead in terms of what may be next depending on node involvement.

redsox · September 2010

Janet,

I agree that most of the material they hand out or that you can find in a general book store is dumbed down and that many doctors start explaining at an elementary level and have trouble adjusting their explanations to match the patient's knowledge base. We do all have different backgrounds with some aspects that may help in different areas. A nurse will understand some clinical information better than most of us do. Someone with research experience may understand statistical design and analysis better than most. ...and so forth for other relevant skills and experience. There is so much information to try to absorb! It does seem to help push most doctors to a higher level explanation if you have read relevant material and acquired some of the vocabulary. I found it is best to take things one step at a time and research what comes next without trying to follow the whole future path at once.

What will help you depends on your background, but some suggestions:

This National Cancer Institute website has good summary information.

http://www.cancer.gov/cancertopics/pdq/treatment/breast/Patient

You will notice that it has a tab for the "Patient Version" and another for the "Health Professional Version". Try the patient part and then if you are still looking for more, try the health professional one. Note the references and try following some of those that seem relevant. You can be the judge of how much you are getting from these articles. You won't understand it the way an oncologist would, but you can tell if you understand enough to keep going. You can search for references at the following website and find copies of the abstracts and often the full articles:

http://www.ncbi.nlm.nih.gov/pubmed

Another good site is:

http://www.nccn.org/clinical.asp

especially the clinical guidelines for breast cancer. You have to register but should be able to access the guidelines after you register. Again follow the references they cite for questions you are interested in.

Beesie · September 2010

redsox, you're right. I simplified my comment about those who have small, low grade tumors generally having low recurrence rates. I agree with you that the most important factor in determining recurrence risk is the size of the margins. In my statement I made an unstated assumption that if you have a small tumor, the odds are good that your surgeon will be able to get good margins. That's probably a fair assumption in many cases, but as you point out, it's not true in all cases. And I also assumed, but didn't state, that for those who have larger areas of DCIS, getting good margins (ideal margins being 10mm) is difficult. That's usually true too, but some women who are large breasted may have good margins after a lumpectomy and as a result, have lower than average recurrence risk. Thanks for making the modification to my point - what you said is more accurate.

I've read some of the thread that you mentioned and honestly, it's astounded me. I don't agree at all that a treatment is worthless if all it does is reduce your recurrence risk, without any apparent impact on survival. Particularly for those who have DCIS or early stage invasive cancer, if you are lucky enough that you don't need treatments like chemo and Herceptin, why leave yourself at high risk of a recurrence? 50% of recurrences of DCIS are not found until they have already become invasive, and while some of these recurrences still won't require additional treatments like chemo, others certainly will. Putting aside the question of long-term survival but just thinking about the treatments that might be necessary if you have a recurrence, why put yourself through that if it can be avoided?

Having said that, I also don't agree with those who say that we should do "everything we can" to reduce our recurrence risk. Every treatment comes with risks and side effects, so I think it's important to weigh the benefit of the treatment (in terms of how much it will reduce your recurrence risk) against the risks of this treatment. Then decide which treatments make sense for you. This benefit/risk equation is different for each of us, because the same treatment will have different benefits depending on your diagnosis and pathology, and it may expose you to different risks, depending on your personal and family health history. Tamoxifen, for example, might make tremendous sense for one women with DCIS (someone at high risk of recurrence) while making little sense (and providing little benefit) to another women with DCIS (someone who had a bilateral mastectomy, for example).

Too often we look for cookie cutter answers on what to do and we try to find our decisions in what someone else did. I always find this concerning, because each of us has a different diagnosis, a different pathology, a different personal health history, and a different tolerance of risk. Every situation is unique and there is no short-cut to making the right decision. You need to learn the facts about your risks, and how each treatment will affect you personally.

redsox · September 2010

Beesie,

I agree with the points you made, especially that recurrence risk is worth reducing for its own sake but not at all costs.

One other item anyone should check out when reading a report of a clinical trial is the eligibility criteria for the study. Would you have met those criteria? If not, do the results apply to you? Often, those results give the best data available but they should be applied cautiously. Patients are usually excluded for a particular characteristic because the investigators think they are different, maybe with worse prognosis or maybe just unknown. Either way, it is hard to evaluate not only for the patient but also for the doctor.

JanetinVirginia · September 2010

Very thoughtful discussion - thanks winterstorm for starting it.

I had thyroid cancer when I was 20 and I'm almost 60 now. Having cancer shapes the rest of your life. However, I can't say that it was 'always' on my mind -- just nagging reminders now & then. I decided long ago to enjoy whatever time 'cancer free' and live life to the fullest - that there would be plenty of time to worry about it if it ever happened again. During that time I was blessed with two children & a great career. That's how I intend to approach it this time too - do what's necessary to buy time & live my life.

Lowrider - I think you are exactly right. Research needs to focus on the "on/off" switches and individualized gene therapies. Everything has happened so fast, I haven't had time to look into who/where that kind of research is taking place. If anyone knows where, please post, because that's where I'm going to place my bet & donate.

Dilly · September 2010

Hello Ladies, you'll find me mostly on the old ladies thread, but also IDC. However had there been this forum 14 years ago, I would have been here, lumpectomied with wide margins and cured. My rb DCIS was found thru my 50th bday screening mammogram, and was 4mm - teeny. Perhaps if there had been some of the newer tests & trials, I might have had rads then, but I also have a fantastically rare autoimmune blistering disease, which makes those decisions very difficult. I was fine with my treatment. I wasn't too worried because it seems like I'm always growing something wierd that the docs biopsy, and end up benign.

Besides annual physicals, blood work, etc, we set up regular mammos and vaginal us, did a screening bone scan & I had regular colonoscopies. In '03, and just beginning menopause late in life, the annual vaginal ultrasound was suspicious but hyst/ooph proved benign. That may have saved my life by removing so much estrogen -- which my recent tumor feasted on. My current (Screening mammo Mar 2010; surgery Apr. 2010) dx is lb IDC <1cm Stage 1b, grade 2, ER 95%+, PR-, Her2-; I had another lumpectomy, rads, and now 5 yrs Anastrazole. Now, 'things' and attitudes are different. My surgeon assures me this is a separate event, not a recurrence of the DCIS. Edited to add: Perhaps I should mention I did the Oncotype test this time, and was squarely in the middle grey area, but in my particular situation, 'statistics' showed chemo would save one life of a hundred, and we decided that the risks of chemo outweighed the relative benefit. Definitely a difficult personal decision for anyone.

I've been told that adjuvantonline can give you some of the statistics you're looking for and that you don't really need to be a doctor to use the service. (google it) Also, someone mentioned on a DCIS post that a lot of research is being done on the West Coast. I think that's true. The support group Breast Cancer Connections in Palo Alto has a large online database and some good help. (google them)

Also, the Humboldt Community Breast Health Project has a link to a DCIS clinical trial being sponsored by Dr. Love. http://www.hcbhp.org/events.shtml#LoveMahoneyStudy

Best wishes to all for long and loving lives; Be Well

redsox · September 2010

Another point about the design of clinical trials --

The burden of proof is on the advocate of an experimental or additional treatment to show that it is worth adding to or changing treatment. That means that trials testing lumpectomy vs. lumpectomy + radiation therapy put the burden of proof on the side of those favoring the additional treatment.

This drives the statistical design of the trials. The statistical design generally assumes the null hypothesis regarding the primary endpoint, that is -- the study is set up to test the hypothesis that the treatment being tested does not (null hypothesis) result in improvement for the primary endpoint. For example, to test the benefit of radiation therapy after lumpectomy a trial will randomize consenting patients to either of two arms of the study, 1) lumpectomy alone or 2) lumpectomy + radiation therapy. The null hypothesis is that there is no difference in the primary endpoint in the two arms of the study. For DCIS studies the endpoint is recurrence rate. (As discussed above, secondary endpoints such as survival will be reported but a finding of no difference does not prove that no difference exists. There is simply not enough statistical power to detect whether a difference exists or not.)

The power calculations based on the primary endpoint lead to a target number of patients for the trial. These are set to have a large enough sample size so that you have a 95% (usually) chance of detecting a specified difference (e.g. 10%) in the primary endpoint.

When the trial is completed and follow-up time (or event rate) has been met, the data is analyzed. To reject the null hypothesis the difference in the rates of the endpoints between the two arms must be great enough that you can be 95% certain that there really is a difference in the rates of the primary endpoint between the two arms of the study. When a clinical trial meets that difficult threshold (statistically significant) it provides the strongest possible evidence that the additional/experimental treatment is beneficial. Because statistical flukes are always possible in any one study, two or even three clinical trials with similar results are preferred to define a standard of care.

When a trial does not show that the difference in the primary endpoint between the arms of the trial is statistically significant the benefit to the additional treatment is generally rejected although there may in fact be some benefit.

For the question of whether radiation therapy after lumpectomy makes a difference in recurrence rates there are four clinical trials that show consistently that radiation therapy reduces the recurrence rate by ~50% and that about half of recurrences are invasive cancer. Since the invasive recurrences carry the threat ultimately of death from breast cancer they are a particular concern.

Beesie · September 2010

redsox, great explanation of research technique, and I say that as someone who's worked on and/or been responsible for dozens and dozens of research studies (non-medical) over the past 30 years.

"For the question of whether radiation therapy after lumpectomy makes a difference in recurrence rates there are four clinical trials that show consistently that radiation therapy reduces the recurrence rate by ~50% and that about half of recurrences are invasive cancer. Since the invasive recurrences carry the threat ultimately of death from breast cancer they are a particular concern. " Yeah, the evidence is pretty compelling and the conclusion seems sort of obvious, doesn't it? But I guess not to everyone.

mathteacher · September 2010

Overall survival is the gold standard.

All you have to do is provide studies that show women who took radiation lived longer of all causes at 20 years out. The evidence we found showed no overall survival value. But you have found other evidence?

I am eagerly looking forward to seeing this evidence.

Thank so much for looking into this.

wyldblumusic · September 2010

Snark deleted in order to keep the peace...

Anonymous · September 2010

Really?

-------------------------------------------------------------------------------------------------------

Not knowing when the dawn is coming, I open every door." -- Emily Dickinson

AlohaGirl · September 2010

This is a very interesting thread, though I must say some of these snarky comments seem unnecessary. Don't we have enough to deal with without bickering with each other?

Mathteacher and Lucy88, I gather that your concern is that 20 years out more (or as many) women are dead as a result of the side effects of radiation as would have been dead as a result of the breast cancer? Seems like a reasonable question to me, but it sounds like Redsox was saying in her original post that there aren't studies (at least relating to DCIS) that would get to that answer. Further, my guess is that any studies on women who had radiation 20 years ago might not accurately reflect the side effects of radiation today since (at least as I understand it) radiation is more refined now than it was 20 years ago. That said, I would think that one could review studies showing what recurrence rates are with and without radiation and you might be able to determine what percentage of women with recurrence end up dying of breast cancer within some time period. (That is to say, look at studies to figure out how many women with DCIS have an invasive recurrence, then look at studies of the survival rate for women with invasive cancer.) And you could look at studies of radiation to determine how frequent serious side effects are (I would think there are studies of this). Probably couldn't give you a complete answer but might give you a sense. I'm not a scientist or a researcher, so I am just speculating about what kind of information is out there, but I have a feeling you might be able to get to something that approximates an answer to your concern even if there isn't a specific study on point.

I had radiation after my lumpectomy and my gut reaction is that in my case it was the right thing to do. In addition to talking to my doctors, I spoke to several friends in the medical field (including a radiation oncologist, a gynecologist, a radiologist, an anaesthesiologist, a general surgeon and a naturopath) to get their thoughts and to have a better sense of the risks of side effects v the risk of recurrence and, while I was nervous about radiation, I was comfortable that the risk of recurrence (possibly invasive, therefore possibly involving more extensive treatments and/or risk of not being able to beat the recurrence) without it would be a bigger concern for me personally than concerns about the risk of radiation would be. For me, it was worth the potential serious side effects of radiation to reduce my risk of recurrence, for even if the recurrence didn't ultimately kill me, it would potentially subject me to much more difficult treatments in order to survive. Probably the fact that I was 40 at the time of diagnosis (so hopefully living 40 or 50 more years) and because my 46 year old cousin was battling a very aggressive Stage IV cancer at the time impacted my decision-making process. Throughout all of my decisions regarding treatment, I've wanted someone to be able to tell me what MY recurrence risk is, what MY risk of side effects from treatments are, etc., and I have been very blessed to have worked with excellent doctors who did their best to provide those answers, but I realize that there are no guarantees and there is no one right decision for everyone (even for everyone with exactly the same diagnosis).

Anyway, I just wanted to share my thoughts. I know I got a little off the subject there. Best wishes to you all. Thank you for this thread!

BikerBabe1 · November 2010

I am bumping this thread because I think the discussion on recurrence rate vs. outcome rate is important information, not because there is a right answer.

I was diagnosed in November. I have not been satisfied with the medical professionals' (those that I have met with) pronouncement that while, for DCIS, a "lumpectomy has a higher rate of local recurrence, the outcome survival rates are the same as for a mastectomy." This hasn't made sense to me given that when a local recurrence happens, 50% is DCIS and 50% is IDC. After conducting my own research (I don't have a research background but I can read a study), I identified this "low statistical power" issue related to the outcome rates. I have not yet made my decision about whether I will have a lumpectomy with radiation or a mastectomy. I am thankful I have a choice. And I am thankful that I can make this decision based upon all the information that is available to me about risk and side effects---including risk related to study design questions etc . . .

I also appreciate each day that this decision is very personal to me--to each of us--and that there is no good choice and that it is truely tough.

Thank you redsox for starting this thread. It continues to be on my mind as I consider my decision.

redsox · December 2011

bumping for dancetrancer

dancetrancer · December 2011

Thank you redsox - very helpful!

Recurrence rates vs. survival rates

Comments

Categories