Please, please, please say it's so...

That was my initial reaction today when I saw news stories about a study presented at a breast cancer conference sponsored by a number of leading organizations with a professional interest in the diagnosis and treatment of breast cancer.

The headlines were pretty clear, to the effect that the study showed the value of screening mammography in women between the ages of 40-49. The accompanying stories suggested that this research essentially repudiated the recommendations of the United States Preventive Services Task Force initially published in November 2009.

You may remember that event, since it created a huge amount of media and public interest when the Task Force suggested that screening mammograms to find breast cancer early should not be done routinely in women between the ages of 40 and 49. This was a change from its prior recommendation, and was in conflict with the opinions of the American Cancer Society and other organizations which conitinued to endorse routine breast cancer screening in this age group.

There has been a lot of water under the dam since then, and there have been additional scientific studies reported about the value of screening mammograms, some of which support breast cancer screening and others which do not.

Enter the reports this week that a study from Michigan suggests that screening mammograms and breast self-examination in women between the ages of 40 and 49 results in earlier diagnosis and less disfiguring treatment.

When I read the headlines and the news reports, I quickly came to the conclusion that there was a possible disconnect between what the reporters concluded from the study and what I thought was scientifically valid.  Mind you, I read these things with my own bias/conflict of interest: I have been a supporter of mammograms for women between the ages of 40-49, and believe they save lives in this age group. But my problem here is whether this particular research actually supports that position, as the press reports suggested.

Guess what? I don't think it does.

First, I went to the actual report on the study which was an abstract posted on the website of one of the organizations that sponsored the meeting.

The abstract-which is the official "record" of the study being reported-was pretty straightforward: the researchers examined information from a statewide breast cancer registry that had data on 5903 women diagnosed with breast cancer in Michigan from 2006-2009. 

Focusing solely on the question of how cancers were diagnosed in women between the ages of 40 to 49, the researchers reported that 48.3% of the breast cancers were found by mammograms and 46.1% by palpation (which means someone-we don't know whether it was the woman, her physician or another health professional-felt the lump. It turns out that in press reports there were comments that in this group 90% of the lumps were felt first by the women, and 10% by their doctors). Compared to women ages 50 and older, the younger women had a greater percentage of their cancers found by palpation.

That's it when it comes to specific information in the abstract about women between ages 40-49.

The abstract does go on to note that for all the women in the study-including women younger and older than age 50-those who had their cancers diagnosed by palpation (as opposed to being diagnosed following a mammogram) more often had mastectomies (total removal of the breast) compared to less-disfiguring lumpectomies (where the cancerous tumor is removed), and that a slight but significantly greater number of women who had their cancers found by palpation had more advanced cancer when they were first diagnosed (stage II vs. stage I).

The authors concluded:

"If screening mammography is omitted in this (40-49) age group, cancers when detected may be of a more advanced stage and result in more mastectomies. This study also supports the use of palpation as a method of detection despite recent recommendations against teaching self breast exams by USPSTF."

Yes, but...

Does that justify the headlines that mammograms in this age group actually save lives? Does it really provide evidence that formal breast self-examination is an effective tool in reducing deaths from breast cancer in women between 40 and 49?

Then came step two in my quest to find out how the headlines came to be.

As I suspected, there was a press release which promoted the findings of the study. The headline from the press release said, "Large Michigan Study Suggests Continued Importance of Self-Exams, Annual Mammography in Breast Cancer Detection, Even in Younger Women."

Reading further into the press release:

"The new recommendations (referring to the November 2009 USPSTF guidelines) were controversial because in many cases they were in conflict with more established approaches, which included annual screening mammograms for women between ages 40-49 (true), and encouraged regular self-exams (not quite true)." (coments in parentheses added by me)

The press release went on to quote some additional data that was not contained in the abstract, such as 40% of the cancers found by palpation were in women under the age of 50. There was also additional information that women whose cancers were found by palpation were more likely to undergo chemotherapy compared to women whose breast cancers were diagnosed by mammography.

The press release contains a quote which basically says that the study validates that women who undergo regular mammography screening present at earlier stages and require less aggressive treatment than women who do not (undergo regular mammography screening). It concludes that the study also shows the value of breast self-examination as an "important public health measure."

You might imagine that-as an advocate of breast cancer screening and the value of early detection-I would be dancing in the streets as a result of this report.

Sadly, I am not.

In fact, if anything, I am concerned that the various representations of this study that I have read missed a fundamental point: this study reports information, but is not designed to answer the questions which the headlines and the press release suggest as the appropriate conclusions, namely that mammography and breast self-examination in women between the ages of 40-49 saves lives.

And that's the key point, the fundamental contention between the "pro" and "not-so-pro" breast cancer screening camps: do screening mammograms between the ages of 40 and 49 save lives? And if so, how many lives are saved at what personal cost of potential harms, such as additional biopsies and the possibility that a cancer might be diagnosed that would never lead to a problem in a woman's lifetime?

The American Cancer Society believes screening mammograms do save lives with acceptable risks and complications. Others-including the USPSTF-disagree, although it is important to note that the Task Force did recommend that women in the 40-49 age range have a careful discussion about the benefits and risks of screening mammography with their health care professional, then make a decision to get screened or not to get screened based on what they feel most comfortable doing.

And, by the way, neither the Task Force nor the American Cancer Society believe that routine formal breast self-examination-as opposed to breast awareness-has been shown to be effective in reducing deaths from breast cancer. Both organizations do recommend that if you feel a lump in your breast you should see a health professional promptly-even if you had a screening mammogram the week before.

Past research studies have not shown that a formal, monthly program of breast self-examination saves lives, and this study wasn't designed to show that it does. Both the American Cancer Society and the USPSTF agree on this point.

Yes, more women between the ages of 40-49 in this particular Michigan study were diagnosed by someone feeling a cancer in their breast than was the case for older women. But does that mean they didn't have a mammogram before they felt the cancer? What if some of them had in fact had mammograms and the cancers grew rapidly in between screenings? There is no information in the study to answer that question.

And then there is that pesky "gold standard" issue: do we know from this study that mammograms would have saved more lives in this age group? Or would it have found more cancers without having an impact on life expectancy?

This is a limited, initial brief abstract based on "backward looking" data in a cancer registry, not a randomized trial which is forward looking, where one group is assigned to get screening, and another group is not.

Although it would be easy to conclude that logically it must make a difference, that's not the way we should be doing our science. It is possible that despite the higher stage at diagnosis for the "palpation" group, the overall death rates from breast cancer might not be reduced by screening mammograms. That is the conclusion of some recently reported studies: increased breast awareness and improved treatments may have had the largest impact on reducing deaths from breast cancer. It's not a conclusion I adhere to, but that doesn't mean that I am going to lean on this particular research to support my point of view.

I know all of that is complicated, and the space in this blog is too limited to go into great statistical detail and give a full explanation of all the issues that so many experts on both sides of the screening mammogram debate have raised over the past number of years.

But-despite my bias in favor of screening-I don't think it is wise to draw sweeping generalizations for the public based on a limited set of information that on close inspection does not support the conclusions that some have suggested.

It isn't easy or simple to write this blog when I am critical of organizations and colleagues I know and respect. At the same time, I think it is important that we clearly understand the information that our research provides, and clearly articulate the implications of our science. We have an obligation to report the conclusions accurately and honestly, and not try to stretch the results to meet some personal point of view or goal, no matter how worthy that might be.

The debate about the value of screening mammograms is difficult enough. We need more clarity if we are going to come to the best conclusions that serve the people who look to us for guidance on complicated issues. Over-interpretation of our science is not going to get us where we need to be on this debate which is vitally important to women everywhere.