More is not always better: MRI and the increasing mastectomy rate

“Early detection of cancer saves lives.”

How many times have you heard this statement or something resembling it? It’s a common assumption (indeed, a seemingly common sense assumption) that detecting cancer early is always a Very Good Thing. Why wouldn’t it be, after all? For many cancers, such as breast cancer and colon cancer, there’s little doubt that early detection at the very least makes the job of treating the cancer easier. Also, the cancer is detected at an earlier stage almost by definition. But does earlier detection save lives? Does finding more disease before surgery that wouldn’t have been found before improve outcomes and survival rates?

In the case of breast cancer, a new imaging modality has brought this question to the fore in a pretty contentious way. If there’s one thing about treating cancer, there is always the push, both philosophical and technological, to find ever smaller pockets of tumor. I wrote about this very issue in great detail a while ago, explaining why ever earlier detection is not always the good thing that most people assume that it is, thanks to lead time bias, length bias, and the Will Rogers effect. The main reason that ever earlier and more sensitive tests to detect cancer are not an unalloyed good is primarily the phenomenon of overtreatment that they often cause, which occurs when small malignant lesions that would never have progressed to life-threatening cancer are treated as though they were cancer. The reason it is so hard to resist overtreatment when such lesions are found is because we have no reliable means of predicting which lesions will progress to cancer. Consequently, patients are subjected to what is in retrospect unnecessary surgery

Other than prostate cancer, nowhere does the impetus towards ever more sensitive tests to detect cancer bump into the fear of overtreatment is the debate over the use of MRI for breast cancer screening and for assessing the extent of disease prior to definitive surgery. Indeed, I’ve written about this issue before at least twice as well. Recently, I learned of a study that suggests that my fears about the increasing use of MRI might have turned out to be true. Here’s what I said:

Another issue to consider is an unintended consequence that is likely to come about if MRIs are done for all patients with breast cancer. For this discussion, remember the term “stage migration.” Most doctors are not going to order an MRI on just the opposite breast in a patient with newly diagnosed cancer. Indeed, prior to the publication of this paper, there was a growing tendency to order MRI for breast cancer to evaluate the extent of disease in the involved breast. That means that these patients are probably all going to get bilateral breast MRI, and this is where the problem of the extreme sensitivity of MRI comes in. There is little doubt that MRI will be more likely to find additional suspicious lesions in the breast with the known cancer an/or to indicate that the cancer goes further than it appears on mammography. As I asked a few days ago (rhetorically, of course), how can this not be good thing? Trust me, there’s a way. One unintended consequence of using MRI willy-nilly to evaluate extent of disease in breast cancer could well be an increase in the percentage of breast cancer patients undergoing mastectomies as opposed to lumpectomies.

What is important to remember here is that the survival rates for patients undergoing lumpectomy and radiation are the same as those for patients undergoing mastectomy. It is true that patients undergoing lumpectomy have a higher rate of local recurrence in the treated breast, but there is no survival difference because those recurrences can be salvaged with mastectomy. Also remember that these results have been validated over three decades with large clinical trials using the “primitive” technology of mammography and ultrasound, along with clinical assessment, as the only means of preoperative assessment of extent of disease in the affected breast. There is no evidence yet that preoperative MRI in any way leads to improved survival in breast cancer. However, there is evidence suggesting an unintended consequence of the increasing infiltration of MRI into the preoperative evaluation of women with breast cancer.

I’m referring to an abstract of a study that is scheduled to be presented in a week and a half at the American Society of Clinical Oncology (ASCO) meeting in Chicago, probably the largest meeting of clinical and academic cancer physicians and researchers in the world, with some 20,000 descending on the Windy City to soak up the latest results of cancer clinical trials. It’s definitely a talk that I will have to make an effort to attend, Trends in mastectomy rates at the Mayo Clinic Rochester: Effect of surgical year and preoperative MRI, a study which after the public release of abstracts to be presented at the meeting has been reported in the news thusly:

Signaling the reversal of a long decline in the use of mastectomy for breast cancer patients, a new study indicates more women are opting for the operation in part because more patients are getting MRI scans that can find additional tumors.

The marked change in care in recent years suggests a deepening of the emotional dilemma that many women face when choosing between mastectomy and more targeted treatments that can preserve the breast.

The study of 5,463 Mayo Clinic breast cancer patients found an increase of 13 percentage points in the use of mastectomy between 2003 and 2006. That rise coincided with a doubling in the use of MRI for such patients, according to the report, which was released Thursday in connection with an American Society of Clinical Oncology meeting to be held in Chicago later this month.

The rebound in mastectomy rates is sparking a debate among oncologists about whether the enhanced sensitivity of MRI, or magnetic resonance imaging, is always a good thing.

Indeed it is. In fact, because there was no strong evidence supporting its use preoperatively and because I believed that nonselective use of preoperative MRI would lead to a lot of overtreatment in the form of mastectomies, it was a long time before I finally caved in as a result of the pressure placed on clinicians by this study and acquiesced to the trend. Now, depending on the quality of the evidence presented in the Mayo Clinic abstract, there may be some ammunition to argue that the routine use of MRI may not be always be justified in the preoperative evaluation of cancer:

In a multivariate model adjusted for age, stage, contralateral breast cancer, and density, both MRI (Odds Ratio (OR): 1.7, p<0.0001) and surgical year (compared to 2003; OR: 1.4 for 2004, 1.9 for 2005, and 1.7 for 2006; p<0.0001) were independent predictors of mastectomy.

True, this is a correlative study, and correlation does not necessarily equal causation. However, inferring causation from this correlation is not unreasonable. The purpose of doing MRIs before surgery is to determine whether or not there are multiple tumors or extension of tumor beyond what can be seen on mammography or ultrasound. Because it is so sensitive, it will often find what it’s looking for. Once again, though, remember the Will Rogers effect and lead time bias. There may be evidence of extra disease there by MRI, but we know from mountains of data that using mammography and ultrasound alone to evaluate extent of disease has led to survival rates every bit equivalent of rates achieved with mastectomy. We also know from detailed pathological studies of mastectomy specimens that there is not infrequently microscopic tumor in the breast more than 2 cm away from the primary tumor. With a high rate of success, radiation “cleans” up any microscopic or tiny macroscopic foci of disease that weren’t included in the lumpectomy. It is unclear what adding MRI can do to improve upon this, but it is quite clear what preoperative MRI can do to influence a clinician’s and woman’s decision. If there is more extensive disease than thought, or if there are additional foci of possible disease (remember, without a biopsy it can’t be concluded that what the additional “spots” are in fact tumor), then often the surgeon has little choice but to recommend a mastectomy. As Dr. Seema Khan puts it:

In some cases, the additional tumors that MRI reveals might be effectively treated with more limited therapies, such as removing the initial lump followed by radiation and hormone treatment, said Dr. Seema Khan, director of the program for early detection and prevention of breast cancer at Northwestern Memorial Hospital. But once a scan uncovers additional tumors, many patients instinctively opt for a full mastectomy.

“There’s a huge question of whether we’re being led down a path of overtreatment by routinely using MRI,” said Khan, who was not part of the Mayo research team.

This is, of course, what I’ve been saying and thinking for some time, ever since MRI became increasingly pervasive. Even now, I (and every other surgeon who treats breast cancer) continue to grapple with the question of what the appropriate role of MRI is. Clearly more studies are needed to clarify the situation, but what do we do in the meantime while those studies are being done? What do we recommend to the women who are our patients? There’s no doubt that MRI is useful in women with a high risk of developing cancer due to family history or BRCA mutations. It also appears to be true that preoperative MRI is useful in young women with dense breasts who develop cancer. Moreover, we as surgeons focus on overall survival rates and forget that a local recurrence in the breast can be devastating to women, even if it is ultimately successfully treated. There are women who would rather sacrifice their breast to decrease the risk of local recurrence to very low levels (it’s never zero) even fully understanding that it won’t increase their chance of long term survival. It’s an individual decision, but the overdiagnosis that MRI can produce is a powerful force pushing women to have mastectomies.

One thing that somewhat confounds this study, though, is that there was a trend towards more mastectomies in women who did not undergo MRI. True, MRI emerged as an independent predictor of mastectomy in a multivariate analysis, but the study isn’t bulletproof. For one thing, in my own anecdotal experience, for reasons that are unclear to me, more women with breast cancer seem to choose mastectomy than before. This could be confirmation bias speaking on my part, but this study suggests that maybe it isn’t. The reason for this increasing preference for mastectomies is not clear to me at all, but there does seem to be a belief out there that somehow doing more surgery will make a cure more likely. I won’t pretend that this belief isn’t shared among some surgeons, evidence be damned.

What this study and the ongoing controversy over the proper role of MRI in the preoperative evaluation of breast cancer is that in cancer, increased sensitivity to detect ever smaller foci of cancer always comes with a price. Always. The benefit is that it can detect disease that may not have been detected on other imaging modalities before, allowing for treatment of the tumor before it becomes advanced. There are cancers that simply do not show up well on mammography or ultrasound until they are quite large. The price, of course, is overdiagnosis and over treatment. The difficult challenge we as clinicians and scientists have is to figure out the proper balance between the two and exactly when the use of this new technology is likely to produce the best chance of survival at the least cost.

After I’ve heard talk, perhaps I’ll comment further.