Q&A: Practice of Overdiagnosing Breast Cancer

For the past decade, screening mammography has been the widely accepted tool for identifying potential breast cancers as early as possible. But, as diagnosis rates have increased, so have the questions about whether the procedure is pinpointing the right cancers instead of simply more cancers.

In an October New England Journal of Medicinestudy, H. Gilbert Welch, MD, MPH, professor of medicine of Dartmouth University Geisel School of Medicine, analyzed data from 1975 to 2012 from the Surveillance, Epidemiology, and End Result (SEER) program to determine tumor size and breast cancer incidence in women age 40 and older. He and his colleagues determined the fatality rates for years before screening mammography popularity grew and after.

The number of early-detected small breast tumors spiked from 36% to 68%, while the number of large tumors dropped from 64% to 32%. Examined against the fatality rates, the specter of overdiagnosis loomed. Diagnostic Imaging spoke with Welch about the likelihood of overdiagnosis among women with potential breast cancers and what it means for screening mammography efficacy.

What are the overarching clinical implications of your study?

I think the overarching message is that, with mammography, the clear call is that most people appreciate the potential, but it benefits very few. That’s one reason we emphasize the very good news that reduced breast cancer mortality largely reflects better treatments, not screening. Mammography comes with the human cost of treating others for disease that would never have been a bother to them. That’s the overdiagnosis problem. It’s not limited to breast cancer screening. Radiologists will be familiar with it in the context of early thyroid cancer, lung cancer, or prostate cancer screening. It’s a general problem in cancer screenings.

How significant is the amount of overdiagnosis?

I think that’s a value judgment. I think it’s quite significant, as we show. For every woman that has breast cancer destined to become large that is detected earlier, about four are diagnosed with cancer that isn’t going to go forward. And, yet, they’re treated for it.

What impact does overdiagnosis have on treatment?[[{"type":"media","view_mode":"media_crop","fid":"53051","attributes":{"alt":"H. Gilbert Welch, MD, MPH","class":"media-image media-image-right","id":"media_crop_9523821813235","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"6601","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"height: 167px; width: 200px; border-width: 0px; border-style: solid; margin: 1px; float: right;","title":"H. Gilbert Welch, MD, MPH","typeof":"foaf:Image"}}]]

There are implications on a bunch of levels. No one would want to be treated for cancer unnecessarily. We do have to recognize that’s a side effect of looking early for cancer. Basic treatments can have side effects: complications and death rates. But, it happens. So, treatment is something that no one wants to take lightly in the context of cancer. It’s also important to think about how overdiagnosis produces misleading feedback. Women who have screenings and treatment for cancers see themselves as breast cancer survivors, naturally. It’s a cycle – the more overdiagnosis screening tests cause, the more people will have to be treated for the disease and will see themselves as survivors. Then, the more popular the screening becomes. The feedback is exactly wrong because it ultimately leads to the understanding of the problem of overdiagnosis as a benefit rather than people experiencing more consequences and harm.

Getting a diagnosis of cancer has important psychological implications. Suicide rates have been shown to go up with people who have new diagnosis – even low-risk prostate cancer. It’s not common, but it happens. It can be a devastating diagnosis. Bankruptcy is a real problem following cancer. So, we want to be really careful about not looking so hard for cancer that we tell way too many people that they have it.

What impact does overdiagnosis have on the relevance and efficacy of screening mammography overall?

This is a bigger problem than just with mammography. You have to give radiologists credit that they’re beginning to grapple with this issue in lung and thyroid cancers. They’re taking a more considered and prudent approach to try to minimize overdiagnosis. That approach is one of not looking so hard for small abnormalities so that those you do find have a higher threshold for biopsy and the wait-and-see approach to monitor whether they’ve grown. I want to be clear that radiologists are actually very sensitive to this general problem. They’re thinking about and working on it with the same attention being paid to the problem with early lung or thyroid cancer. The general notion that we’ve put out there is the best test is the one that finds the most cancers. But, when it comes to overdiagnosis, the best test is the one that finds the cancers that really matter.

Are there recommendations for what can be done to avoid overdiagnosis?

First, don’t look so hard for cancer. Recognize the efforts that provide gentler screenings and screenings less frequently. Those are recommendations that reflect the interest to mitigate the problem of overdiagnosis. I think making sure of the diagnosis is a valuable use of time. It’s really beginning to be appreciated by radiologists in other venues, and the same is probably needed in breast cancer. If you have a small abnormality, the question shouldn’t be how fast do you biopsy, it should be how can you check another film in three to six months to see if that abnormality has grown.

What other take-away messages do you have?

The general message is that screening has a mixture of benefits and harms. The process of looking for a cancer early is not the best one at full throttle. Moderation is called for. You must also recognize that value judgments are involved. No one can give a single right answer. That’s important because having screenings should be a choice for patients – not a public health imperative. Not something that everyone has to do.

I think we’re in the middle of a paradigm shift and course correction. In medicine, these things take time. I think, with some of the best intentions, doctors originally believed that early detection was the right strategy no matter what, and they looked as hard as they could for early forms of cancer. That is changing and helping. More attention will be paid to minimizing the number of overdiagnoses and more attention will be paid to making sure the patient knows the full story – and that they get a real choice.