An Expert Interview With Laura J. Esserman,
October 30, 2013
Medscape: I understand that the Viewpoint article inJAMA [5]is the forerunner of a longer paper that will expand on your recommendations.
Dr. Esserman: Yes; that is currently under review.
Medscape: Could you elaborate on what you hope will happen as a result of your recommendations?
Dr. Esserman: What we are trying to do is take the conversation from a contentious process about being for or against screening to being more about understanding the nature of what happens when you screen. When you screen, you find the spectrum of cancer, but you will necessarily find more of the more indolent cancers. This is something that we really did not anticipate when we started screening.
Initially, we observed that early-stage cancers had much better outcomes than late-stage cancers. It was an easy leap to the assumption that if we could just catch cancer earlier, we would prevent more aggressive treatment and we would prevent cancer deaths. The first big screening opportunity was for cervical cancer -- which turned out to be, in large part, a slower-growing cancer with not quite as wide a spectrum as the other diseases that we see. Not only did we reduce the incidence of cancers that were killing people, so the mortality rate dropped, but we also detected precancerous lesions. So we reduced the incidence of invasive cancer as well as late-stage disease.
But as we started to screen for other cancers, such as prostate and breast cancers, we found that there is not just one kind of cancer. There are fast-growing and slow-growing types, but initially they all were treated the same. We did not have a way to distinguish indolent from consequential disease. That is a problem that everyone understands in prostate cancer. But we are trying to point out that it happens in almost every screening condition -- that this is a part of screening.
So we want to make screening better, to focus on how to ensure that we recognize cancer while also developing the tools to reassure people that when we detect the indolent kind of cancers that are not the kind of aggressive cancers that are going to kill people, that we won't overtreat. That was our key message, and we have a number of recommendations that we believe will help solve this problem.
I think the US Preventive Services Task Force (USPSTF) guideline on lung cancer screening[8] is perfect timing for all this, because it explains that you have to focus on a high-risk population and cannot go after every lesion that you see. In fact, nodules of 4-6 mm diameter in the lung have about a 0.15% chance of being malignant.[9] There are other ways to manage such lesions rather than just going in and biopsying everything. We should start thinking about having thresholds and try to be a little more thoughtful about what is the right balance.
Many people will say, "What about that one person you miss?" but if you are following someone you are not going to miss it, and most of these on the lower end of the spectrum are not rapidly growing, so you are not putting patients' lives at risk. We want people to say, "Wait a minute; we can do things a different way."
One of my coauthors, Brian Reid, cites a perfect example in Barrett esophagus. People have been very aggressive in screening and biopsying and then treating these patients, because they thought that Barrett esophagus was a complication of reflux that progressed to esophageal adenocarcinoma. However, this wasn't effective in reducing the incidence or mortality of esophageal adenocarcinoma, because most people with Barrett esophagus don't progress or die of esophageal cancer. Our research challenge is to determine the biological differences between slowly progressive disease and nonprogressive conditions and the time window for early detection. Rapidly progressive disease may be difficult to prevent with screening. That is another challenge.
Medscape: Could you elaborate further on the type of conversation that you want to start with patients?
Dr. Esserman: What we are concerned about -- and I know this because we have run studies on ductal carcinoma in situ (DCIS)[10] -- is that the public has to be aware that the word "cancer" does not mean, "I'm going to die tomorrow, I have to have something done in 2 weeks." That is actually very rare. The public has to understand that there is a conversation. As we learn more, there are options to do things differently. Dr. Thompson, for example, offers men with Gleason grade 3 prostate cancer a watch-and-wait approach. He has just enrolled his 1000th patient and has excellent results. If there is no conversation about this, then none of things are even going to be successful.
Medscape: The public has been educated for 30 years that regular screening is necessary. There were hysterical responses when the latest USPSTF mammography guidelines[11] and prostate cancer early detection guidelines[12] recommended less screening. There was also enormously negative reaction in some professional quarters. It seems to be difficult to persuade a large proportion of people to change. Do you really think your recommendations can start to change that?
Dr. Esserman: That is what we are hoping for, so that people will not be panicked when they see the recommendations that screening should be every other year, for example. There are a lot of reasons for that. We can start with the baseline of every other year, which is what every other country does, and then look for populations that might have specific benefits for more frequent screening, and this could be accomplished with a risk-based screening approach. We should understand the concepts of risk-based screening.
People shouldn't say that it's terrible to change this. They should consider that when they buy a product like a computer, they expect that to change and improve every year. So why in the world are we content with doing what we did 30 years ago as far as screening is concerned?
Medscape: How do you see improvements in your own field of breast cancer?
Dr. Esserman: For example, if a woman who is a BRCA mutation carrier chooses to be screened, we would screen her every 6 months. We want to start to integrate some of the single nucleotide polymorphisms (SNPs) that are now coming out of the Breast Cancer Association Consortium.[13] We are also starting work on comprehensive models to change the way we assess people, to find characteristics that indicate we should start screening at age 40 years. People who are at risk for hormone-positive cancers should be offered prevention strategies. We have started to identify cohorts of people who might be more at risk for estrogen receptor-negative cancer. Now we are thinking about what we can do to actually lower the risk and not just screen.
I lead a big cohort, the Athena Breast Health Network, involving over 150,000 women receiving breast care, and their doctors, specialists, and researchers, at the 5 University of California medical centers, and that is actually what we are in the process of doing.[14] We are trying to put together a comprehensive program that automates risk assessment at the time a woman comes in for screening. We are examining how we put this into more of a learning system where we learn, organize, and start refining our screening and prevention recommendations, asking what we need to do things differently if we want to make things better. We should be excited about wanting to make things better, not feel as though change is a problem.
Medscape: Every time a different guideline has come out recommending reducing the amount of cancer screening. it has generated wide discussion, but the amount of negative reaction suggests that it is difficult to change current attitudes. Some studies have shown that despite guideline recommendations, it is difficult to bring about change in practice.[15]
Dr. Esserman: There are already a number of things that we could fix that are on that low end of the spectrum. For example, for low-grade DCIS, the Oncotype DX DCIS test (Genomic Health, Redwood City, California) shows that these small lesions are associated with a 3.7% risk for cancer in 10 years[15]; that is less than the risk in someone with atypia. These probably are atypia, so let's call them that: atypical lesions. There is no reason for us to upgrade the risk. And there is no urgency to treat now when the risk is really over a 10- to 20-year period. That is where we could start. There is plenty of work to do at the edges without touching anything that might be more controversial.
Medscape: So, it is a matter of being able to identify who is really at high risk?
Dr. Esserman: I think that because breast cancer and prostate cancer are common, we have to do a little bit better to understand who is at risk for what type of cancer. That is going to be the next challenge. In a patient at risk for an indolent cancer, screening is not going to make a big difference because the tumor can be excised whenever it shows up. In patients at risk for a very fast-growing tumor, prevention is a better strategy because it is always hard to get it just in time. In the small group at very, very high risk, we should screen very intensively or very early. We just need to take a step back and start thinking about some other strategies and start testing and learning.
Without putting people at harm, there is a lot of room for improvement as well as things that we can leave alone when we screen, to make screening better. Especially, there is no urgency to find low- or intermediate-grade DCIS. A lot of the calcification biopsies we would leave alone, and we would only really make an intervention if there was significant change.
Medscape: Are there any recent advances in molecular biology, such as markers or diagnostic tests, that will make it easier to identify patients at very low risk?
Dr. Esserman: One of the things that Dr. Laura van 't Veer (leader of the Breast Oncology Program at the UCSF Helen Diller Family Comprehensive Cancer Center) and I are working on is trying to validate an ultra-low signature for breast cancer. We are trying to find a reliable tool to classify some of these tumors as extremely low risk. We have a proposed molecular test, based on a new threshold for the 70-gene test in Dr. van 't Veer's original cohorts. We have subsequently shown that the fraction of low- and ultra-low-risk cancers is higher in tumors detected by screening mammography.[16]
We are validating this in a randomized cohort from Sweden, and maybe by the end of the year we will know if that is true; it could be a huge advance. I am calling upon the scientific community and the diagnostic companies to understand that this is an important issue. Hopefully the public will pay attention to it as well.
Medscape: The implications for cost savings must be huge. A recent study calculated that Medicare was spending almost as much on screening for breast cancer as on treatment,[17] and a presentation at this year's ASCO Genitourinary Cancer Symposium showed that the cumulative net cost of prostate cancer overtreatment in men aged ≥ 66 years is $32 million per annum.[18]
Dr. Esserman: We have an article about the costs of screening that will be coming out later this year in the Archives of Internal Medicine. It shows that if we screened more appropriately, it would be infinitely more cost-effective.
Medscape: People get upset when costs are mentioned, fearing that lives are being sacrificed to save money through cutbacks.
Dr. Esserman: It is different if it is billions of dollars. If you compare following the USPSTF guidelines vs not following them, we are not talking about a $100 million difference; we are talking about several billion dollars difference. But because it is not our checkbook, nobody knows and nobody cares. In countries that really control their costs, such as the United Kingdom and The Netherlands, they don't screen the way we do because it is not the best way to use the test scientifically, nor is it the right way to use resources.
We cannot have it all ways. So with the high cost of healthcare, let's not pay for things that don't have value. I'm not saying not to screen or not to pay for it. What I am saying is that we should do less of it and use it more appropriately.
Medscape: I assume that electronic healthcare records would help with automated screening programs.
Dr. Esserman: I think that in the United States, we should think about doing screening more like they do in Europe, where you invite people to screen and you can track them, and you actually have it as a public good and you are really organizing it.
Medscape: You seem very optimistic about this. Do you think the hysterical reaction and opposition to reducing screening that was apparent a few years ago has calmed down now?
Dr. Esserman: I hope there is a change. I don't know for sure, but I think I can say that the comments that I have gotten personally have been much more positive this time, so I think there is a shift.
