Perspectives on Overdiagnosis, Overtreatment of Subclinical Thyroid Cancers

Louise Davies, MD, MS, of the VA Outcomes Group and the Dartmouth Institute for Health Policy and Clinical Practice, is an expert on thyroid cancer overdiagnosis and active surveillance. She spoke with Cancer Network about overdiagnosis and overtreatment of subclinical thyroid cancers.

Dr. Davies is an associate professor at the Geisel School of Medicine at Dartmouth and chief of otolaryngology–head & neck surgery at the Veterans Affairs Medical Center in White River Junction, Vermont. As a Fulbright Global Scholar for 2017–2018, she recently studied thyroid cancer overdiagnosis and surveillance at Kuma Hospital in Kobe, Japan and Oxford University in England. She also served on an expert task force about thyroid cancer screening after nuclear accidents, for the International Agency for Research on Cancer.

-Interviewed by Bryant Furlow

Cancer Network: Most of the rise in thyroid cancer diagnoses over recent decades has involved very small tumors. What has driven the increased rate of these diagnoses?

Dr. Davies: In the US, more technology was introduced for working up thyroid nodules-ultrasound in the 1980s and fine needle biopsy in the 1990s. Even though each introduced technology was touted at the time as a way to prevent unnecessary surgeries, they have coincided with a rise in the number of surgeries. Imaging rates overall have also increased, for example the use of MRI and CT scanning to image other body areas, and these scans often happen to include the thyroid gland, uncovering small nodules.

We have data that strongly suggest that the main important risk factor for detection of small thyroid cancers is access to healthcare, for example people with insurance, and people who use more healthcare in general.

In South Korea, things started going crazy with the use of thyroid ultrasound, starting in about 2000-and have just gotten completely out of control. The rates of thyroid cancer [there] are 10 times higher than in places that have not experienced the same problems, such as the Nordic countries. I met with South Korean researchers about it recently and although some interventions have induced minor decreases in the rate of thyroid surgery, the trend downward has been modest.

In 1999, the South Korean government instituted a cancer screening program [as part of its National Cancer Control Program] because cancer is the leading cause of death there. The screening program covered gastric, liver, colorectal, breast, and cervical cancers. Thyroid cancer was not on the list. But what happened was, as they instituted this nationwide screening approach and emphasized early detection of other cancers, some places started offering inexpensive retail thyroid-ultrasound screening that was not endorsed by South Korean public health authorities.

Cancer Network: When exactly did thyroid cancer diagnoses start their sharp rise?

Dr. Davies: The incidence started rising in the early 1990s, at least in the United States. When we published our paper in JAMA in 2006, we presented a very simple, straightforward analysis with a single take-home point about overdiagnosis-so it broke through and got people thinking about overdiagnosis as the potential cause. What was so striking was that the incidence had tripled but mortality was completely unchanged from what it has been for 30 years or so.

That picture has become a little more complicated in recent years.

Cancer Network: How so?

Dr. Davies: I’ve been following the trends, and one thing that has been happening is mortality from thyroid cancer looks like it’s going up a tiny bit. For quite a while, that depended on how you analyzed the numbers because the changes were quite small. Then last year, [Dr. Hyeyeun] Lim et al published a paper in JAMA asserting a clear increase in mortality. The change in mortality was not significant for those diagnosed with early-stage or regional spread, but mortality was increased for those with widespread distant metastasis at the time of diagnosis. They saw that as reason for worry that there is a new risk factor for thyroid cancer and that it’s not all overdiagnosis.

We wrote a letter to the editor because we felt they had not considered all of the possibilities. I do think it’s possible that there is a new risk factor or that people are developing more aggressive disease-higher-mortality disease.

There have been other cancers where we’ve seen overdiagnosis [but] where mortality went up and then back down when the detection rates declined. Prostate cancer is one example. When prostate cancer screening was very popular in the 1980s and 1990s, incidence skyrocketed, and the mortality line went up but then went back down again as screening rates declined. The trend line both rose and fell in about 30 years, so it was unlikely to be some new aggressive form of cancer that came and went from the population.

More likely, it was attribution bias-attributing patients’ deaths to something in their medical record that isn’t really the cause of death. It’s a well-known issue in cancer epidemiology that we always consider in our analyses and try to account for in our inferences.

Cancer Network: Has overdiagnosis led to widespread overtreatment?

Dr. Davies: Thyroidectomy rates have gone up. For thyroid cancer, once you’ve been diagnosed with a small thyroid cancer, if you opt for surgery, then the greatest harm probably comes from removing an entire thyroid gland when the guidelines suggest that’s not necessary. Removing just half is generally appropriate.

Removing half of the gland for appropriate early-stage cancers is guideline-concordant. A key benefit is that most patients won’t need to take thyroid hormone replacement pills. That’s a big benefit. If you take out the whole gland, the patient needs to take a pill every day, and not everybody feels normal on thyroid replacement. It also means regular blood tests, prescription renewals-it’s a different way to live your life. And identifying a person as having cancer has ramifications for psychological quality of life.

With some provisions of the Affordable Care Act potentially at risk of being rolled back, insurers would be able to create policies that charge different rates for people with preexisting conditions, or possibly to exclude them. This means that in the US, if this change occurred, simply having a cancer diagnosis-including if you are on active surveillance for a low-risk cancer-would affect your healthcare costs and insurability.

And of course, there are low but never zero risks with surgery-mainly damage to the voice, due to injury of the recurrent or superior laryngeal nerves, and the risk of permanent hypoparathyroidism when the whole thyroid gland is removed.

Cancer Network: Which patients are candidates for active surveillance instead of surgery?

Dr. Davies: The clinical threshold now, based on data from Japan, is that cancer not exceeding 1 cm and papillary type cancers are potential candidates for monitoring rather than immediate surgery. Additional criteria are that the cancer cannot be located against the recurrent laryngeal nerve or attached to the trachea in a shape that suggests invasion is present or likely to occur, and no nearby lymph nodes have cancer in them. So, basically, 1 cm or less in the middle of the thyroid gland, no lymph node metastases-those are good candidates for active surveillance. 

Older patients are also candidates, because small thyroid cancers grow less often in older people. The ideal candidate is described as a person 60 years of age or older. Adults aged 18 to 59 years are considered appropriate for active surveillance if they meet the tumor criteria. Children younger than 18 years are not candidates. Patients must also be willing to follow up with regular ultrasound checkups, and the people [clinicians, technicians] taking care of them need to feel comfortable doing and/or interpreting ultrasound images. 

The 5-year data emerging from the Fukushima Health Management Survey on the prevalence of thyroid cancer in children and adolescents that has come out in 2017 and 2018 suggests this concept might need to be revisited in the future. The understanding of thyroid cancer in children and adolescents is still emerging, and the Fukushima data will be very enlightening.

An open research question is whether it is safe to monitor thyroid cancers larger than 1 cm but smaller than 2 cm. There are cohort studies under way, and we plan to test this via a randomized clinical trial to compare active surveillance to surgery for early-stage thyroid cancers.

Cancer Network: How frequently must patients come in for ultrasound under active surveillance?

Dr. Davies: It varies. For example, the Memorial Sloan Kettering Protocol requires an ultrasound every 6 months for 2 years, then switches to annual scans. Kuma Hospital in Japan, where active surveillance started, does one every 6 months for the first year, and then switches to [performing scans] once a year unless there are reasons to check more often.

Cancer Network: Are there biomarkers or liquid biopsies for noninvasive monitoring of thyroid cancer?

Dr. Davies: There are not yet approved or standard blood tests to monitor thyroid cancer.

Cancer Network: Your team tried to create a registry for incidentally detected thyroid nodules-“incidentalomas.”

Dr. Davies: Yes, but it didn’t work out. We discovered that you cannot make a registry of incidentalomas in the way we wanted because a good registry must capture all cases, and we discovered that radiologists have widely varying behavior [in terms of] how often they write in their reports that they identified a thyroid nodule. We found inconsistencies in reporting. Also, it’s hard to tell from a medical chart why a physician did or did not take action based on a radiology report finding. We found in this study that if physicians did not act on incidental thyroid findings, they did not reliably document why or how that decision was made.

Cancer Network: What should clinicians know and communicate to patients about thyroid cancer overdiagnosis?

Dr. Davies: When I talk about overdiagnosis, I’m not saying that nobody with early-stage thyroid cancer should have surgery. I’m saying people need to understand the facts for themselves and make decisions that match their personal goals and values. Some people with small cancers do not want surgery, but some do. They should be able to make that decision for themselves, and we should make sure they have the opportunity to understand the risks for themselves. My goal in doing research in this area and talking about it is to help our field as a whole get better at supporting treatment decisions that align with patients’ own value systems and risk tolerance.