‘Epidemic’ of cancer diagnoses attributable to better thyroid imaging

The incidence of thyroid cancers in the United States has nearly tripled in the past 4 decades, especially among women. However, the increase can largely be attributed to detection of low-risk asymptomatic cancers, most discovered with improved imaging.

At 4.9 cases per 100,000 people in 1975 rising to 9.2 per 100,000 in 2002 and to 14.3 per 100,000 in 2009, researchers concluded that “there is an ongoing epidemic of thyroid cancer in the United States,” in a 2014 study that evaluated data from the National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) database.

“It does not seem to be an epidemic of disease, however,” the researchers wrote. “Instead, it seems to be substantially an epidemic of diagnosis.”

Most of the cases in 2009 were papillary thyroid cancer (12.5 per 100,000), and the increased incidence of these cancers among women (from 4.6 to 19 per 100,000) was three times that among men (from 2.2 to 5.9). During the same period, the mortality rate from thyroid cancer has remained stable.

“Although the rate of thyroid cancer has tripled, the rate of death has not changed at all, again indicating that this epidemic of diagnosis doesn’t really reflect an underlying shift in the number of people who are likely to become ill from thyroid cancer,” Barbara Burtness, MD, professor of medicine (medical oncology) at the Yale University School of Medicine and Yale Cancer Center, told Endocrine Today.

Improved imaging

Ultrasound imaging is currently the gold standard in diagnostic imaging for thyroid cancer, which has high survival rates but also a high risk for recurrence.

According to Bryan Haugen, MD, of the University of Colorado Anschutz Medical Campus, nuclear imaging with radioactive iodine, which has been used since the 1940s, is effective in identifying hot and cold nodules.

“These tests can basically tell you if the thyroid likes to take up a lot of iodine. It can explain why a thyroid may be overactive, and it can also assess a thyroid nodule,” he said. “Some nodules that take up a lot of iodine are hot nodules, and those are almost never cancer. Most commonly, thyroid nodules take up less iodine than the surrounding normal thyroid, and those are cold nodules, which carry about a 10% risk of being a thyroid cancer.”

However, most of the new cancers are small so nuclear imaging is not effective for evaluating them.

“That’s where ultrasound came in as sort of the gold standard, it’s probably been going on since the 1960s but really didn’t take off again until maybe the late 1980s or 1990s when we really had better ultrasound imaging,” Haugen said. “We can really see a lot now with ultrasound.”

According to Stephanie L. Lee, MD, PhD, ECNU, an Endocrine Today Editorial Board member, nodules must be relatively large to be seen on nuclear imaging whereas ultrasound can identify much smaller lesions. However, she explained that the first part of nearly every algorithm is to evaluate thyroid-stimulating hormone. If TSH is low, suggesting hypothyroidism, a thyroid scan would be the next test to be performed; if the nodule is functional the evaluation can be stopped. If a nodule is discovered to be a cold nodule, it is recommended that thyroid ultrasound be the next step of evaluation to assess the sonographic risk of malignancy.

Thyroid problems may also be discovered when patients undergo imaging for other reasons.

“The increase in detection is, in part, because we’re doing so many images of other things, and incidentally they’re finding thyroid nodules,” Lee said. “It’s a large proportion of the new nodules that are being diagnosed.”

“We talk about imaging that’s directed specifically to look at the thyroid, and then we talk about imaging that is ordered to evaluate another organ,” Susan J. Mandel, MD, MPH, a professor of medicine at the Hospital of the University of Pennsylvania, told Endocrine Today. “The thyroid is the innocent bystander. In other words, if you order an ultrasound to image the carotid, you can’t image the carotid without imaging the thyroid because it’s right next door.”

Updated guidelines

Updated American Thyroid Association management guidelines released in October now recommend that clinicians use ultrasound images in patients with normal or elevated TSH levels to determine which nodules should be biopsied by evaluating the characteristics of the nodule.

“There are a whole bunch of things we look for to find what we call higher suspicion and lower suspicion nodules,” Haugen, who chaired the guideline task force, said. “It used to be in the old days we’d just [discriminate] based on size. But now we do it more based on the ultrasound characteristics to determine if it is a high-risk nodule or a low-risk risk nodule. Then we decide whether we biopsy or not, based on those characteristics and size.”

The previous ATA guideline, published in 2009, recommended aspirating most nodules larger than 1 cm, as well as some nodules measuring 6 mm to 10 mm in patients considered at high risk for cancer who also have worrisome characteristics identified by sonogram.

Lee said with the rising rates of incidental nodules it is impossible to biopsy every nodule larger than 1 cm since by age 65 years 50% of the population has a thyroid nodule.

“There are not enough endocrinologists in the world to do that,” she said. “This is what led us to rethink our guidelines and ask which nodules are the highest risk. Which ones do we really have to biopsy? It’s made it more complicated. You actually have to think about these nodules; you can’t just biopsy everything. But, on the other hand, it has helped stop us from performing biopsies on a lot of people with benign nodules.”

According to Mandel, who also served on the task force, the new guideline includes recommendations based upon reviews of all of the literature published in the past 5 years.

“The new guidelines have extensively reviewed prior literature that has examined sonographic patterns rather than individual features,” she said. “We tried to make a tool that was accurate and easily accessible and applicable to an endocrinologist or a thyroidologist in practice. As opposed to describing individual features — what is or what isn’t present — we tried to say, ‘Here, these are pictures of nodules that fall into five different general categories.’”

According to Mandel, these sonographic patterns can be associated with the risk for malignancy, and the new guideline covers five categories, including cyst, very low suspicion, low suspicion, intermediate suspicion and high suspicion. For example, a pure cyst would be the benign pattern.

“[Simple cysts] are not very common,” Lee said. “Almost all nodules have internal structures, but if you have one that is purely fluid, that does not need to be biopsied at all.”

The second category, “very low suspicion,” encompasses nodules with less than a 3% chance of being cancerous.

“It’s a pattern that we call either spongiform or mixed-cystic solid, where the solid part of the nodule still looks like a sponge and is not calcified,” Mandel said.

The third category, “low suspicion,” has a 5% to 10% chance of malignancy.

“These are what we call predominantly solid isoechoic or hyperechoic nodules with no calcification,” Mandel said.

A nodule of “intermediate suspicion” pattern would pose a 10% to 20% risk for malignancy, which is most commonly a hypoechoic solid noncalcified nodule with smooth regular borders.

The final category, “high suspicion,” is associated with a more than 70% change of cancer.

“This includes a number of different types of ultrasound findings,” Mandel said. “It includes hypoechoic or dark nodules with irregular borders as opposed to smooth borders, hypoechoic solid nodules with small calcifications called microcalcifications and nodules associated with neck lymph nodes that are abnormal on ultrasound.”

Increased cancer incidence

According to Burtness, one of the main reasons for the increase in number is aggressive diagnosis, with the increases primarily driven by smaller cancers.

“There’s been good data to show that papillary thyroid cancer is present at autopsy and the person never became symptomatic from it in their lifetime,” she said. “With the increasing use of high-resolution ultrasound and fine-needle aspiration and molecular diagnostics, there is a more aggressive pursuit of indeterminate lesions.”

“We are facing an epidemic of diagnosis in thyroid cancer,” Juan P. Brito, MBBS, assistant professor of medicine at the Mayo Clinic in Rochester, Minnesota, told Endocrine Today. “Now that we know where all these new cases are coming from, we can design strategies to identify patients with thyroid cancer who can benefit from our treatment without condemning other patients to unnecessary tests, treatment, suffering and costs.”

In a study designed to evaluate changes in the incidence of thyroid cancer and mortality, Brito and colleagues found that the number of new cases of diagnoses during the past decade can be attributed to low-risk asymptomatic cancers. In the past decade, half of the thyroid cancer cases were diagnosed in people without symptoms of thyroid disease, according to Brito.

“We are spotting more cancer, but they are cancers that are not likely to cause harm,” he said. “Their treatment, however, is likely to cause harm, as most thyroid cancers are treated by surgically removing all or part of the thyroid gland. This is a risky procedure that can damage a patient’s vocal cords or leave them with lifelong calcium deficiencies.”

Brito added that there is no clear benefit in routinely looking for thyroid cancer lesions in asymptomatic patients, and clinicians should not routinely order a thyroid ultrasound in patients with abnormal thyroid function tests.

Surveillance vs. treatment for low-risk cancers

According to Burtness, there are no clear benefits to treating subcentimeter, incidentally discovered cancers if there is no evidence of extrathyroidal extension or metastases.

“Once they are diagnosed, there are many different ways that the patient can be managed,” she said. “Some of the risks could be mitigated if different treatment choices were made.”

Patients treated with total thyroidectomy have an increased risk for speech and swallowing problems, hypoparathyroidism and neck pain.

“The first question being asked for these small lesions is whether there is any advantage of total thyroidectomy over lobectomy?” Burtness said. “A previous study using data from the SEER database showed that for the lower-risk cancers, there was no survival benefit for total thyroidectomy over lobectomy.”

The use of remnant ablation also comes into question, and according to Burtness, there is “no firm proof that it improves survival in low-risk disease.”

The typical approach of surgery, remnant ablation and supratherapeutic thyroid replacement in people younger than 40 years with a subcentimeter lesion and favorable demographic characteristics, can lead to adverse outcomes that might be avoidable with less scrutiny, according to Burtness.

“If you’re going to do the alternative [to surgery] — surveillance with ultrasound every 6 months — there are some data starting to emerge that more ultrasound abnormalities will turn out to be false-positives and not actually a progression to cancer,” Burtness said. “The state of the current evidence is if you’re not going to intervene surgically right away, most people would recommend that you do surveillance ultrasound until you establish stability of the lesion and then back off on the frequency of ultrasound. There’s still a lot of work to be done to figure all of this out. Maybe we’ll learn molecular profiles that can help us sort out the patients who are unlikely to progress.”

Similarly, researchers are questioning the value of ultrasound imaging in patients after surgery.

In a recent study published in Thyroid, researchers concluded that postoperative surveillance every 3 years after 5 years of annual surveillance is cost-effective in patients with low-risk papillary thyroid cancer compared with annual surveillance. Continued annual surveillance cost $5,239 and yielded 22.49 quality-adjusted life-years. Switching to 3-year surveillance cost $2,601 less and was associated with 0.01 fewer quality-adjusted life-years than annual surveillance. An incremental cost of $260,100 per quality-adjusted life-year was found with annual surveillance.

“Tapering postoperative surveillance to 3-year intervals after 5 years of annual surveillance is more cost-effective than annual surveillance for patients with low-risk [papillary thyroid cancer] with excellent response to initial therapy,” the researchers wrote. “Given the very favorable prognosis in this subset of patients, and the added costs of perpetual annual surveillance with essentially equivalent outcomes, clinicians could consider using a less frequent long-term follow-up interval.”

Lee said she believes that the rising incidence of thyroid cancer has “brought the question to the forefront of what do we do with these small low-risk cancers.”

“Do we need to diagnose them? Do we need to treat them? Do we need to do surgery? It’s bringing up a huge area of research to the forefront,” she said. – by Amber Cox

• References:
• Brito JP, et al. Thyroid. 2015;doi:10.1089/thy.2014.0594.
• Cooper DS, et al. Thyroid. 2009;doi:10.1089/thy.2009.0110.
• Davies L, Welch GH. JAMA Otolaryngol Head Neck Surg. 2014;doi:10.1001/jamaoto.2014.1.
• Haugen BR Md, et al. Thyroid. 2015;doi:10.1089/thy.2015.0020.
• Wu JX, et al. Thyroid. 2015;doi:10.1089/thy.2014.0617.

• For more information:
• Juan P. Brito, MBBS, can be reached at Mayo Clinic, 200 1st St. SW, Rochester, MN 55905; email: brito.jaun@mayo.edu.
• Barbara Burtness, MD, can be reached at Yale Cancer Center, 35 Park St., New Haven, CT 06511; email: Barbara.burtness@yale.edu.
• Bryan Haugen, MD, can be reached at University of Colorado Anschutz Medical Campus, 1635 Aurora Ct, Aurora, CO 80045; email: Bryan.Haugen@ucdenver.edu.
• Stephanie L. Lee, MD, PhD, ECNU, can be reached at Boston Medical Center, 88 E. Newton St., Endocrinology Evans 201, Boston, MA 02118; email: stephanie.lee@bmc.org.
• Susan J. Mandel, MD, MPH, can be reached at the Hospital of the University of Pennsylvania, 3400 Spruce St., Philadelphia, PA 19104; email: Susan.Mandel@uphs.upenn.edu.

Disclosures: Brito, Burtness, Lee and Mandel report no relevant financial disclosures. Haugen reports receiving research funding from Genzyme and Veracyte and a one-time honorarium.

Will the use of molecular diagnostics begin to outweigh the use of ultrasound characteristics in decision making about thyroid nodules?

Assessing thyroid nodules for malignancy before surgery remains difficult in many cases. Molecular diagnostic testing has recently emerged as a robust strategy to improve preoperative risk stratification.

For indeterminate cytology nodules, the Afirma gene expression classifier (Veracyte) has demonstrated a negative predictive value sufficient to inform surgical recommendations. Suspicious sonographic features are less likely in nodules with indeterminate cytology and have not been found to be associated with Afirma results. In a recent study, 12% of malignant nodules (with indeterminate cytology) lacked any suspicious sonographic feature. Combining multiple suspicious sonographic features identified all malignant nodules, but these almost all possessed BRAF or NRAS mutations. Expanded assessment of genetic alterations using next-generation sequencing (ThyroSeq v2, CBLPath) may offer even more accurate preoperative testing.

While suspicious sonographic features are associated with malignancy, most do not predict patient outcomes. In contrast, BRAF V600E is associated with papillary thyroid cancer mortality and is independently predictive of recurrence. New molecular insights will continue to advance our understanding of papillary thyroid cancer behavior and may become important factors in the decision to perform initial lobectomy vs. total thyroidectomy or lymph node dissections.

There is increased emphasis on sonographic characteristics to determine the need for fine-needle aspiration biopsy. However, fine-needle aspiration is recommend for nodules bigger than 1 cm if they have an “intermediate” or “high” suspicion sonographic pattern, providing a wide estimated malignancy risk of 10% to 90%.

The information obtained from molecular assessment has already begun to outweigh that of sonographic features for indeterminate cytology nodules and will only become more important to the clinical management of thyroid nodules and cancer. Other molecular modalities, such as miRNA assessment, appear promising and additional avenues, such as epigenetic changes, remain reservoirs for further discovery.

Trevor E. Angell, MD, is an endocrinologist at Brigham and Women’s Hospital. Disclosure: Angell reports no relevant financial disclosures.

Molecular markers will not change the use of ultrasound.

Ultrasound is what guides me to use molecular markers or not. Ultrasound is still the most important test we have for evaluating patients with thyroid nodules. It’s the first stratification tool we have to use if a nodule is suspicious or not, and we can tell with a reasonable amount of certainty. The new American Thyroid Association guidelines give us criteria to differentiate high risk vs. low risk and even the intermediate risk nodules. When I’m evaluating a patient, I ask myself what I’m going to do with the nodule if it comes back benign, indeterminate or malignant. If I have a nodule that has a 10% risk for malignancy just based on its ultrasound appearance, I’ll do a fine-needle aspiration. If the patient is trying to avoid surgery, I will collect genetic testing at the time of initial fine-needle aspiration and then send it for molecular testing if it is indeterminate on cytology. However, not all molecular panels rule out the possibility that a nodule is malignant. I personally use the Afirma gene expression classifier (GEC; Veracyte) because it is designed to identify benign nodules. It performs very well in a low-risk nodule for ruling out cancer. If the Afirma GEC comes back benign, I do not need to perform any additional diagnostic testing on that nodule.

Ultrasound helps with the initial evaluation of the nodule and helps predict the probability of malignancy. I determine whether I will use the GEC after a nodule comes back from cytology as indeterminate. If the nodules look low risk by ultrasound, I will want to do a molecular test because I want to try to keep those patients out of surgery and the Afirma GEC test can help solidify that stance. On the other hand, if a nodule has a high-risk appearance by ultrasound and indeterminate cytology, I do not need molecular testing to tell me how to manage that patient. Their risk of malignancy is already high based on the sonographic pattern, and surgery is the next step in the evaluation.

Jennifer Sipos, MD, is associate professor of medicine and director of the Benign Thyroid Disorders Program in the division of endocrinology, diabetes and metabolism at The Ohio State University School of Medicine. Disclosure: Sipos reports no relevant financial disclosures.

• References:
• Alexander EK, et al. N Engl J Med. 2012;doi:10.1056/NEJMoa1203208.
• Alexander EK, et al. J Clin Endocrinol Metab. 2014;doi:10.1210/jc.2013-2482.
• De Napoli L, et al. Thyroid. 2015;doi:10.1089/thy.2015.0311.
• Haugen BR, et al. Thyroid. 2015; doi:10.1089/thy.2015.0020.
• Keutgen XM, et al. Clin Cancer Res. 2012;doi:10.1158/1078-0432.CCR-11-2487.
• Labourier E, et al. J Clin Endocrinol Metab. 2015;doi:10.1210/jc.2015-1158.
• Nikiforov YE, et al. Thyroid. 2015;doi:10.1089/thy.2015.0305.
• Xing M, et al. JAMA. 2013;doi:10.1001/jama.2013.3190.
• Xing M, et al. J Clin Oncol. 2015;doi:10.1200/JCO.2014.56.8253.
• Zhu QL, et al. AJR Am J Roentgenol. 2015;doi:10.2214/AJR.14.13984.