March 09, 2018
7 min read

Molecular testing for thyroid cancer can reduce unnecessary surgeries

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Mark Lupo
Mark A. Lupo
Hossein Gharib
Hossein Gharib

Most thyroid nodules are benign, and cytologic specimens collected through fine-needle aspiration can usually be diagnosed as malignant or benign based on their appearance under a microscope. In cases where the status of a thyroid cytology sample is not clear, further testing using molecular markers may be important for ruling out thyroid cancer and preventing unnecessary surgery in these patients. The high negative predictive value of the currently available molecular thyroid tests is extremely valuable, according to Mark A. Lupo, MD, founder and medical director of the Thyroid & Endocrine Center of Florida.

“From a clinical endocrinology standpoint, if you’ve got testing that supports a high probability that the nodule is benign — less than 5% chance that it’s cancer — that, for a clinician, is great,” Lupo said. “You can then take that patient out of the surgery route and just watch over time. From a diagnostic standpoint, that negative predictive value is the first thing we’re looking for.”

Commonly used tests

The two molecular thyroid tests in common use are the Afirma test (Veracyte) and the ThyroSeq test (CBLPath).

The Afirma test has been shown to be reliable across 20 studies, according to the company’s website. Afirma also offers malignancy classifiers to guide physicians in choosing between surgery and observation.

The ThyroSeq test has been in use for thyroid nodules and cancer for more than 10 years. It was launched in 2007 at the University of Pittsburgh Medical Center as a seven-gene panel. CBLPath’s website reports that the ThyroSeq v3 “decreases the largest number of avoidable diagnostic surgeries.”

Data supporting the efficacy of the ThyroSeq v3 were presented at the 2017 American Thyroid Association annual meeting. The double-blind, multicenter, international study found that the ThyroSeq v 3 performed dependably across various types of cancerous and benign nodules and in populations with variable disease prevalence.

According to Hossein Gharib, MD, professor of medicine at the Mayo Clinic College of Medicine and Science in Rochester, Minnesota, and past president of the American Thyroid Association, both of these tests offer a 95% negative predictive value, which means they are highly accurate in ruling out malignancy. However, he added that neither test features a particularly high positive predictive value.

“We like the positive predictive value to be high; in other words, when we do the test and it is positive, we want to say that there is a 90% chance that this is cancer, so let’s go to surgery,” he said. “These tests don’t have that high a probability, so in that sense, they are a bit more limited.”

According to Lupo, the positive predictive value can range substantially between the two tests. For example, he said the Afirma test might indicate only a 40% to 50% chance that the nodule is cancerous.

“That means half of those patients going for surgery still have a benign nodule,” Lupo said. “That’s a little bit frustrating sometimes for patients and clinicians, to have had surgery despite there being no cancer.”

Based on the ThyroSeq test’s positive predictive value, approximately two-thirds of patients who have surgery do have cancer or precancer, whereas about one-third might still be benign, Lupo said.


Because ThyroSeq provides results specific to the mutation or fusion present, the positive predictive value varies based on these results, Lupo said. For example, the positive predictive value may be approximately 50% for RAS, but 99% for BRAF V600e.

In terms of diagnosis, Lupo said, the tests’ greatest diagnostic benefit comes from their highly accurate negative predictive value.

“Diagnostic-wise, the most important thing is the negative predictive value, which is very helpful,” he said.

Gharib said the current molecular tests are always being updated for greater accuracy, and that it is likely the positive predictive value of these tests will improve.

“These tests have been modified, and so the ones that are available now are better than the ones available 5 or 7 years ago,” he said. “They add more genetic information. They built it into the tests, and so the newer versions are a bit more sensitive. It’s highly likely that a couple of years from now we’ll have a couple of good tests.”

Lupo said the two tests are currently used primarily to guide surgical decision-making, adding that in cases where the need and extent of surgery have been established, molecular testing does not contribute anything to the decision. In the future, however, such testing may aid surgical decision-making by providing insight into prognosis, he said.

“If you have a patient, for example, with an abnormal lymph node that you find on ultrasound and you biopsy it and it’s part of the cancer, then you know you need to do a total thyroid removal and lymph node dissection. Molecular testing is not going to change that,” he said. “Another example would be if you had a nodule that was 4 cm or bigger that was indeterminate by cytology, particularly in an older man. You would probably go ahead and do at least a lobectomy on that patient.”

Lupo said that knowledge of the types of mutations present may be valuable in guiding the extent of surgery.

Other tests

In addition to ThyroSeq and Afirma, other molecular tests are currently available in the United States. Interpace Diagnostics has developed a two-part combination test consisting of the ThyGenX oncogene panel and the ThyraMIR microRNA classifier.

“One is a genetic panel screen that has your usual fusions and mutations, such as RAS and RET/PTC, and they’ve added TERT, as well,” Lupo said. “There is also a microRNA panel. It’s more of a classifier, and it recognizes a benign vs. suspicious appearance.”

Lupo said the ThyraMIR panel has shown positive initial data from one study, demonstrating a 97% negative predictive value and a 68% positive predictive value.
“It has not had a further validation multicenter study yet, so we’re waiting on more data.”

Rosetta Genomics also has developed a pure microRNA classifier, the RosettaGX Reveal. According to a company press release, RosettaGX Reveal has a 99% negative predictive value and is the first thyroid test that works on stained fine-needle aspiration smears. It can be used on the same cytology slides that were used in the initial diagnosis.

Afirma and ThyroSeq, you have to take an extra sample from the nodule with a needle, and then put that in a special tube for testing,” Lupo said. “Rosetta has done something interesting — they are able to take the actual slide that the pathologist finds indeterminate or concerning, and scrape that slide and do the testing right off the cellular material that the pathologist was concerned about. That is a neat concept.”


Prognostic uses

Molecular markers may also be used to determine the aggressiveness of a cancerous thyroid nodule. BRAF, RET and TERT mutations detected through molecular testing may indicate a need to treat the nodule more aggressively, according to Gharib.

“In terms of prognostic value, some thyroid cancers have a BRAF mutation, which is known to be associated with tumors that are more likely to recur, more likely to metastasize, and more likely to cause problems,” Gharib said. “The theory is, if you can do this test at the beginning of your treatment program and the test is positive, you must take a more aggressive attitude toward removing, treating and following the tumor. This is something that is currently not used much, but probably will enter into our algorithm for practice in the next few years.”

Gharib added that the RET marker has also been useful in screening the family members of patients with medullary thyroid cancer.

“Patients who have the mutation either have the disease or will get it,” he said. “Those who don’t will never get it, and so it has been used in clinical practice for a long time now and is very helpful. I expect that BRAF and other mutations will also become useful in the coming years.”

A future of customization

The advent of molecular testing for thyroid cancer has had a significant impact on the practice of endocrinology, particularly in terms of guiding the course of treatment for indeterminate nodules. In particular, Lupo said, molecular testing is cutting back on unnecessary surgeries.

“A general rule of thyroid nodule and cancer management is that less is more, so molecular testing has enabled us to confidently take probably half the patients out of a surgical pool who had a test that looked benign with a high negative predictive value, and watch them over time,” he said. “These are patients with indeterminate cytology who would have gone to surgery otherwise.”

Lupo said of the approximately half a million thyroid nodules biopsied each year, about 20% to 30% are indeterminate.

“Historically, those patients were being referred for surgery, and if we can take half of them out of that surgery direction, that’s a major change for endocrinologists,” he said.
As is the case with many areas of medicine, the future of thyroid cancer testing and treatment will likely involve personalized treatment plans based on molecular or genetic testing, Lupo said. Currently, physicians are using clinical risk to individualize treatment plans.

“If you have a patient who has a history of radiation therapy to the neck or a family history of thyroid cancer, you know ahead of time that that person is at higher risk,” Lupo said. “Whereas a person who has a nodule that was found incidentally during carotid ultrasound that looks low-risk, but is indeterminate on biopsy, would not be as much of a cause for worry.”

Lupo said current risk stratification thoroughly considers patient and ultrasound characteristics and appearance of the biopsy under the microscope.

“So even before we get to molecular, we’re beginning to drill down to a more granular level that helps get people risk-stratified away from or toward surgery,” he said. “Once you add molecular prognostic factors that are better understood, it’s going to be a tremendous improvement in the way we manage patients.” – by Jennifer Byrne


Nikiforov Y, et al. Short call oral abstract #5. Presented at: 87th Annual American Thyroid Association Meeting, Oct 18-22, Victoria, BC, Canada.

For more information:

Hossein Gharib, MD , can be reached at 200 First St. SW, Rochester, MN 55905; email:

Mark A. Lupo, MD can be reached at 3050 Bee Ridge Road, Sarasota, FL 34231; email:

Disclosures: Lupo reports he has received research funding from Veracyte and Rosetta. Gharib reports no relevant financial disclosures.