Molecular Test of Thyroid Nodules With Gene-Expression Classifier Shows Mixed Results

Nancy A Melville

November 05, 2015

ORLANDO, Florida — New reports detailing results of "real-world" use of the Afirma gene-expression classifier (GEC; Veracyte) molecular test — which is designed to help better classify indeterminate thyroid nodules and therefore reduce thyroid surgeries — show a mixed bag of results.

These include some in conflict with published reports of the test's strong negative predictive value, and one suggesting that the GEC test may in fact be having an opposite effect of increasing surgeries.

The Afirma GEC molecular test is one of a number of such products available in the United States.

"The presence of molecular testing itself may not substantially reduce the number of unnecessary thyroidectomies but rather escalate the usage and cost," senior author of one paper, Allen Ho, MD, of Cedars-Sinai Medical Center, in Los Angeles, California, told Medscape Medical News following presentation of his results at the recent 2015 International Thyroid Congress and Annual Meeting of the American Thyroid Association (ITC/ATA).

Molecular Testing Means Pathologists Nuance Diagnosis

In comparing a total of 4292 fine-needle aspirations on indeterminate thyroid nodules before and after the introduction of the GEC in July 2013, Dr Ho and colleagues found no significant differences in the rate of surgeries among patients with indeterminate (Bethesda III or IV classified) nodules.

The incidence of nodules classified as indeterminate, meanwhile, increased significantly following the introduction of the GEC, while the incidence of benign nodules significantly decreased.

"In the absence of other notable factors, these data suggest that [GEC] affects the distribution of Bethesda diagnoses — it may cluster fine-needle aspiration interpretation to Bethesda III/IV, where molecular testing can be used," said Dr Ho and colleagues.

While the Bethesda classification system indicates that Bethesda III diagnoses should make up no more than 3% to 6% of all fine-needle aspirations, Cedars-Sinai's rates are much higher, Dr Ho said. "We and many other institutions are clearly above this recommended metric."

He knew of no other studies reporting similar trends of cluster interpretation involving thyroid nodules stemming from GEC but said there have been similar reports in other diagnoses, such as cervical cancer.

"It has been shown in the cervical-cancer literature that cytopathology 'clustering' of indeterminate diagnoses increased with the national incorporation of a similar classification system," he noted. The effect may be a logical result of the availability of molecular testing, Dr Ho added.

"Molecular testing provides more flexibility for [pathologists] to nuance their diagnosis," he explained. "Naturally, this may compel them to hedge toward an indeterminate diagnosis if it is available to them."

In an effort to offset that effect, Dr Ho and his team have placed a stronger focus on looking beyond the results of a single test, he explained. "We continue to selectively use Afirma and find it valuable, just not in isolation."

"We rely heavily upon discussion with our cytopathologists, radiologists, and ultrasonographers [and] occasionally use additional molecular testing assays, which may jointly confer a better positive predictive value."

"We are also working on obtaining other molecular testing that confers more granularity — as in the mutations that are more concerning, rather than whether it is simply 'suspicious' or 'benign,' " he added.

Duke Uses GEC Only After Two Consecutive Indeterminate Diagnoses

Meantime, Xiaoyin "Sara" Jiang, MD, an assistant professor of pathology with Duke University Medical Center, in Durham, North Carolina, who coauthored another study that was presented at the meeting on Duke's experience, agreed that the Afirma GEC test may prompt some clinicians to more readily fall back on molecular testing, but in general, the increasing options for nodule classification are only helping to improve accurate diagnosis.

"I think, in general, more information is always a good thing, and it's exciting to have these molecular tools where we can better understand the molecular characteristics of nodules and risk-stratify patients based on specific molecular mutations," she told Medscape Medical News.

"It's possible that even prior to surgery we can start stratifying patients according to more or less aggressive disease, so it can help us to stratify who needs this kind of treatment."

Dr Jiang's study, described as "the largest single-institutional study on Afirma GEC utilization to date," showed results consistent with published reports on the GEC, with strong negative predictive value, offering confidence in benign results, and weaker positive predictive value.

Among 115 indeterminate thyroid nodules that were evaluated with Afirma GEC, 45% were classified as benign and 50% as suspicious, with six showing no result due to low mRNA content.

Of the nodules classified as suspicious on GEC, 77% were treated surgically, and final pathology showed that 39% were indeed malignant.

Of 6% of nodules classified as benign that were treated surgically, all were benign on final pathology.

Most of the malignancies (71%) were follicular variant of papillary thyroid carcinoma, while the rest were Hurthle cell carcinomas (18%), or the classic variant of papillary thyroid carcinoma (6%).

Dr Jiang noted that Duke's policy is to utilize GEC only after two consecutive indeterminate diagnoses have been made on the same thyroid nodule. In those cases, the GEC's strong negative predictive can be highly beneficial, she said.

"If the repeat diagnosis continued to show an indeterminate result, many or most patients would have been offered surgery or a lumpectomy, but with GEC results showing the nodule to be benign, we can be confident of that and not only reduce costs but offer improvements in terms of quality-of-life measures relating to being spared the surgery."

Further Prospective Multi-institutional Studies Needed

Another study on the GEC looking at the longer-term outcomes of benign results further supported the test's negative predictive value, showing that among 98 patients who had data for at least 36 months following a benign GEC test result, only 17 (17.3%) had subsequently undergone surgery, with 88% of surgeries taking place within the first 2 years of the test result.

The most common reasons for surgery included having a large nodule size, compressive symptoms, or an enlarging nodule, said first author Jennifer A Sipos, MD, director of the Benign Thyroid Program at the Ohio State University School of Medicine, in Columbus.

"If the majority of patients ultimately went on to have surgery in spite of a benign GEC, the test ultimately would not be particularly helpful," she told Medscape Medical News. "We found, instead, that [most] patients did not have surgery during prolonged follow-up."

Still another study, however, evaluating follow-up on 109 indeterminate fine-needle-aspiration results of 101 patients that were subject to GEC testing, in fact showed weaker reliability of GEC benign results.

Among 39 nodules with benign GEC, 17 had surgical intervention for various indications, and 10 were confirmed to have benign final pathology, for a negative predictive value of 58.5% — significantly lower than published reports.

Of 56 nodules that had GEC test results of suspicious, 43 underwent surgery and 20 of those had a confirmed malignancy on final pathology, for a positive predictive value of 46.5%.

"The high malignancy prevalence in our practice was associated with a significantly lower negative predictive value of GEC in our practice compared with what was previously reported," wrote the authors, led by Emad Kandil, MD, chief of endocrine surgery at Tulane University School of Medicine, in New Orleans, Louisiana.

"These results show that physicians should be careful in using GEC testing in guiding surgical decision making. Future further prospective multi-institutional studies are needed," they concluded.

The studies' authors had no relevant financial relationships.

2015 International Thyroid Congress and Annual Meeting of the American Thyroid Association; Orlando, Florida. Posters 272, 822, 829, 838, presented October 19-22, 2015.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: