BOSTON — New data on the latest versions of two top genomic sequencing classifiers (GSC) demonstrate the increasing ability of molecular testing to assist in the management of patients with thyroid nodules of indeterminate cytology.
Several molecular tests are currently available to help address the challenge of thyroid nodules for which fine-needle aspiration cytology yields indeterminate results (Bethesda class III or IV). Of these, only 5% to 30% are malignant.
The aim of the classifiers is to identify those with low enough risk for malignancy that surgical removal isn't necessary, and also to guide management of those that do test positive.
New findings from a validation study of the Afirma GSC (Veracyte) showing a 36% increase in specificity compared with the previous Afirma gene expression classifier (GEC) were published online May 23 in JAMA Surgery by Kepal N. Patel, MD, of New York University Langone Medical Center, and colleagues.
And on May 19 here at the American Association of Clinical Endocrinologists (AACE) 2018 Annual Scientific & Clinical Congress, Peter M. Sadow, MD, PhD, of Massachusetts General Hospital and Harvard Medical School, Boston, presented data for Veracyte's newly launched companion Afirma Xpression Atlas, an RNA sequencing-based test that provides additional genomic information for test-positive samples.
Also at the AACE meeting, three single-center validation studies of the Afirma GSC were presented as late-breaking posters.
And in another AACE late-breaking poster, real-world data were presented for the ThyroSeq V3 Genomic Classifier (University of Pittsburgh/CBLPath), showing comparable results to those of a recent pivotal multicenter, double-blind study that demonstrated both its improved ability to identify low-risk samples as well as further classify those testing positive. The findings were presented by the test's developer, Yuri E. Nikiforov, MD, PhD, of the University of Pittsburgh, Pennsylvania, and colleagues.
Two Horse Race as Latest Test Versions Narrow Competition
The Afirma GSC was introduced in the US in July 2017 and the ThyroSeq V3 in November 2017. The two new versions and new Afirma add-on have moved the competing brands closer in terms of both the "rule-out " and "rule-in" information they provide, thyroid specialist David A. Cohen, MD, of Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, told Medscape Medical News in an interview.
"It's like a horserace. With every new update, they're enhancing the sensitivity, specificity, positive predictive value, and negative predictive value, and they’re able to report on more mutations associated with malignancy. At this point, the sensitivity and specificity are likely very similar between the two."
But in an editorial accompanying the JAMA Surgery report, Peter Angelos, MD, PhD, of the University of Chicago, Illinois, expressed several caveats about the use of these tests in general. First, he points out that clinicians need to know the prevalence of cancer in their patient population, as it affects the sensitivity and negative predictive value of the tests.
Second, Angelos observed there are no data for use of these tests in pediatric patients. And furthermore, "Clinicians must be careful in deciding when to use the test. If a patient has an indication for surgery other than cytology, then there is little value in pursuing any molecular testing...There will still be many patients who require diagnostic surgery."
A 36% Increase in Specificity With Afirma GSC Versus Older Test
The newly published blinded validation study of the Afirma GSC included data from 49 US centers over 2-years for 191 nodules ≥ 1 cm classified as indeterminate (Bethesda III or IV) on fine-needle aspiration biopsy and with adequate residual RNA. Surgical specimens were also obtained and histopathological diagnosis made by an expert panel blinded to the genomic data. Cancer prevalence was 24%.
The GSC correctly identified 41 of 45 malignant samples as suspicious, yielding a sensitivity of 91.1%, and 99 of 145 nonmalignant samples as benign, yielding a specificity of 68.3%. That 91% sensitivity is comparable with the previous Afirma GEC test, and the specificity represents an improvement of 36%, Angelos points out in his editorial.
The negative predictive value of the GSC was 96.1%, and the positive predictive value, 47.1%. The GSC performed similarly between Bethesda III and IV categories.
Of the 26 samples with Hürthle cell histology — a particularly challenging group to classify — the GSC sensitivity was 88.9% and specificity was 58.8%.
That's a big improvement over the previous Afirma GEC, Cohen stresses.
"Up until the GSC, the main knock against the Afirma GEC was over-calling the Hürthle cell lesions as high risk. The low-risk report was reliable, but if they called it high risk it was still likely to be benign. Now with the GSC they've greatly improved it to very close to where the ThyroSeq is."
In poster presentations at AACE, similar results were seen at individual US centers. At the Memorial Center for Integrative Endocrine Surgery in Hollywood, Weston and Boca Raton, Florida, R. Mack Harrell, MD, and colleagues found that they performed 19% fewer surgeries in Afirma GSC- versus GEC-tested patients, as the GSC identified just 34% of indeterminate nodules as suspicious vs 54% with GEC.
Afima Atlas Add-on Yields More Information for Surgery
The same company's add-on, the new Afirma Xpression Atlas, detects 761 DNA variants and 130 RNA fusion pairs in over 500 genes linked to thyroid cancer, some of them associated with more aggressive thyroid cancers or higher-stage disease. It is designed to be used with the Afirma GSC for thyroid nodules destined for surgery to provide further genomic data on the nodules' actively transcribed genes about risk of malignancy, diagnosis, neoplasm type, prognosis, and cell-signaling pathway activation, all of which can inform clinical management.
"For patients already going to surgery for suspected thyroid cancer, the Afirma Xpression Atlas provides additional genomic information that physicians and patients may use in decisions about the surgical approach," said Sadow in an Afirma statement.
"Additionally, as more is understood about gene alterations and their impact on thyroid cancer development, prognosis, and tumor pathway dependence, this RNA sequencing information may help further guide patient care," Sadow said.
ThyroSeq V3 Real-World Data Confirm Randomized Trial
For its part, the ThyroSeq V3 genomic classifier tests for more than 12,000 DNA variants and over 120 gene fusions, as well as multiple copy number alterations and gene expression alterations.
In a data analysis of the first 3783 consecutive fine-needle aspiration samples with indeterminate cytology (Bethesda III, 81%; IV, 15%; V, 4%) tested by ThyroSeq V3 gene classifier from November 1, 2017 to March 1, 2018, fully informative samples were obtained from 3584 samples (95%).
Of those, 62% yielded "negative" and 38% "positive" results. Of the latter, 0.7% were actually parathyroid and 0.14% medullary thyroid carcinoma.
Among the remaining 1348 test-positive nodules of follicular thyroid-cell origin, 5.5% had a high-risk molecular profile, 16.6% had BRAF V600E-like mutations/fusions, 54.2% had RAS-like mutations/gene fusions, 14.2% had copy number alterations, and 4.7% had gene expression alterations.
The expected probability of cancer (or noninvasive follicular thyroid neoplasm with papillary-like nuclear features) would be 100% for the nonfollicular thyroid, BRAF-like alterations, and other high-risk alterations. In contrast, the malignancy risk among copy number alterations is just 59%, and among RAS-like alterations, 62%.
Thus, all patients in the former group would be referred to surgery, whereas those in the latter two groups could be followed or sent for lobectomy rather than total thyroidectomy, Nikiforov told Medscape Medical News. "Among those that are malignant, it differentiates by specific genetic signature...This gives physicians more granular information to manage patients."
Choosing Between Tests: Surgeon Preference, Cost, and Ease of Use
Asked to comment on the Afirma test, Nikiforov pointed out that despite Veracyte's much larger advertising budget, many of the large academic centers send their samples for ThyroSeq testing.
"You must be doing something right if you didn't invest [millions of dollars] in advertising and people are still using you and like you," he observed.
Gregory Dodell, MD, assistant clinical professor of medicine, endocrinology, diabetes, and bone disease at the Icahn School of Medicine at Mount Sinai, New York City, works with a surgeon who uses ThyroSeq. "As a clinician, when I get that report, it's great. It's so comprehensive and it makes me feel I'm doing the right thing for the patient."
Dodell doesn't do the biopsies himself, so the surgeon's preference is important.
"If I send him my patients, I want him to use the test he's most comfortable making decisions on. I always admire and trust a surgeon who says a patient doesn't need surgery...He uses this information to spare the patient."
Cohen, who has experience with both tests, noted, "If I'm choosing between the two, I'm also thinking about patient access. Ability to obtain a test and cost of a test play a big role. If the two are very similar, and one is cheaper or easier for the patient to do than another, that weighs heavily."
Indeed, Angelos advises in his editorial, "Clinicians who choose to use such molecular tests must do so in a careful fashion while continuing to rely on experience and clinical judgment to advise their patients."
Patel has received speaker's honoraria from and Sadow has served as a consultant for Veracyte. Nikiforov is the owner of intellectual property related to ThyroSeq. Angelos, Harrell, Cohen, and Dodell have reported no relevant financial relationships.
Materials provided by Medscape.