Miriam E Tucker

June 01, 2016

ORLANDO, Florida — In cytologically indeterminate thyroid nodules, the mere presence of a RAS mutation may not be as predictive of malignancy as has been previously reported, a new systematic review suggests.

The RAS mutation is the most common abnormality in mutation panel testing, accounting for at least 50% of all positive results in cytologically indeterminate thyroid nodules. And thus, as part of a seven-gene panel test, the predictive value of RAS influences the performance of the overall panel, Mark A Lupo, MD, of Florida State University College of Medicine, Sarasota, explained here at the American Association of Clinical Endocrinologists 2016 Annual Meeting.

In his new systematic review of 19 studies, two-thirds of patients with RAS mutations were found to have malignant tumors following surgery, in contrast to the 80% that has been quoted repeatedly in the literature. Moreover, the RAS mutation's positive predictive value was higher in the 15 studies in which the pathologist knew the patients' mutation status than in the remaining studies, when the pathologist was blinded, suggesting bias in the histologic assessment, he said.

And, he added, other work has suggested that even when a RAS mutation is associated with cancer, it tends to be less aggressive. In fact, RAS has been commonly identified with encapsulated follicular variant of papillary thyroid carcinoma (EFVPTC), which in April 2016 was renamed noninvasive follicular thyroid neoplasm with papillarylike nuclear features (NIFTP) and deemed to be noncancerous.

"I think in the future with RAS we're going to ask whether there are other mutations that make us worry more, whereas…RAS alone as a mutation in a [thyroid] nodule probably does not have the positive predictive value we thought, and even if they're cancerous those nodules tend to be fairly indolent," Dr Lupo told Medscape Medical News in an interview.

While he still recommends surgery for most RAS-positive patients, Dr Lupo said that he now more often recommends a lobectomy rather than total thyroidectomy, as long as no other worrisome signs are present.

"We have to take the clinical information together to make the decision. RAS alone doesn't tell the whole story."

Session moderator David C Lieb, MD, associate program director of the endocrinology fellowship program at Eastern Virginia Medical School, Norfolk, said that the study "reminds us that we need to use caution in interpreting the results from molecular diagnostic tests, and it's not black and white. You have to interpret it in the setting of your patient population and also in that individual patient."

Also asked to comment, Yuri E Nikiforov, MD, PhD, professor of pathology and director of the division of molecular anatomic pathology at the University of Pittsburgh School of Medicine, Pennsylvania, told Medscape Medical News: "There is significant variability in diagnosing cancer in thyroid nodules carrying RAS mutations.

"The difference in my opinion, and as evident from the literature, is due to the fact that many of these nodules are tumors progressing to cancer through the stages of precancer or early cancer, currently called NIFTP, and they have only partially developed nuclear features of papillary cancer."

As for the gene panel test, he said, "RAS should be used with other gene markers, which together bring the sensitivity of the assay to a very high level…RAS contributes approximately 30% sensitivity to a larger panel, which overall has a sensitivity of greater than 90%."

Dr Nikiforov, who spearheaded the name change of EFVPTC to NIFTP, also agrees that lobectomy is the right procedure for RAS-positive patients who don't have other mutations more strongly associated with malignancy.

"This is very appropriate based on the fact that most of RAS-positive cancers are low-grade encapsulated tumors at the time of surgery, and their removal by lobectomy is sufficiently curative for the patient. But…these nodules should be removed."

Positive Predictive Value Varies

Before presenting the systematic review data at the AACE meeting, Dr Lupo summarized his own practice's experience with 18 RAS-positive patients in the past 3 to 4 years. Seventeen had indeterminate (Bethesda class III or IV) nodules and 15 underwent surgery. Of those, 10 (67%) were histologically benign.

Of the remaining five, three were EFVPTC (a repeat histological exam would be required to reclassify them as NIFTP, Dr Lupo pointed out). The remaining two had additional mutations predictive of malignancy. "As these driver mutations start adding up, they become more clinically relevant than single mutations," he noted.

The systematic review included 19 studies published through December 2015 that reported RAS mutations in cytologically indeterminate thyroid nodules with histological confirmation following surgery. In all, there were 2099 cytologically indeterminate nodules, including 328 with RAS mutations.

The overall positive predictive value for RAS positivity was 67.7%. Just four of the 19 papers stated that the pathologist was blinded, while the other 15 either didn't say or stated that the pathologist was unblinded. And of the studies with unblinded pathologists, six were from Dr Nikiforov's institution, the University of Pittsburgh Medical Center (UPMC), accounting for 163 of the 328 RAS-mutated nodules.

Of the total 288 nodules in the studies with unblinded pathologists, the reported malignancy risk was 68.8%, compared with 60.0% of 40 nodules in the four blinded studies. While those numbers didn't allow for statistical significance, "the trend suggests knowing that there's a mutation present may bias to thinking a cancer is present," Dr Lupo said.

Moreover, the mean positive predictive value reported for nodules with RAS mutations in the six UPMC studies was actually significantly higher than in the 13 non-UPMC studies, 84% vs 51.5% (P < .01).

In response to that, Dr Nikiforov said, "Part of the reason that in the UPMC studies a higher percentage of nodules carrying RAS mutations were called cancer is likely to be that the pathologists, knowing the mutational status of the nodule, paid particular attention to it, submitted all sections of the nodule for microscopic examination, and did the exam very carefully looking for invasion and other microscopic features of cancer. I would think that such…thorough examination is actually beneficial for the patient.

"In fact," he added, "I believe that pathologists should know the mutational status before they perform histopathologic examination, as this would help establishing more accurate diagnosis in these nodules."

But Dr Lieb pointed out that such biases could apply to clinicians as well. "If I see a patient and I know they have a particular mutation, once they've already been diagnosed with cancer and had their surgery, I have to be careful with how I interpret the result of that specific mutation, as to whether I say the person needs radioactive iodine because of the mutation. I think people make some of those decisions when there aren't as much data to support doing that just yet, because we really don't know."

Dr Lieb also said the whole concept that one mutation doesn't necessarily mean a patient has cancer but may be "at risk" for cancer is "a shifting paradigm."

He added, "What we want is a test that says you should have surgery or a total thyroidectomy vs a lobectomy…and a test that says who should definitely have radioactive iodine afterward too.…Or a test that says we don't need to worry and the patient doesn't need to come back again.…We're not there yet."

Veracyte, manufacturer of the Afirma Gene Expression Classifier, assisted with data preparation for the abstract but did not fund the study. Dr Lupo has received funding for speaking, consulting, and/or research from Veracyte, Asuragen, Interpace, and Rosetta Genomics. Dr Nikiforov's employer, UPMC, has granted CBL Path a license to market UPMC's ThyroSeq trademark for commercial use, but he receives no direct or indirect compensation related to CBLPath. He is a consultant for Quest Diagnostics. Dr Lieb has no relevant financial relationships.

For more diabetes and endocrinology news, follow us on Twitter and on Facebook.

American Association of Clinical Endocrinologists 2016 Annual Meeting; May 28, 2016; Orlando, FL. Abstract 1168


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.