Gene-expression classifier: A new diagnostic test for indeterminate FNAs

Cytologically indeterminate fine needle aspiration results are common. From 15% to 30% FNA biopsies of thyroid nodules are indeterminate or suspicious, including atypia or follicular lesions of undetermined significance (AUS/FLUS); follicular or Hürthle cell neoplasms; and FNA suspicious for malignancy. Because the risk for malignancy in an indeterminate FNA may be as high as 30%, most of these patients were traditionally advised to proceed with surgical resection, followed by completion thyroidectomy if malignancy was confirmed.

The problem with this approach is that most such nodules (more than 70%) would be determined post-operatively to not be cancer. In hindsight, surgery was unnecessary. If only there were a means to determine which nodules were at higher vs. lower risk for cancer before surgery, the number of unnecessary thyroid surgeries performed could be reduced.

In the June issue of the New England Journal of Medicine, Alexander and colleagues reported the results of a prospective, multicenter study (VERA001) validating a novel diagnostic test, the gene-expression classifier, that measures the expression of 167 genes associated with thyroid cancer. This new diagnostic tool has shown promise in improving preoperative risk assessment of indeterminate thyroid nodules.

Of the 265 indeterminate nodules in the study, 85 were malignant. The gene-expression classifier correctly identified 78 of the 85 nodules as suspicious (92% sensitivity; 95% CI), with a specificity of 52%. The negative predictive values were as follows:

Atypia (or follicular lesion) of undetermined clinical significance (95%).
Follicular neoplasm or lesion suspicious for follicular neoplasm (94%).
Suspicious cytologic findings (85%).

Because of the high sensitivity and negative predictive value, this means that cytologically indeterminate nodules with benign gene expression classifier have a low risk for cancer and may be followed conservatively by observation instead of surgery, as we would usually recommend for any nodule that has benign cytology.

In the accompanying editorial, Jameson suggested that if the gene-expression classifier test was used in clinical decision-making, the number of surgeries in the United States for nodules with cytologically indeterminate FNA might possibly be decreased by as much as one-third.

Duick et al report the clinical utility of benign gene-expression classifier on the clinical decision-making of thyroid nodules with indeterminate FNA cytology results. They found that approximately one surgery was avoided for every two gene-expression classifier tests run on thyroid FNAs with indeterminate cytology.

What does this mean for clinical endocrinologists?

It means that soon not every thyroid nodule with an indeterminate FNA will need to go on to surgery. Gene-expression classifier may not eliminate the need for surgery in all nodules with cytologically indeterminate FNAs; however, this new diagnostic tool has the potential for changing how thyroidologists manage thyroid nodules in the future.

In our own practice, the majority of FNAs with suspicious or indeterminate cytology have been subsequently found to be benign on gene-expression classifier. This has markedly decreased our recommendations to proceed with surgery that might potentially have been found to be unnecessary after the fact.

For more information:

Alexander EK. N Engl J Med. 2012;367:705-715.

Jameson JL. N Engl J Med. 2012;367:765-767

Duick DS. Thyroid. 2012;doi:10.1089/thy.2012-0180.

Disclosure: I was one of the principal investigators for the VERA001 study reported above sponsored by Veracyte Inc.