Boston, MA – Each year, tens of thousands of patients have all or part of their thyroids removed to rule out cancer because of suspicious, but uncertain, cytology test results. In the majority of cases, the suspicious thyroid nodules are determined to be ultimately benign.
Now, new research led by Brigham and Women’s Hospital (BWH) finds that a novel, genomic diagnostic test that measures the expression of 167 genes has shown promise in improving pre-operative risk assessment by re-classifying otherwise indeterminate results from thyroid biopsies as either benign or suspicious. These findings are published early online in the New England Journal of Medicine.
“Our findings show that the gene expression test can substantially reclassify otherwise inconclusive results from thyroid biopsies, said Erik Alexander, MD, lead author of the paper and a physician-researcher in the Division of Endocrinology, Diabetes and Hypertension at Brigham and Women’s Hospital. “Currently, indeterminate thyroid nodules are usually referred for thyroid surgery given the unanswered question of thyroid cancer. While all care should be personalized, our findings suggest that this test has the potential to drastically reduce unnecessary surgery and allow physicians to monitor many patients in a more conservative fashion.”
Thyroid nodules are common and ultrasound-guided needle biopsies (fine needle aspiration) have been shown to accurately identify about 65-75 percent of nodules as benign. Approximately five-10 percent of diagnostic biopsies are malignant. The remaining biopsies produce indeterminate results – occurring about fifteen to thirty percent of the time. For these patients, there remains substantial concern for thyroid cancer, though the diagnosis is uncertain. Because of the concern for cancer, in most cases, all or part of the thyroid is removed for final diagnosis. However, the nodule is ultimately benign in seventy to eighty percent of cases. For these patients, surgery was not needed and the patient was unnecessarily exposed to the cost, risk and morbidity associated with this intervention.
In this research study, researchers enrolled 3,789 patients and collected 4,812 thyroid samples from nodules larger than 1 cm and evaluated the effectiveness of a novel gene expression test in 265 thyroid samples that were cytologically indeterminate (suspicious for cancer, though not conclusive) from 49 different academic and community hospital sites around the United States. Samples were simultaneously collected for the standard-of-care cytology analysis, as well as one or two additional needle sticks for inclusion in the study. If the cytology result was indeterminate, the study sample was then analyzed using the gene expression test. Thyroid surgery was performed based on the judgment of the treating physician who was blinded to the test results. At completion of the study, the gene expression test results were compared to final histopathology diagnosis (the gold-standard diagnosis) provided by two blinded pathology experts following their review of the surgical tissue sample.
For all cytologically indeterminate nodules, the researchers found that when the gene expression test was benign, histopathological analysis of the nodule proved it benign 93 percent of the time. And when applied to lower risk subgroups of indeterminate thyroid samples labeled “atypia (or follicular lesion) of an undetermined significance” or “follicular neoplasm,” the accuracy improved to 94 and 95 percent, respectively.
“This very high negative predictive value is comparable to that of a cytologically benign cytology result. Therefore, such a result will allow clinicians to recommend a more conservative approach of watching waiting in lieu of diagnostic surgery,” said Alexander. “It is estimated that over 50 percent of current thyroid surgeries are performed unnecessarily (for cytologically indeterminate, though histopathologically benign nodules), and therefore could be impacted by the gene expression test.”
The study was co-led by Bryan R. Haugen, M.D., professor of medicine and pathology, and Head, Division of Endocrinology, Metabolism & Diabetes at the University of Colorado. Researchers from the Perelman School of Medicine at the University of Pennsylvania, The Ohio State University College of Medicine, Centro Diagnostico Italiano, University of Washington School of Medicine, University of Cincinnati College of Medicine, Johns Hopkins University School of Medicine and Veracyte, Inc. contributed to the study.
The genomic test, Afirma Gene Expression Classifier, is manufactured by Veracyte, Inc. The study was funded by a research grant by Veracyte, Inc. to the institutions of the study investigators. The co-principle investigators, in conjunction with a steering committee, had complete control of the study design, analysis, and reporting of results.