R. Rokosh1, A. Kundel1, T. C. Hill1, J. Ogilvie1, K. Patel1 1New York University School Of Medicine,Department Of Surgery,New York, NY, USA
Introduction: The Afirma® Gene Expression Classifier (AGEC) has been shown to identify benign thyroid nodules among those classified as cytologically indeterminate with a negative predictive value of 94-95%. This test therefore has the potential to help avoid unnecessary surgery on Bethesda III and IV nodules that are ultimately found to be benign. Our study aimed to assess the clinical utility of the AGEC molecular assay for thyroid nodules with indeterminate cytopathology at a tertiary referral center.
Methods: Retrospective analysis of all indeterminate thyroid nodules evaluated with GEC from September 2012 to December 2014 at a large tertiary referral center was performed. Cytologic and AGEC diagnosis were compared with final surgical pathology in corresponding samples. A prevalence of malignancy of 40%, as established at our institution, was used to estimate performance characteristics with Bayes Theorem.
Results: Over the course of 27 months, 154 patients with indeterminate nodules by FNA who underwent AGEC testing were identified. Of these, AGEC classified 104 (67.5%) as suspicious, 43 (27.9%) as benign, and 7 (4.6%) as non-diagnostic. Of the 104 suspicious AGEC patients, 71 underwent thyroidectomy (2 patients were operated on elsewhere), and ———43/69 (62.3%) had malignant final pathology. Of the 43 benign AGEC patients, 8 underwent thyroidectomy and 1/8 (12.5%) had malignant final pathology. Based on these data, our AGEC sensitivity is 97.7% and specificity is 21.1%. Given the 40% prevalence of malignancy at our institution, the estimated negative predictive value of AGEC in our practice is 93.3%. Our mean follow-up time for this study was 7.8 months.
Conclusion: This study confirms that patients with suspicious AGEC have a high likelihood of having a malignancy and should undergo surgery. Our study suggests that practice-specific cancer incidence within the indeterminate cytopathology (Bethesda III-IV) group should be calculated at each institution to evaluate its unique NPV of Afirma analysis. Thus, the clinical utility of a benign AGEC result in surgical decision-making varies with disease prevalence, which is unique to each practice.
