12351 Real World Performance of the Afirma Xpression Atlas in Bethesda V and Vi Thyroid Nodules: A Private Practice Experience

Abstract Disclosure: M.S. Tkach: None. E. Coughlin: None. A. Pinero-Pilona: None. Fine Needle Aspiration (FNA) is a cornerstone of the diagnosis of thyroid cancer. The Bethesda System for Reporting Thyroid Cytopathology (BS-TCP), developed in 2010, and genomic sequencing classifiers (GSC) now allow clinicians to objectively identify the risk of malignancy of thyroid nodules (“TN”) prior to surgical interventions. For further characterization, the Afirma Xpression Atlas (XA) was created and launched in May 2018, utilizing RNA sequencing to evaluate fusions and variants. As of 2023, the Afirma XA panel includes 905 genomic variants and 235 fusions from 593 genes. Post-validation studies for the Afirma XA have focused on large academic centers, but applicability or extrapolation to smaller endocrinology practices is less clear, potentially due to different patient population and thyroid cancer prevalence. This study was designed to analyze and measure the performance of the Afirma XA in BS-TCP V and VI TN classified as GSC suspicious for malignancy in a small endocrinology practice. We conducted a retrospective cohort analysis comparing all BS-TCP V and VI TN with their respective final surgical pathology results between June 2018 and December 2023. 76 TN meeting the criteria for BS-TCP V or VI classification were identified. We excluded any TN (12 total) for which pathology data was unavailable; reasons ranged from lack of patient follow-up to patient's death with no autopsy available. Forty-seven thyroid TN were therefore included, and the BS-TCP VI category was the predominant cytopathology discovered, representing 78.7% of the TN in this group. Afirma XA was able to identify genomic alterations in 78.7% of all included TN (thus 78.7% is its sensitivity). Sensitivity of XA was found to be 70% in BS-TCP V (n=10) and 81.1% in BS-TCP VI (n=37). 83% of the analyzed TN were from female patients and 17% were from male patients; there was no statistically significant difference between XA frequency in TN from male vs. female patients. 100% of all BS-TCP V and VI TN (n=47) as determined by the cytopathologist were associated with cancer on pathology; thus, the sensitivity of the cytopathology itself is 100%, and positive predictive value of XA in both groups was 100%. Age and nodule size were not significantly associated with XA presence or absence overall. XA was overwhelmingly comprised of BRAF :p:V600E c.1799T>A mutations: 70.2% of of XA findings showed this mutation (50% in V and 75.6% in VI). Our percentage of XA on BS-TCP V and VI nodules reflects similarity to prior large studies in academic centers. Based on this data, we conclude that XA is a valuable tool for high-risk TN classification and treatment guidance; it can both aid in surgical planning and determine options for potential pharmacotherapy. We conclude that in our study, the relation between XA and cytopathology in BS-TCP V and VI nodules correlates well with data from large academic center studies, even in a small private practice. Presentation: 6/3/2024