Thyroid nodule molecular profiling: The clinical utility of Afirma Xpression Atlas for nodules with Afirma Genomic Sequencing Classifier-suspicious results

Background: The Afirma Genomic Sequencing Classifier uses whole transcriptome RNA sequencing to identify thyroid nodules as benign or suspicious. The Afirma Xpression Atlas became available in 2018 and reports findings across 593 genes, including 905 variants and 235 fusions. When an alteration is identified, its risk of malignancy and associated neoplasm type is listed. We report the results of Afirma Xpression Atlas testing at our institution during its first 2 years of clinical use. Methods: All patient charts with indeterminate thyroid nodules and Afirma Xpression Atlas results at our institution were reviewed. Thyroid nodule characteristics, cytology, Afirma Genomic Sequencing Classifier results, Afirma Xpression Atlas results, and final histopathology were reported. Results: Afirma Xpression Atlas was performed on 136 indeterminate nodules since May 2018, and 103 met inclusion criteria. Forty-three nodules had positive Afirma Xpression Atlas results, and of these, 83.7% were follicular cell-derived thyroid cancer on surgical histopathology. This is similar to the overall 82.5% positive predictive value among Afirma Genomic Sequencing Classifier-suspicious indeterminate nodules during the same time period. Of the 60 nodules with negative Afirma Xpression Atlas, 73.3% were follicular cell-derived thyroid cancer on surgical histopathology. Conclusion: Afirma Xpression Atlas positivity is predictive of follicular cell-derived thyroid cancer, but its positive predictive value is similar to that of Genomic Sequencing Classifier-suspicious results alone at our institution, which is higher than previously published. Specific mutations likely predict follicular cellderived thyroid cancer with higher accuracy, but our current sample size of any given mutation is too small to evaluate this further. Larger studies are needed to determine whether Afirma Xpression Atlas results predictably inform the risk of malignancy and tumor characteristics in thyroid nodules. (c) 2021 Elsevier Inc. All rights reserved.