BRAF V600E mutation is insufficient for accurate diagnosis of thyroid neoplasms, although it has been widely used in clinical settings. We aimed to construct a messenger RNA (mRNA)-based model to identify thyroid cancer in indeterminate thyroid nodules in the absence of BRAF mutation. Our cohort consisted of 88 papillary thyroid cancer (PTC) patients with BRAF V600E substitution, 103 BRAF mutation negative PTC cases, and 96 subjects with benign thyroid nodules. Based on the mRNA levels of MET proto-oncogene, receptor tyrosine kinase (MET), TIMP metallopeptidase inhibitor 1 (TIMP1), transforming growth factor alpha (TGFA), Cbp/p300 interacting transactivator with Glu/Asp rich carboxy-terminal domain 1 (CITED1), and fibronectin 1 (FN1) in thyroid specimens, model building was conducted to develop a gene expression-based classifier to differentiate benign from malignant thyroid nodules. In addition, external validation was performed using the thyroid carcinoma (THCA) dataset from the Cancer Genome Atlas (TCGA). Dataset-based bioinformatic analysis confirmed the elevated expression levels of these five genes in PTC compared to normal thyroid tissue. Overall, the binary classified model correctly identified 84 of 88 BRAF-mutated PTC and 75 of 103 BRAF mutation negative PTC as malignant tumors, respectively. For PTC carrying BRAF V600E mutation, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the curve (AUC) were 0.955, 0.969, 0.966, 0.959 and 0.987, respectively. More importantly, the diagnostic performance of the approach in determining PTC without BRAF mutation was presented as the sensitivity of 0.728, specificity of 0.969, PPV of 0.962, NPV of 0.769, and AUC of 0.871. We developed and validated a gene expression classifier to assist in the diagnosis of BRAF V600E-negative PTC cases. This tool is poised to become a useful adjunct to individualized management of thyroid cancer in clinical practice.
Wang et al. (Tue,) studied this question.