Next-Generation Sequencing Effectively Detects Cancer in Thyroid Nodules

The ThyroSeq v2 next-generation sequencing assay was highly accurate at diagnosing cancer in thyroid nodules with follicular neoplasm or those suspected of follicular neoplasm cytology.

Yuri E. Nikiforov, MD, PhD

The ThyroSeq v2 next-generation sequencing (NGS) assay was highly accurate at diagnosing cancer in thyroid nodules with follicular neoplasm or those suspected of follicular neoplasm cytology, according to Yuri Nikiforov, MD, presenting at the 2014 Annual Meeting of the American Thyroid Association.

The most common approach used to evaluate thyroid nodules is fine needle aspiration (FNA) cytology, but about 20% to 30% of FNA results have indeterminate cytology. A common indeterminate diagnosis, follicular (or oncocytic) neoplasm/ suspicious for a follicular (or oncocytic) neoplasm (FN/SFN), carries a risk of cancer of 15% to 30% and is typically managed with surgical lobectomy.

"In this study, we tested whether the most complete next-generation sequencing panel of genetic markers can significantly improve cancer diagnosis in these nodules," said Nikiforov, vice-chair for Molecular Pathology and director, Division of Molecular and Genomic Pathology at the University of Pittsburgh Medical Center.

NGS uses massively parallel sequencing of clonally amplified DNA templates to sequence up to 106 to 109 bases with a sensitivity of 3% to 5% at a cost of about $0.05 to $1.00 per 106 bases, and is able to investigate multiple regions of the genome up to the whole genome.

The objectives of this study were to create the most complete panel of genetic markers for thyroid FNAs based on NGS and to explore the panel performance for cancer diagnosis in nodules with FN/SFN (Bethesda System Diagnostic Category IV) cytology.

To do this, a total of 143 consecutive FNA samples with the cytologic diagnosis of FN/SFN and known surgical outcome were evaluated. Of these, 91 samples were studied retrospectively, and 52 samples were studied prospectively. The performance characteristics of the two cohorts were compared by testing for similarity in true positive fraction (TPF), false positive fraction (FPF), positive predictive value (PPV), and negative predictive value (NPV), and were found to be similar.

Molecular analysis was performed using targeted NGS on Ion Torrent PGM (318 chip) using the ThyroSeq v2 panel that simultaneously tests for point mutations in 13 genes, including BRAF, NRAS, HRAS, KRAS, PTEN TP53, TSHR, GNAS, CTNNB1, RET, PIK3-CA, AKT1, TERT, and 42 gene fusions involving the RET, BRAF, NTRK1, NTRK3, ALK, PPARG, and THADA genes. The cell type composition of FNA samples and the expression of oncogenes resulting from gene fusions were assessed by the expression of 15 genes.

In evaluating the rate of informative testing, 99.3% of samples worked for mutations, and 98% of samples worked for gene fusion and gene expression analysis. The overall analytical sensitivity was 3% to 5% of mutant alleles (6% to 10% of cells) with mutation and 1% of cells with gene fusion.

Using histology, 39 malignant and 104 benign nodules were detected, for an overall cancer rate of 27.3%. In these samples, the NRAS (n =16), KRAS (n = 6), TERT (n = 4), and TSHR (n = 3) genes were those that were most commonly affected by point mutations. These molecular alterations in FNA samples were associated with an 81%, 83%, 100%, and 33% risk of cancer, respectively. Two nodules displayed concurrent NRAS and TERT mutations. Three gene fusions were identified, including THADA (n = 5), PPARG (n = 4), and NTRK3 (n = 2). All three gene fusions were associated with a 100% risk of cancer.

In the overall cohort, the 60-gene ThryoSeq v2 panel performed well with a high sensitivity (90%; 95% CI, 80%—99%) and specificity (93% CI, 88%–98%), resulting in a PPV of 83% (95% CI, 72%–95%) and a NPV of 96% (92%–100%) in nodules with FN/SFN cytology. The overall accuracy of the panel was 92% (88%–97%).

"High NPV and PPV is expected in populations with variable disease prevalence (14% to 34%)," Nikiforov explained. This comprehensive genotyping panel may also have additional benefits in terms of both tumor prognostication (BRAF, TERT, and TP53) and in identifying targets for therapy (BRAF/RAS, ALK, NTRK3).

"The results of this study indicate that comprehensive genotyping of thyroid nodules using a broad NGS panel provides a highly accurate diagnosis for nodules with FN/SFN cytology and should facilitate optimal management of these patients," the researchers concluded.


View more from the 2014 ATA Annual Meeting

Related Videos
View All
Related Content