Next Generation Genomic Sequencer Improves Thyroid Cancer Diagnosis

Written by Kerri Wachter

Afirma Gene-sequencing technology first to distinguish thyroid cancer subtypes, including those of Hürthle cells and medullary thyroid carcinomas.

With Quan-Yan Duh, MD, Gregory Randolph, MD, and Angela M. Leung, MD, MSc

Genomic sequencing technology combined with machine learning could reduce the need for diagnostic surgery for some patients with thyroid nodules considered indeterminate by cytopathology, and improve surgical planning for others,1 according to findings presented at the American Thyroid Association 87th Annual Meeting in Victoria, BC, Canada. The authors demonstrated the efficacy of the Afirma Genomic Sequencing Classifier (GSC) to distinguish benign Hürthle cell adenomas from Hürthle cells carcinomas and to identify medullary thyroid cancer.

Genomic Sequencing Classifier is the next generation of the Afirma Gene Expression Classifier, which “identifies, by using RNA sequencing and machine learning algorithms, genomically benign thyroid nodules among those with indeterminate FNAB [fine needle aspiration biopsy] to prevent unnecessary diagnostic surgery,” according to information provide by the manufacturer, Veracyte, Inc. 

Minimizing Surgery for Hürthle Cell Adenomas

In the case of Hürthle cell carcinomas, which account for about 20% of all indeterminate thyroid nodules, fine needle aspiration biopsies are often cytologically indeterminate. Pathologic examination of surgical specimens is necessary to distinguish benign Hürthle cell adenomas from malignant carcinomas.1

“Although many patients with such lesions are recommended for thyroid surgery, only a minority will be diagnosed with a Hürthle cell carcinoma,” said Angela M. Leung, MD, an endocrinologist and assistant professor of medicine at the University of California/Los Angeles.  Dr. Leung moderated the session but was not involved in any of the studies.

“The challenge, thus, is to recommend thyroid surgery only when needed, in order to avoid the unnecessary surgeries for those patients with Hürthle cell adenomas,” Dr. Leung told EndocrineWeb.

The Afirma Genomic Sequencing Classifier appears capable of identifying those patients with benign Hürthle cell adenomas, who might be able to avoid surgery. In an analysis of 26 Hürthle-cell fine needle aspiration samples, the test was 89% sensitive for malignancy and 59% specific for benign nodules,1 said Quan-Yan Duh, MD, professor of surgery and chief of endocrine surgery at the University of California San Francisco, in presenting the data at ATA.

“Our findings suggest that the Afirma GSC significantly improves Hürthle cell classification, which means that more patients with benign nodules should be able to avoid unnecessary diagnostic surgery,” said Dr. Duh.

Confirming Medullary Thyroid Cancers  

Gregory Randolph, MD, FACS, FACE, who is Claire and John Bertucci Endowed Chair in Thyroid Surgical Oncology at Harvard Medical School in Boston, Massachusetts, presented results from a study of 211 indeterminate thyroid nodule fine needle aspiration (FNA) samples,2 at the American Thyroid Association 87th Annual Meeting. The Afirma test for medullary thyroid cancer was interestingly 100% sensitive and specific in identifying medullary thyroid cancer, said Dr. Leung. 

Not only is medullary thyroid carcinoma (MTC) rare, accounting for only 1-2% of all thyroid cancers in the United States, it is difficult to diagnose.2

“MTCs secrete calcitonin, different than the thyroid cancers derived from thyroid follicular cells, which secrete thyroglobulin. Screening for elevated serum calcitonin levels in every thyroid nodule is usually not regarded as cost-effective in the United States,” Dr. Leung said.

“When combined with the uncertainties in the histological diagnosis of MTC following fine needle aspiration biopsies, these factors overall contribute to the increased chances of missing a diagnosis of medullary thyroid cancer, preoperatively,” she said.

Imaging Studies Inform Surgical Strategy

Confirming a nodule is MTC in advance will help surgeons plan appropriately, said Dr. Randolph.

“This includes ordering the necessary imaging studies to inform whether cancer has spread regionally to cervical lymph nodes. It also enables the surgeon to anticipate potential complications such as hypertension that can accompany MTC in certain cases from undiagnosed adrenal tumors which may co-exist,”2 he said.

In order to develop the Hürthle cell carcinoma classifier, a differential expression analysis was performed on 21,162 genes, of which the most significant 1,408 differentially expressed nuclear and mitochondrial genes were selected for differentiation of Hürthle from nonHürthle samples.1,2

All 13 mitochondrial transcripts were highly ranked and were all selected,1 according to Dr. Duh. A training set of 119 Hürthle cell carcinoma-negative and 199 positive samples were used for cross-validation. Next, neoplasm-specific features were identified to develop an algorithm to detect Hürthle cell neoplasms. The resulting classifier improved specificity over the previous generation Afirma Gene Expression Classifier by 47%.1,2

“The results of this study add to the existing literature regarding the use of molecular testing panels to assess the likelihood of Hürthle cell carcinomas among cytologically indeterminate thyroid nodules,” said Dr. Leung, “and it improves upon the performance of the existing Genomic Sequencing Classifier to be extended also for use in patients with Hürthle cell carcinomas.”  

Afirma GSC Demonstrated High Sensitivity and Specificity

A similar process was used to develop an algorithm to identify medullary thyroid carcinomas. The initial training involved 483 FNA samples, of which 21 were MTC.2 Selection of features involved an additional 87 tissues. A total of 108 differentially expressed genes were identified, with five ultimately included in the GSC.

Dr. Randolph noted that the Afirma GSC’s high sensitivity and specificity demonstrated in the study “should empower surgeons and their patients with the information they need for optimal treatment.”

Three other studies of the Afirma GSC were also presented at the ATA meeting as posters. The researchers found high sensitivity and specificity in detecting the BRAF V600E mutation and for distinguishing parathyroid from nonparathyroid tissue.3

The BRAF V600E gene mutation was associated with the growth and spread of cancer cells and its presence or absence in thyroid nodules may lead to better treatment planning.3 Because parathyroid glands may be mistaken for thyroid nodules, the ability to identify this tissue in indeterminate FNAs could help reduce the number of unnecessary surgeries.4

Lastly, in another poster presentation, researchers verified the efficacy of the Afirma Gene Sequencing Classifier for thyroid nodule FNA samples with as little as 5 ng of RNA.5 In addition, the test functioned well even with contaminants present in the sample.

There were no relevant disclosures to report.

 

 

Sources

  1. Duh Q, Angell TE, Babiarz J, et al. Development and Validation of Classifiers to Enhance the Afirma Genomic Sequencing Classifier Performance Among Hürthle Cell Specimens. Oral abstract presented at the American Thyroid Association 87th Annual Meeting. October 18-22, 2017,  Victoria, BC.
  2. Randolph G, Angell TE, Babiarz J, et al. Clinical Validation of the Afirma Genomic Sequencing Classifier for Medullary Thyroid Cancer. Oral abstract presented at American Thyroid Association 87th Annual Meeting. October 18-22, 2017,  Victoria, BC.
  3. Angell TE, Babiarz J, Barth N, et al. Clinical Validation of the Afirma Genomic Sequencing BRAF V600E Classifier. Poster presented at the American Thyroid Association 87th Annual Meeting. October 18-22, 2017,  Victoria, BC.
  4. Sosa J, Angell TE, Babiarz J, et al. Clinical Validation of the Afirma Genomic Sequencing Parathyroid Classifier. Poster presented American Thyroid Association 87th Annual Meeting. October 18-22, 2017,  Victoria, BC.
  5. Hu Z, Choi Y, Liu T, et al. Analytical Performance of Afirma GSC: A Genomic Sequencing Classifier for Cytology-Indeterminate Thyroid Nodule FNA Specimens. Poster presented at the American Thyroid Association 87th Annual Meeting. October 18-22, 2017,  Victoria, BC.