The detection of thyroid nodules is something common in clinical practice. Most thyroid nodules are not serious and have no symptoms, and only a small percentage of them are cancerous. However, it is not always possible to confirm the latter with standard procedures: it has been estimated that 20% of detected nodules are classified as indeterminate using standard techniques. In these cases, surgery is chosen to remove them. Once removed, between 65% and 75% of these nodules are benign, according to the final histological evaluation. In other words, they are cases in which surgery could have been avoided if a better diagnostic method had been available.
This is what scientists from the CSIC's Centro de Investigación en Nanomateriales y Nanotecnología an the Fundación para la Investigación y la Innovación Biosanitaria del Principado de Asturias (FINBA) have done. The team has developed a method to improve thyroid cancer diagnoses during preoperative evaluation. It is based on the correlation between the methylation level of various CpG regions of DNA and the malignancy of a large sample of thyroid nodules.
DNA methylation is a chemical modification involved in gene expression and gene silencing. Altered methylation patterns are known to be associated with cancer development. One type of such alteration is found in the CpG regions.
What this team has done is to apply machine-learning calculations to analyse a large number of genetic data from fine-needle aspiration biopsies (FNAs). The scientists have analysed the data, firstly with classical analysis techniques and afterwards with artificial intelligence techniques. As a result, from the initial 850,000 biomarkers they have identified three biomarkers which predict the malignancy of a given sample with high specificity and sensitivity.
This is an affordable method, sensitive and specific, that improves the ability to diagnose thyroid cancer during pre-operative evaluation. Therefore, many surgeries can be avoided in the cases which are difficult for diagnostic with current techniques (scientists estimate that up to 60% cases could avoid surgery). The team is now looking for interested companies to develop specific kits and implement them in hospitals for thyroid cancer diagnosis.
Contact:
Xavier Gregori
Deputy Vice-Presidency
for Knowledge Transfer-CSIC
Tel.: +34 93 887 60 04