Thyroid Cancer Coverage from Every Angle
Advertisement
Advertisement

Using AI to Determine the Genetic Risk of Thyroid Nodules

By: Joseph Fanelli
Posted: Thursday, March 12, 2020

Testing the ability of artificial intelligence (AI) to stratify thyroid nodules, machine learning demonstrated the potential to define the genetic risk of thyroid nodules, according to the results of a study presented in JAMA Otolaryngology—Head & Neck Surgery. John Eisenbrey, PhD, of Thomas Jefferson University, Philadelphia, and colleagues concluded that these findings show the diagnostic applications of machine learning interpretation and critically afford physicians additional information to cytogenetics with undetermined cytologic results.

“Machine learning is a low-cost and efficient tool that could help physicians arrive [at] a quicker decision as to how to approach an indeterminate nodule,” Dr. Eisenbrey said in a Thomas Jefferson University press release. “No one has used machine learning in the field of genetic risk stratification of thyroid nodule on ultrasound.”

In this retrospective diagnostic study, the authors reviewed the electronic medical records of 121 patients who underwent ultrasonography and molecular testing for thyroid nodules from January 1, 2017, through August 1, 2018. Nodules were classified as high or low risk based on the results of an institutional molecular testing panel for thyroid risk genes. The authors applied a machine-learning algorithm to ultrasound images of the patients’ thyroid nodules.

Among the 134 lesions identified from the patients, 638 diagnostic ultrasonographic images were selected, and 556 were used for the training model, 74 for validation, and 53 for testing. The machine-learning model successfully identified genetically high-risk thyroid nodules with a specificity of 97% and a positive predictive value of 90%. The model exhibited a negative predictive value of 74.4% and an overall accuracy of 77.4%.

“There are so many potential applications of machine learning,” Dr. Eisenbrey added. “In the future, we’d like to make use of feature extraction, which will help us identify anatomically relevant features of high-risk nodules.”

Disclosure: For the full disclosures of the study authors, visit jamanetwork.com.

Advertisement
Advertisement


By continuing to browse this site you permit us and our partners to place identification cookies on your browser and agree to our use of cookies to identify you for marketing. Read our Privacy Policy to learn more.