Alberta Company Creates Innovative AI Tool for Thyroid Scans

A new tool has been developed by an Alberta company MEDO, which combines current thyroid scanning methods with artificial intelligence to simplify the classification of thyroid nodules.

Thyroid nodules are common, and need to be classified to investigate any indications of thyroid cancer and other issues. Thyroid scans involve an ultrasound probe of the neck which takes multiple pictures of the thyroid gland. A sonographer measures and characterises any nodules that are found in order for a radiologist to identify whether they are benign or possibly malignant. This subjective interpretation can produce errors, especially for people with complex thyroid glands. If a benign nodule is misconstrued as malignant, this can result in unnecessary biopsies. In the case of a false negative, malignancy may be undiscovered.

The need for efficient and accurate classification has been met by MEDO’s artificial intelligence software, which takes new and previous scans to classify nodules efficiently and reliably. The tool has recently been approved by the United States’ Food and Drug Administration. It is hoped that this approval will be followed by similar approvals in Singapore and eventually Europe.

MEDO was co-founded by a computer scientist, technology strategist and two radiologists, whose cross-disciplinary expertise have allowed them to create deep learning solutions to revolutionise ultrasound technology. The company is also developing tools for lung and fracture ultrasounds.

The company has received support and mentorship from the University of Alberta’s Health Hub. The Health Hub supports start-ups in the health and life-science industries. The University of Alberta is ranked in the top 5 of research universities in Canada.