A machine-learning algorithm that analyses thyroid-nodule ultrasound images can correctly identify cancer as accurately as a thyroid biopsy with fine-needle aspiration, a study shows.

US clinicians say their AI tool could help endocrinologists and radiologists in choosing which thyroid nodules should undergo biopsy, especially those who may not review a large volume of thyroid ultrasounds.

“We believe this is a good next step to improving patient care and avoiding unnecessary procedures,” lead researcher Dr Nikita Pozdeyev, from the University of Colorado in Aurora, said in media statement.

The findings were presented at the Endocrine Society’s annual meeting, ENDO 2022, in the US earlier this month.