Can AI correctly identify malignant, benign breast lesions?

A novel artificial intelligence-based model could help clinicians diagnose early-stage breast cancer.

In a study published in BMC Medicine, researchers developed a deep-learning model using radiologist-specific knowledge from radiologic reports and the preoperative breast magnetic resonance images of more than 1,500 patients with benign or malignant pathology-confirmed breast lesions — with the goal of enhancing the diagnostic interpretability on breast MRI scans. The diagnostic accuracy of the new AI model was compared with a black-box deep-learning model.

Compared with the black-box model, the new AI model had a comparable area under the receiver operating characteristic curve. The model also demonstrated higher diagnostic accuracy compared with a majority of the radiologists included in the study, improved inter-reader agreement and helped radiologists downgrade previously diagnosed malignancies to benign lesions.

The findings revealed the potential clinical applications of the novel AI model in the classification of early breast cancer and reduction of unnecessary biopsies.

Read more: BMC Medicine

The article presented here is intended to inform you about the broader media perspective on dentistry, regardless of its alignment with the ADA's stance. It is important to note that publication of an article does not imply the ADA's endorsement, agreement, or promotion of its content.