Researchers find errors in AI-supported medical imaging

by | Jul 10, 2020 | medical malpractice

Medical centers in Louisiana and throughout the country use many advanced imaging techniques to diagnose patients. Artificial intelligence has emerged as a tool in MRI, CT and nuclear magnetic resonance imaging machines, but its accuracy has come into question after a research team tested the algorithms employed by imaging software during image reconstruction.

Small changes equal big errors

The researchers ran tests on the AI-supported imaging software. The tests were designed to evaluate how small patient movements, small structural changes and repeated sampling influenced final images. Test results showed that AI algorithms injected blurred details into final images after processing small alterations in input data. Repeated sampling added to the deterioration of images. Overall, significant errors emerged in patient images, and the researchers concluded that AI, in its current form, posed a risk of misdiagnosis.

Potential to harm patients

The research team acknowledged that AI could someday improve medical imaging but cautioned about its exclusive use at this point in time. The deep learning systems within AI proved to be unstable, so physicians relying on AI-produced images could misdiagnose people and select the wrong courses of treatment. Such mistakes could harm patients through the delay of care or inappropriate care.

Financial consequences of misdiagnosis

A misdiagnosis or delayed treatment could leave you with more medical expenses than if you had been treated correctly the first time. You may also experience prolonged pain or even disability when the failings of a medical professional or institution harm you. Damages are potentially recoverable after medical negligence, but these cases sometimes require litigation in order to pursue a settlement. You may get answers about your legal rights by speaking with a personal injury attorney.