Mayo Clinic AI model predicts ALS

Rochester, Minn.-based Mayo Clinic researchers have developed an artificial intelligence model that can predict amyotrophic lateral sclerosis and anticipate patient survival using data from F-wave nerve conduction studies.

The model was trained on F-wave responses from 46,802 patients, including 5,329 with motor neuron disease in the training set, according to an Aug. 29 news release. In the test set, 689 patients with ALS were matched with 689 controls. The AI used time-frequency characteristics of the responses to classify patients and estimate outcomes.

The model outperformed clinical annotations and identified factors associated with decreased survival, including older age at onset, family history of ALS and high model classification probability. It also helped identify bulbar-onset cases likely to progress, according to an Aug. 29 study in Brain.

Though the model was not trained to distinguish ALS from similar conditions, exploratory analysis showed it could differentiate ALS from disorders such as inclusion body myositis and cervical and lumbar radiculopathy.

Researchers emphasized that the model is not intended to make clinical diagnoses but to support decision-making with individualized probability estimates.

The post Mayo Clinic AI model predicts ALS appeared first on Becker’s Hospital Review | Healthcare News & Analysis.