AI Analysis of Chest Radiographs Can Triage Acute Chest Pain Syndrome

Model including deep learning predictions significantly improved discrimination of the composite outcome.

HealthDay News Deep learning (DL) analysis of chest radiographs may improve triage of patients with acute chest pain (ACP) syndrome in the emergency department, according to a study published online Jan. 17 in Radiology.

Márton Kolossváry, M.D., Ph.D., from Massachusetts General Hospital in Boston, and colleagues examined whether a DL analysis of the initial chest radiograph can help triage patients with ACP syndrome. To predict the 30-day composite end point, including acute coronary syndrome, pulmonary embolism, or aortic dissection, and all-cause mortality, a DL model was trained on 23,005 patients based on chest radiographs. Performance between models was compared using the area under the receiver operating characteristic curve (AUC; model 1: age + sex; model 2: model 1 + conventional troponin or d-dimer positivity; model 3: model 2 + DL prediction) in internal and external datasets.

The researchers found that compared with models 1 and 2, model 3, which included DL predictions, significantly improved discrimination of those with the composite outcome (AUC, 0.62, 0.76, and 0.85 for models 1, 2, and 3, respectively). Overall, 14 percent of patients could be deferred from cardiovascular or pulmonary testing for differential diagnosis of ACP syndrome using model 3 compared with 2 percent of patients using model 2, when using a sensitivity threshold of 99 percent. In different age, sex, race, and ethnicity groups, model 3 retained its diagnostic performance.

“We were able to provide more accurate predictions regarding patient outcomes as compared to a model that uses age, sex, troponin or d-dimer information,” Kolossváry said in a statement.

One author disclosed financial ties to the pharmaceutical and medical device industries.

Abstract/Full Text

Editorial (subscription or payment may be required)