Tympanic videos classified into AOM versus no AOM had sensitivity of 93.8 percent, specificity of 93.5 percent
By Elana Gotkine HealthDay Reporter
THURSDAY, March 7, 2024 (HealthDay News) — An artificial intelligence decision support tool to interpret videos of the tympanic membrane can predict diagnosis of acute otitis media (AOM), according to a study published online March 4 in JAMA Pediatrics.
Nader Shaikh, M.D., from the University of Pittsburgh School of Medicine, and colleagues developed and internally validated an artificial intelligence decision-support tool to interpret videos of the tympanic membrane and enhance accuracy in the diagnosis of AOM. Otoscopic videos of the tympanic membrane captured using a smartphone during outpatient clinic visits at two sites in Pennsylvania were analyzed; 1,151 videos from 635 children (most younger than 3 years) were included in the study.
The researchers found that diagnostic accuracy was almost identical for the deep residual-recurrent neural network and the decision tree network. Tympanic membrane videos were classified into AOM versus no AOM categories with a sensitivity and specificity of 93.8 and 93.5 percent, respectively, with the finalized deep residual-recurrent neural network, while the decision tree model had corresponding sensitivity and specificity of 93.7 and 93.3 percent. Bulging of the tympanic membrane aligned with the predicted diagnosis most closely; in the test set, bulging was present in all 230 cases in which the diagnosis was predicted to be AOM.
“With appropriate training, this tool could be used by a wide range of medical personnel to enhance teaching of otoscopic examination, discussion with colleagues, documentation in the electronic health record, and discussion with parents,” the authors write.
Editorial (subscription or payment may be required)
Copyright © 2024 HealthDay. All rights reserved.
