Ultrasound AI could help standardize trauma care
Trauma patients who present with poor blood flow and suspected abdominal hemorrhage are well served by emergency physicians using AI-augmented FAST imaging.
The acronym stands for Focused Assessment with Sonography in Trauma. The thumbs-up comes from the University of Kentucky, where researchers retrospectively tried the AI assist for FAST using patient data acquired over an eight-year period ending in 2022.
The setting was a quaternary care level-1 trauma center that performs around 3,500 adult trauma evaluations per year.
Senior author Zachary Warriner, MD, and colleagues had their work published in May by the Journal of Trauma and Acute Care Surgery [1].
The team sought to specify the role AI might play in interpreting abdominal images acquired with FAST protocols.
4 CNNs, 1 impressive performance
To separate strengths from weaknesses, they investigated the technology’s adeptness along two lines—the adequacy of the visualization and the accuracy of any positive findings.
“Adequate” FAST images were those that offered diagnostic-quality looks at the liver and kidney or the spleen and kidney.
Successful “positive” results correctly detected abnormal fluids associated with blunt force trauma to the abdomen.
Using more than 6,600 images from 109 cases, the researchers put four separate convolutional neural networks through their paces for the two tryouts of interest.
On analysis they found the AI models achieved a collective 88.7% accuracy, 83.3% sensitivity and 93.6% specificity for the “adequacy” test cohort.
The “positive” cohort hit 98.0% accuracy, 89.6% sensitivity, and 100.0% specificity.
AI aids accuracy
Meanwhile they found AI augmentation improved the accuracy and sensitivity of the “positive” models to 95.1% accurate and 94.0% sensitive.
The team’s overall conclusion:
AI can detect positivity and adequacy of FAST exams with 94% and 97% accuracy [respectively], aiding in the standardization of care delivery with minimal expert clinician input. AI is a feasible modality to improve patient care imaging interpretation accuracy and should be pursued as a point of care clinical decision-making tool.”
Study abstract here, full text behind paywall.
Dave P. has worked in journalism, marketing and public relations for more than 30 years, frequently concentrating on hospitals, healthcare technology and Catholic communications. He has also specialized in fundraising communications, ghostwriting for CEOs of local, national and global charities, nonprofits and foundations.