Researchers at the University of California at Irvine and NIH's Imaging Biomarkers and Computer-Aided Detection Laboratory in Bethesda, Md., developed a software algorithm that can distinguish specific spinal fractures, according to HealthImaging.
The team looked at 104 scans, with 94 containing fractures and 10 with no fractures.
Radiology published the study in its January edition.
Here are five observations:
1. Through quantitative analysis, the algorithm has the capability to diagnose specific fractures of the thoracic and lumbar spine.
2. Joseph Burns, MD, PhD, and his team reported the algorithm demonstrated a 92 percent "sensitivity for fracture detection and localization of the correct vertebra."
3. The study found a false-positive rate of sensitivity for fracture locations to the correct vertebra of 1.6.
4. The researchers concluded the algorithm is not ready for clinical application yet, but it has promise to help radiologists with fracture classification.
5. The team believes the algorithm may also reduce fracture report discrepancies.