Retained surgical instruments can lead to prolonged hospitalization and significant complications, but the methods currently used to avoid such incidents are less than ideal. According to a new study published in the American Journal of Roentgenology, the combination of a newly developed computer-aided detection (CAD) algorithm and a radiologist displayed strong diagnostic accuracy in the detection of randomly placed radiopaque μTags. Could this be the key to reducing retained surgical instruments?

The authors used data from 700 thoracoabdominal radiographs—410 with a μTag and 290 without—to test the CAD algorithm’s effectiveness. Radiologists reviewed radiographs determined to be negative by the algorithm.

Overall, the CAD algorithm had one false-positive result. Its sensitivity was 79.5 percent and its specificity was 99.7 percent. Combining the algorithm with one “failsafe” radiologist increased sensitivity to 98.5 percent and resulted in 47 percent of the radiographs not requiring an immediate review.

These findings, the authors explained, could potentially keep the number of retained surgical instruments much lower.

“In a hypothetical future state, if the CAD system identifies a radiopaque μTag while operating in high-specificity mode, a surgeon could immediately remove that object without waiting for a radiologist to identify it,” wrote lead author Matthew S. Davenport, MD, department of radiology at the University of Michigan Hospital and Health Systems in Ann Arbor, and colleagues. “The μTag is sufficiently small to be embedded in a wide variety of iatrogenic objects and can be detected reliably despite confounding instruments in the surgical field. These results suggest that the proposed combined CAD-radiologist algorithm may help dramatically reduce the risk and occurrence of retained surgical instruments.”