Pairing an alert system with CAD software halves time-to-treatment for pneumothorax
Combined with an electronic notification system (ENS), computer-aided detection (CAD) programs targeted at identifying pneumothorax on imaging can significantly reduce the time it takes to initiate treatment.
A new analysis of a commercially available deep learning-based CAD and ENS indicates that use of the system can nearly halve the time to treatment (TTT) after signs of pneumothorax are identified on chest radiographs. By expediting communication between radiologists and referring providers when suspicious findings are identified, the system has great potential to improve clinical outcomes in real world scenarios, authors of the analysis suggested.
“Pneumothorax (PTX) is among the conditions recommended by the American College of Radiology for immediate communication by radiologists to physicians within minutes to avoid patient decompensation. However, in clinical practice, it is not always possible for radiologists to promptly and accurately interpret CXRs, which can lead to delays in diagnosing PTX,” Si Nae Oh, MD, with the Department of Family Medicine at the National Health Insurance Service Ilsan Hospital in the Republic of Korea, and co-authors explained.
Numerous studies have signaled that CAD systems can improve detection rates in time sensitive cases, such as those related to PTX. But less is known about how these improvements affect the timeline of patient management, the group noted.
Researchers sought to determine the system’s utility in both spotting PTX and initiating communication between providers in real clinical settings. To do this, they implemented the duo in an 818-bed general hospital, having 33 physicians and their residents utilize both CAD and ENS and another 155 providers use CAD only. Time-to-treatment figures from before and after implementation were compared to determine whether the combination of programs resulted in improvements in care.
More than 600,000 chest X-rays were included in the analysis, with just 2.0% positive for PTX. Compared to the CAD only group, the CAD and ENS group saw a significant reduction in the time-to-treatment with supplemental oxygen, dropping from 277.8 minutes to 143.8 minutes. The system did not, however, result in significant reductions in time-to-treatment for more invasive treatment options, such as aspiration and tube-thoracostomy.
The authors pointed to previous studies that analyzed the use of CAD in PTX detection, noting that those assessments did not show a reduction in time-to-treatment. While CAD has proven to be beneficial for detecting PTX, CAD alone can only do so much for improving outcomes. Incorporating an electronic alert system improves the communication of findings, leading to expedited treatment and therefor outcomes as well.
The group suggested that future work could analyze the system as an independent reader in clinical settings.
The study abstract is available in the Journal of the American College of Radiology.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.