“As COVID-19 continues to rapidly spread around the world, it is critical that we have all the solutions on the table, as healthcare systems and providers may become overwhelmed with symptomatic patients with severe or worsening respiratory status,” said Moshe Becker, CEO and co-founder of RADLogics, in an interview with MD+DI. He explained that there will also be the need to monitor patients with functional impairment and/or hypoxemia after they recover from COVID-19.
To that end, RADLogics has developed an AI-powered system that supports chest imaging for the medical triage of patients with a high probability of COVID-19, and symptomatic patients with severe or worsening respiratory status. A recent study found that this CT image analysis algorithm achieved a high level of accuracy for detecting COVID-19 on CT. The study found that the algorithm could differentiate 157 patients with and without COVID-19 with a 0.996 AUC (plus, 98.2% sensitivity and 92.2% specificity).
“The algorithm is supported by our own workflow software program that enables rapid deployment of the solution at multiple hospitals, using commercial cloud-computing resources through Amazon Web Services that can process up to one million CT studies per day,” said Becker. “It integrates within minutes into a comprehensive, seamless, and secure workflow to augment acute-care teams with deep clinical insight and actionable data within seconds,” he continued, noting that it can also be installed on premise. The system requires no extra steps, and the findings slide into the current workflow, so radiologists do not have to leave their current screens to review the results.
Although non-contrast thoracic CT is not recommended as a front-line test, the system aims to improve patient care by providing doctors with a suggested “Corona Score,” which measures the percentage of lung volume that is infected with disease. “In turn, this provides caregivers with a better sense of the seriousness of the illness, thus allowing doctors to better triage the patients based on those that urgently need intensive care, such as ventilator support,” said Becker.
To date, RADLogics has deployed this AI-powered imaging solution in China, Russia, and Italy. Since the initial study, the company has processed hundreds of more patient cases in these countries, and has found similar results and performance. “In addition, based on the results to date, we’ve seen significant demand,” said Becker. “We’ve engaged hospitals and providers on six of seven continents (with the exception of Antarctica). To meet this demand, we have actively engaged major OEM distribution partners and hope to have distribution agreements in place soon that will enable more rapid deployment of our solution.”
According to Becker, RADLogics has applied for an Emergency Use Authorization with FDA, and he said the company is in active dialogue with the agency. “In addition, we have taken a similar approach with the EU to secure a CE Mark, and we are also engaging priority countries directly,” Becker said.
Results of the study are available on arXiv.org, and it has been submitted to the Radiology Society of North America for review.