Before he started using the artificial intelligence application, Jefferson Einstein, part of Jefferson Health Health System, faced the growing challenge in managing the case of pulmonary embolism.
Challenge
The growing volume of imaging and manual flows of work – common problems in many healthcare systems – limit the speed and efficiency of PE identification and triance, especially when time was critical.
Customary processes often tried to keep pace with the growing demand for imaging, which included many disciplines and modality. The staff recognized by AI can potentially fill the gaps in these work flows, ultimately helping to achieve the goal of ensuring the highest standard of patient care.
APPLICATION
“At Jefferson Einstein, our culture of innovation and commitment to remaining” on the bleeding “of progress led us to examine how AI can revolutionize P management, improving work flows, increasing diagnostic accuracy and improving patient care,” said Dr. Avi Sharma, director and collaborator of AI in Jefferson.
He said that Health It Vendor Aidoc technology appeared as an ideal tool to satisfy this critical need.
“As the Global Partners of Arc – the Digital Innovation Center at the Sheba Medical Center – and early adoploators of artificial intelligence in radiology, we have already implemented the AIDOC platform in diagnostic radiology systems,” he continued. “It gave us confidence that AIO suppliers can also optimize our flows of the work of the reacting team on lung embolism (PERT).
“AI flags of sellers suspected PES in real time, creating a proactive process by automatically identifying potential cases and notifying PET through a mobile application,” he added. “This ability enabled timely treatment and quick mobilization of multidisciplinary PET for urgent interventions.”
In addition, AIDOC technology has easily integrated with existing systems, including Fuji Synaps Pacs and Cerner EHR, ensuring minimal disruption of work flows, strengthening both diagnostic radiologists and intervention radiologists for more effective work, he noted.
“The choice of a consistent, adapting partner was crucial, and for years we grew up with Aidoc, implementing more systems to further increase patient care,” he said.
Fulfillment of the challenge
The implementation of AI technology has revolutionized how Jefferson Einstein manages PE. Pert now receives notifications via mobile alerts, enabling intervention radiologists to assess patients’ imaging and laboratory results from anywhere and at any time. Real time updates provide trouble -free cooperation between departments, significantly reducing delays.
“AI starts an improved process: he warns intervention radiologists who will then mobilize the necessary resources to distinguish between qualified patients to the operating room,” Sharma explained. “This reduced the treatment time and increased the trust of clinicians by improving communication and making decisions.
“We also included artificial intelligence in our educational efforts,” he added. “For example, residents of radiology are trained in the system during orientation, supporting early knowledge of these tools.”
RESULTS
The results so far can be divided into three main areas. First, an boost in clinically appropriate intervention.
PET intervention indicators increased by 73.8%, increasing from 0.84% (17 interventions/2 022 CTPA) before AI to 1.46% (32 interventions/2191 CTPA). CTPA is computed tomography of pulmonary angiography.
“This shows the power of artificial intelligence in effective playing critical cases and ensuring that high-risk PE patients will receive timely care,” Sharma explained. “In fact, this means faster, more precise answers when every minute matters.”
Secondly, shortening time to treatment.
“In the last study, accepted to SIR 2025, we assessed the impact of the system on shortening the time of examination to the needs of patients with acute pulmonary embolism subjected to percutaneous thrombiectomy,” Sharma explained. “The study showed that the implementation of the alert significantly reduced the total exam time by 20%-148 minutes to 119 minutes.
“The alert system improved interdisciplinary cooperation between radiology, rescue and critical teams, increasing timely care,” he continued. “In the case of patients, these minutes can mean a difference between life and long -term complications.”
Third, increased efficiency of radiology flow.
“Our radiology department recorded measurable profits in critical times of communication,” he noted. “Analyzing data on performance from early users, we showed a significant improvement in performance among radiologists. This has created a positive feedback loop, because the stories of success encourage more of our team to accept AI in their work.
“This approach includes an analysis of the implementation data before and after AI, presentation of arrangements at staff meetings, encouraging peer-to-peer discussion on the benefits of AI and adapted educational approaches,” he added.
Tips for others
Sharma said that if his peers from other hospitals and healthcare systems are considering AI, they should jump immediately.
“The best time to start using AI was yesterday,” he joked. “Focus on identifying pain points in work flows and choose AI systems that integrate with existing systems such as PACS and EHR Platforms.
“Building a culture of trust and curiosity is crucial,” he continued. “Engage early adoptions who are happy to discover new technologies and encourage them to share their success stories. Use data to demonstrate measurable results and share them widely in your organization. When colleagues see the proof of the impact of the tool, they will be much more willing to try.”
He added that ongoing education is also necessary, especially for radiologists who will shape the future of healthcare in AI-Ai-supporting AI.
“Finally, remember that AI is a long-term investment in patient care,” he summed up. “It’s not just about roi. It’s about raising the care standard around the world. AI offers a safety network that provides the highest level of care for every patient.”
Watch now: why you can’t just lose artificial intelligence on the CIO or CTO to make AI director