Improving Patient Safety Through Smarter IVC Filter Follow-Up
R&E Foundation grant demonstrates how an AI-powered alert system increases retrieval rates and reduces long-term complications
Inferior vena cava (IVC) filters can prevent life-threatening pulmonary embolism (PE) in patients with deep venous thrombosis (DVT) who have contraindications to anticoagulation.
When no longer indicated, IVC filter retrieval is recommended to reduce potential long-term complications. However, IVC filter retrieval rates remain low in the overall population.
“Persistently low IVC filter retrieval rates are driven by both systemic and clinical gaps. Systemically, many institutions lack standardized follow-up protocols, sufficient staffing, and dedicated clinic visits, making it difficult to track patients and ensure timely filter removal,” said Adam Fang, MD, vascular and interventional radiologist at MedStar Health and clinical associate professor of radiology at MedStar Georgetown University Hospital in Washington, DC.
“Clinically, patients and referring physicians are often unaware that IVC filters are intended for temporary use and that leaving them in place can cause complications," Dr. Fang said. "Existing follow-up mechanisms are insufficient because they rely on informal communication and do not provide structured reminders or active patient tracking, leading to many patients being lost to follow-up and contributing to persistently low retrieval rates.”
To address this, Dr. Fang and his team developed a Filter Alert System (FAS) using AI with support from a 2022 Joan Eliasoph, MD/RSNA R&E Foundation Research Seed Grant. This platform applied real-time natural language processing to identify patients with IVC filters who presented to their hospital.
“Unlike traditional methods, which rely on manual chart reviews or clinician follow-up and can be time-consuming or prone to missed cases, FAS automatically scans radiology reports and sends real-time alerts to a secure email account,” Dr. Fang said. “This automation enables more efficient identification of patients who may need follow-up care.”
Using AI to Find What Falls Through the Cracks
The team first validated FAS using a test dataset, then integrated it into clinical workflows and evaluated its impact on follow-up, retrieval rates, dwell times and complications.
“The AI-driven FAS can identify patients with IVC filters with remarkable accuracy, detecting them in 99.7% of finalized CT reports, with strong sensitivity (85.7%), specificity (99.9%) and positive predictive value (94.7%),” Dr. Fang said.
Integrating the FAS into routine clinical workflows has been shown to significantly improve IVC filter follow-up and retrieval rates. Over a nine-month period, a standardized protocol was used to identify eligible patients and coordinate follow-up with providers and patients. During this time, follow-up appointments rose from 21.6% before FAS integration to 58.0%, while filter retrievals increased from 21.6% to 44.0%.
“These findings demonstrate that incorporating FAS into clinical practice can enhance patient monitoring, facilitate timely filter removal and reduce the number of patients lost to follow-up,” Dr. Fang said.
Early identification and timely retrieval of IVC filters offer important benefits for patients. Prompt removal reduces the risk of serious complications, including filter fracture, migration, embolization, vascular injury and venous thrombosis. Filters left in place for extended periods can also make retrieval more technically challenging, sometimes requiring advanced procedures.
“By facilitating earlier removal, patients face fewer long-term risks and improved safety. These practices align with guidance from the U.S. Food and Drug Administration recommending that IVC filters be removed once protection from pulmonary embolism is no longer necessary,” Dr. Fang said.“These developments point to a broader shift toward AI-supported, data-driven care models that enhance clinical workflows and raise the standard of patient management across health care systems.”
— ADAM FANG, MD
Automating Follow-Up for Temporary Implantable Devices
The FAS highlights the potential of AI-driven tools to transform patient monitoring and follow-up for implantable devices. By providing a standardized, automated, real-time method to identify patients in need of care, FAS improves efficiency and patient safety.
“Its application could extend beyond IVC filters to other implantable or temporary medical devices, enabling proactive monitoring, timely removal and more effective use of health care resources,” Dr. Fang added. “These developments point to a broader shift toward AI-supported, data-driven care models that enhance clinical workflows and raise the standard of patient management across health care systems.”
Dr. Fang credits the R&E Foundation grant for supporting this research.
“It provided me with essential funding to collect preliminary data, enabling me to explore novel research questions and gain practical experience in study design, data collection and analysis. The grant also offered opportunities for mentorship and exposure to a collaborative research environment, which shaped my approach to clinical investigations,” Dr. Fang concluded. “Without this support, the initial work and insights that now form the foundation for larger, prospective studies would not have been possible.”
For More Information
Learn more about R&E Foundation funding opportunities.
Read previous RSNA News stories on R&E Foundation-funded research: