Incidental findings are health issues found by accident during medical scans done for other reasons. For example, if a patient has a CT scan after an injury, doctors might find lung nodules or signs of inflammation unrelated to the injury. These findings can help doctors diagnose and treat diseases earlier than usual.<\/p>\n
<\/p>\n
About 30% of chest CT scans show unexpected lung nodules. But across the United States, only about 30% of these cases get proper follow-up. This gap puts patients at risk of missing early signs of lung cancer, which is the leading cause of cancer death in the U.S. In 2021, over 131,000 people were expected to die from lung cancer. Finding lung cancer early allows treatments that can increase survival rates, which are very low when cancer is found late.<\/p>\n
Two health systems \u2014 Sutter Health in California and Atrium Health in North Carolina \u2014 have created programs to find lung cancer early by tracking incidental findings in CT scans.<\/p>\n
<\/p>\n
Sutter Health’s Approach<\/strong> <\/p>\n This program doubled early lung cancer detection, with over 70% of cases now found in early stages (I or II). Early detection helps because treatments like surgery or targeted radiation work better then.<\/p>\n <\/p>\n For difficult cases, a team of specialists discusses the best plan. Nurse navigators contact patients, make sure they keep appointments, and set up referrals. This team work helps patients follow care plans and feel less worried.<\/p>\n <\/p>\n In 2024, Sutter Health also used a similar AI system to find nearly 200 brain aneurysms by accident, showing this method can help beyond lung health.<\/p>\n <\/p>\n Atrium Health’s Efforts<\/strong> <\/p>\n After starting this system, Atrium Health received about 2,000 alerts and found 488 patients needing more tests; about 70 patients saw lung specialists. They found some early-stage lung cancers this way.<\/p>\n <\/p>\n Atrium Health also started using robotic bronchoscopy, a less invasive tool that helps doctors take accurate biopsies from hard-to-reach nodules. This lowers risks compared to traditional methods and improves sampling accuracy.<\/p>\n Besides lung cancer, AI technology now helps find other diseases automatically during imaging.<\/p>\n <\/p>\n Sacroiliitis Detection via AI in CT Scans<\/strong> <\/p>\n Researchers in Israel created SIJ-grade, a fully automatic AI system that detects and grades sacroiliitis from CT scans taken for other reasons. The system uses different AI tools to find the sacroiliac joint and measure disease severity.<\/p>\n <\/p>\n Testing showed the system was 92% accurate and had 95% sensitivity, sometimes doing better than human experts. This technology can find sacroiliitis earlier without extra scans or radiation.<\/p>\n Automation in detecting incidental findings and managing follow-up is key to turning AI results into better care. These systems work best when AI tools connect smoothly with electronic health records (EHR) and communication systems in healthcare.<\/p>\n <\/p>\n Enhancing Radiologist and Clinician Workflows<\/strong> <\/p>\n AI systems also send closed-loop alerts for follow-ups, remind teams automatically, prioritize cases by risk, and assign tasks to nurse navigators and specialists. This process helps make sure no patients are forgotten.<\/p>\n <\/p>\n The Impact on Clinician Retention and Efficiency<\/strong> <\/p>\n Coordination Through Multidisciplinary Teams<\/strong> <\/p>\n Managing Incidental Findings of Varying Significance<\/strong> <\/p>\n Rules like the European Union AI Act and US FDA guidelines require AI tools to be clear and controlled by humans, keeping patient safety in mind. Tools like MorphEUS help doctors decide when to act on uncertain findings and when to wait and watch.<\/p>\n <\/p>\n
Sutter Health trained over 500 radiologists in 28 hospitals and 80 imaging centers to carefully watch for lung nodules found during scans. They used AI tools linked with Epic\u2019s Result Tracker to alert doctors about nodules as small as 5 mm. The alerts follow rules from the Fleischner Society, which recommends how to handle nodules based on their size and patient risk.<\/p>\n
Atrium Health uses AI software that scans radiology reports for words like \u201clung mass\u201d or \u201cpulmonary lesion.\u201d Nurse navigators check these alerts daily to decide if patients need follow-up. Small nodules (under 6 mm) are watched by primary care doctors. Larger nodules go to lung specialists.<\/p>\nAdvances in Automatic Detection of Other Diseases Through Imaging<\/h2>\n
Sacroiliitis is a joint inflammation connected to a disease called ankylosing spondylitis. It causes back pain and disability, especially in young adults. Early diagnosis is important but hard because its symptoms can look like common back pain. It often takes about seven years to diagnose correctly.<\/p>\nThe Role of AI and Workflow Automations in Early Disease Detection<\/h2>\n
AI helps radiologists by marking small signs of disease during image review. For example, John Muir Health and UPMC use AI with ambient listening to save clinicians time on charting. This cuts down on paperwork after work hours, helping reduce burnout.<\/p>\n
John Muir Health saw a 44% drop in doctor turnover after using AI tools for documentation and workflows. Spartanburg Regional Healthcare saved 9,000 nursing hours per year by involving nursing leaders in setting up AI-driven workflows. These changes help not just patient care but also staff satisfaction and work speed.<\/p>\n
Standard communication and follow-up rules with AI help reduce delays and mistakes. Sutter Health uses multidisciplinary tumor boards to plan tough cases, helped by AI tracking. Nurse navigators schedule visits and contact patients using AI tools, helping to keep follow-ups on track.<\/p>\n
More incidental findings have raised questions about how to handle uncertain results, sometimes called \u201cincidentalomas.\u201d These can cause worry, extra tests, or costs that may not be needed.<\/p>\n