An AI-assisted lung cancer screening program increased screening uptake and led to earlier-stage cancer detection across urban and rural centers, according to findings from a 6-year analysis presented at the 2026 ASCO Annual Meeting (Abstract 109).  

“This study demonstrated that an AI-assisted lung cancer screening program can indeed improve screening uptake and improve early-stage cancer detection,” said Jun Zhang, MD, PhD, of the OSF HealthCare Cancer Institute in Peoria, Illinois. “A multifaceted approach combining centralized navigation and innovative screening technologies can further improve screening uptake.” 

Background 

Although low-dose CT screening has been shown to reduce lung cancer mortality among high-risk individuals, screening uptake nationwide remains well below the 40% to 50% threshold needed to meaningfully reduce mortality at the population level.  

Dr. Zhang suggested that low screening uptake may stem from uncertainty among primary care physicians about eligibility criteria, limited patient awareness, geographic and insurance-related barriers to access, and organizational fragmentation that can impede coordination between screening, diagnosis, and treatment. 

The research team hypothesized that a multifaceted, phased, AI-integrated health system intervention could achieve high, sustainable lung cancer screening rates while promoting earlier-stage diagnosis.  

Study Methods 

The OSF HealthCare project combined AI-assisted electronic health record alerts, centralized navigator coordination, quality dashboard integration, radiology workflow automation, and blood-based screening.  

Because the health-care system spans a broad geographical area across Illinois and Michigan—including seven urban tertiary hubs and nine rural community access sites—AI-enabled tools were used  to support system-wide implementation of the program. To extend screening services in rural areas, the system adopted a multifaceted hub-and-spoke model that centralized resources and operations.  

Interventions were implemented in phases between October 2019 and late 2023. The first phase involved consolidating navigators into a centralized team responsible for managing screening registries, patient scheduling, and care coordination. In mid-2021, the health-care system adopted the Radloop natural language processing engine to automate actions based on radiology reports and American College of Radiology Lung-RADS classification. Later that year, AI-driven electronic health record alerts were deployed to identify high-risk patients eligible for screening. In early 2022, OSF HealthCare adopted the expanded U.S. Preventive Services Task Force (USPSTF) eligibility criteria for lung cancer screening, which included adults aged 50 to 80 years with a smoking history of at least 20 pack-years who currently smoked or had quit within the previous 15 years. By mid-2022, screening metrics had been integrated into the health-care system’s Ambulatory Quality Dashboard, providing real-time performance data at both the facility and provider levels. Finally, in late 2023, the system launched a pilot program using the FirstLook Lung blood-based ctDNA screening test from DELFI Diagnostics, targeting high-risk individuals and those who had not complied with recommended screening.  

Findings 

Between 2020 and 2025, the lung cancer screening rate at OSF HealthCare increased from 18.2% to 42.8%. Over the same period, the national screening rate, based on American Lung Association annual reports, rose from 14.5% to 19.5%. By 2025, OSF HealthCare’s screening rate exceeded the national rate by 23.3 percentage points.  

Screening volumes across the OSF HealthCare system increased throughout the study period, despite disruptions associated with the COVID-19 pandemic. Although screening rates briefly declined during the initial lockdown in 2020 and subsequent variant surges in 2021 and 2022, overall screening volumes continued to rise. 

Dr. Zhang noted that lung cancer screening rates within the OSF HealthCare system continued to grow year over year, outpacing both the national and Illinois state averages despite the challenges posed by the COVID-19 pandemic.  

“As we know, if early screening is effective, then we should see stage migration, meaning that we're going to detect more early-state lung cancer and less late-stage lung cancer,” Dr. Zhang explained. “In fact, that's the case.” 

The proportion of lung cancers diagnosed at an early stage (stage I) increased from 30.9% in 2020 to 44.6% in 2025, while the proportion diagnosed at a late stage (stage IV) decreased from 41.0% to 34.2%. This corresponded to an annual increase of 2.4 percentage points in stage I diagnoses and an annual decrease of 1.18 percentage points in stage IV diagnoses. In contrast, national averages changed more modestly during the same period. By 2025, stage I disease accounted for 28.1% of lung cancers nationally, whereas stage IV disease accounted for 34.2%. 

When analyzed by setting, the proportion of early-stage lung cancers increased by 2.14 percentage points annually in urban facilities and by 4.62 percentage points annually in rural facilities. Over the same period, the proportion of late-stage lung cancers decreased by 1.11 percentage points per year in urban settings and by 1.20 percentage points per year in rural settings.  

“So, put together, we can say that both urban and rural areas in the OSF system benefitted from this early screening platform, but rural areas benefitted even more,” Dr. Zhang said. 

Expert Insights 

ASCO Discussant Caroline Chung, MD, MSc, FRCPC, CIP, Professor of Radiation Oncology and Diagnostic Imaging and Co-Director of the Institute for Data Science in Oncology at The University of Texas MD Anderson Cancer Center, placed the findings in a broader context.  

She emphasized the importance of evaluating AI tools in cancer care within their broader clinical context, rather than relying solely on model performance metrics, and she commended the authors for examining the program’s impact in urban and rural settings.  

“The time [that] was taken to compare urban vs rural is a really important piece...and when it comes to lung cancer screening it’s also clinically applicable—where are you going to put your efforts, where do these interventions lie?” Dr. Chung said. “I want to call this out in terms of really thinking through the different aspects of the data that you can dig into it terms of relevance.” 

DISCLOSURES: Dr. Zhang reported a consulting or advisory role with AstraZeneca, BridgeBio, Daiichi Sankyo, DAVA Oncology, Fosun Pharma, Hengrui, IDEOlogy, Johnson & Johnson, Novocure, and Regeneron; he also reported being on the speakers’ bureau for AstraZeneca, Daiichi Sankyo, Johnson & Johnson, MJH Life Sciences, Regeneron, and Sanofi Genzyme. Additionally, Dr. Zhang reported receiving research/support from AbbVie, Ankyra Therapeutics, Astellas, AstraZeneca, BeiGene, BMS, BridgeBio, Champions Oncology, Eli Lilly, Frontier Medicine, Genentech, InnoCare Pharma, Janssen, Kahr Medical, Merck, Mirati, Nerviano Medical Sciences, Nilogen, Novartis, and Zai Lab. No other study authors reported any relationships of interest.  

Dr. Chung reported a consulting or advisory role with Convergent Radiotherapy and Radiosurgery, and has received research funding from RaySearch Laboratories and Siemens Healthineers. Additionally, she has a patent application and has received travel expenses from Elekta.

ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO^®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.