Many Lung Cancer Patients Miss Screening Window

“Most patients who ultimately develop lung cancer would not have qualified for low-dose CT (LDCT) screening under current U.S. Preventive Servies Task Force (UPSTF) criteria,” said Omar Cespedes-Gomez, MD, MPH, in a Sunday session examining the real-world reach of lung cancer screening.

Dr. Cespedes-Gomez, whose co-authors include Maedeh Sharifian, MD, and Gelareh Sadigh, MD, from the Department of Radiological Sciences at the University of California, Irvine Medical School, noted that their study findings highlight major challenges in both guideline scope and electronic medical record (EMR) documentation.

Incomplete Data Hampers Early Detection

The team retrospectively analyzed data from 1,930 patients from 50 to 80 years of age who were diagnosed with lung cancer between 2021 and 2024. They assessed which patients would have qualified LDCT lung cancer screening based on the 2021 UPSTF criteria and evaluated how completely smoking histories were recorded in the electronic medical record (EMR).

“Only 13.2% of patients who ultimately developed lung cancer met the 2021 USPSTF eligibility criteria at the time of diagnosis,” Dr. Cespedes-Gomez said. “Among those eligible, only 9.4% had completed a screening LDCT within the two prior years.” “Even among those identified as eligible, very few actually underwent screening before diagnosis, highlighting how challenging it is to identify at-risk individuals”

Omar Cespedes-Gomez, MD, MPH

“Only 61% of patients had a structured smoking history documented,” Dr. Cespedes-Gomez said. “Among those, just 13.2% met screening eligibility criteria, and a mere 9.4% of those eligible had undergone LDCT screening in the two years before their diagnosis.”

The findings underscore how limitations in both data capture and guideline design may prevent early detection. “More than 87% of patients lacked the data needed to assess screening eligibility,” Dr. Cespedes-Gomez said. “This reveals a major and unexpected gap in EMR data completeness given how heavily current guidelines depend on smoking exposure to define risk.”

Drawing attention to demographic factors, Dr. Cespedes-Gomez noted that female, Asian and insured patients were more likely to have complete structured smoking data documented, reflecting a documentation bias. 

Conversely, when assessing eligibility, female, Asian and Hispanic patients were less likely to meet the 2021 USPSTF screening criteria, reflecting structural disparity. “These disparities remind us that gaps in EMR data quality and criteria design often mirror broader inequities in health care access, communication and documentation,” he added. 

Incomplete data not only hinders risk assessment but also weakens the effectiveness of EMR-based screening tools. “Accurate smoking data enable EMR systems to automatically identify patients who qualify for screening and trigger alerts or order prompts,” Dr. Cespedes-Gomez said. “Improving this documentation transforms the EMR into a proactive, population-level screening tool that ensures all eligible patients are systematically identified and offered screening.”

The research team is expanding their analysis nationally and developing AI-driven methods to improve documentation. “We are working on natural language processing (NLP)-enabled EMR verification that uses natural language processing to automatically extract smoking history from unstructured clinical notes,” Dr. Cespedes-Gomez said. “By converting free-text information into structured data, we can better identify patients eligible for lung cancer screening.

Access the session, “Eligibility for Lung Cancer Screening Among Patients Diagnosed With Lung Cancer: A Retrospective Analysis of EMR Documentation and USPSTF Criteria,” (S4-SSCH02-2-2) at RSNA.org/MeetingCentral