Single CT scan all it takes for new deep learning model to predict lung cancer risk
A new deep learning model is able to predict cancer risk using just a single low-dose CT (LDCT) lung cancer screening exam.
New research presented this week during the American Thoracic Society 2025 International Conference details the capabilities of a DL model called Sybil, which was developed with the help of data acquired during the National Lung Cancer Screening Trial (NLST). Experts believe Sybil could improve lung cancer risk stratification by offering patients more personalized screening strategies.
The DL model was tested on more than 21,000 scans of patients ages 50 to 80 who had voluntarily undergone LDCT. Scans took place between 2009 and 2021, with patient outcomes followed through 2024.
Unlike traditional risk models, Sybil was tasked with predicting who would develop lung cancer in the future based on imaging alone—a task the model was quite skilled at, according to researchers. Sybil performed well at predicting lung cancer risk at both the one- and six-year marks.
“Sybil’s value lies in its unique ability to predict future lung cancer risk from a single LDCT scan, independent of other demographic factors that are conventionally used for risk stratification,” corresponding author Yeon Wook Kim, MD, a pulmonologist and researcher at Seoul National University Bundang Hospital in Seongnam, Republic of Korea, and colleagues noted.
The team is optimistic the model could be used to more accurately screen for patients who have the highest risk of developing cancer, allowing low-risk patients to go longer between screenings.
“Sybil demonstrated the potential to identify true low-risk individuals who may benefit from discontinuing further screening, as well as to detect at-risk groups who should be encouraged to continue screening,” the group suggested.
The team added that the tool could be especially beneficial in areas where instances of lung cancer in nonsmokers are on the rise, such as Asia.
Learn more about Sybil here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.