In never-smokers, deep learning assessment of single baseline low-dose computed tomography (CT) scans demonstrated a 79 percent AUC for predicting lung cancer up to six years later, according to new research presented today at the American Thoracic Society (ATS) 2025 International Conference.
Utilizing single low-dose computed tomography (LDCT) scans, a deep learning model may offer significant prognostic capability for predicting future lung cancer risk up to six years, even in non-smokers, according to research presented today at the American Thoracic Society (ATS) 2025 International Conference.
For the study, researchers assessed the use of the deep learning model Sybil for predicting lung cancer risk in an initial cohort of 21,087 Asian individuals (ranging between 50 to 80 years of age) who had baseline LDCT screening scans, conducted from January 2009 to December 2021, with follow-up continuing until June 2024. A subsequent stratified analysis included 4,611 people with > 20 pack years of smoking history, 5,378 ever-smokers with < 20 or unknown pack years and 11,098 never smokers, according to the study.
“(The deep learning model) Sybil demonstrated good performance in predicting future lung cancer in an Asian screening cohort comprised of individuals with differing risk profiles. Our findings suggest the potential to develop personalized LCS strategies using Sybil, especially for the low-risk group in this population,” noted Yeon Wook Kim, M.D., a pulmonologist affiliated with the Seoul National University Bundang Hospital in Seongnam, South Korea, and colleagues.
Overall, the study authors found that deep learning assessment of single baseline LDCT scans achieved an 86 percent AUC at one year and a 74 percent AUC at six years for predicting lung cancer.
“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,” noted Yeon Wook Kim, M.D., a pulmonologist affiliated with the Seoul National University Bundang Hospital in Seongnam, South Korea.
For never smokers, the deep learning model retained the 86 percent AUC at one year and offered a 79 percent AUC for predicting lung cancer at six years, according to the study authors. Out of the 257 study participants diagnosed with lung cancer within a six-year period from the original LDCT scan, the researchers pointed out that 115 people in this group were never smokers.
“Sybil demonstrated good performance in predicting future lung cancer in an Asian screening cohort comprised of individuals with differing risk profiles. Our findings suggest the potential to develop personalized LCS strategies using Sybil, especially for the low-risk group in this population,” added Kim and colleagues.
(Editor’s note: For related content, see “Chest CT Research Reveals at Least One Lung Nodule in 42 Percent of Non-Smokers,” “CT Study: Modified Lung-RADS Model Offers Enhanced Prognostic Assessment of Pure Ground-Glass Nodules” and “Can AI Facilitate Single-Phase CT Acquisition for COPD Diagnosis and Staging?”)
Reference
1. Kim YW, Oh J, Park M, et al. Validation of Sybil deep learning lung cancer risk prediction model in Asian high- and low-risk individuals. Am J Respir Crit Care Med. 2025;211:A5012. https://doi.org/10.1164/ajrccm.2025.211.Abstracts.A5012 .
Considering Breast- and Lesion-Level Assessments with Mammography AI: What New Research Reveals
June 27th 2025While there was a decline of AUC for mammography AI software from breast-level assessments to lesion-level evaluation, the authors of a new study, involving 1,200 women, found that AI offered over a seven percent higher AUC for lesion-level interpretation in comparison to unassisted expert readers.
SNMMI: Can 18F-Fluciclovine PET/CT Bolster Detection of PCa Recurrence in the Prostate Bed?
June 24th 2025In an ongoing prospective study of patients with biochemical recurrence of PCa and an initial negative PSMA PET/CT, preliminary findings revealed positive 18F-fluciclovine PET/CT scans in over 54 percent of the cohort, according to a recent poster presentation at the SNMMI conference.
Could an Emerging PET Tracer be a Game Changer for Detecting Hepatocellular Carcinoma?
June 23rd 2025In addition to over 90 percent sensitivity in detecting hepatocellular carcinoma (HCC), the glypican-3 (GPC3) targeted PET tracer 68Ga-aGPC3-scFv appeared to be advantageous in identifying HCC tumors smaller than one centimeter, according to pilot study findings presented at the SNMMI conference.