CT-based radiomics and ablation margin quantification for predicting local tumor progression in Stage I NSCLC after radiofrequency ablation.
Authors
Affiliations (7)
Affiliations (7)
- Department of Radiology, The First Medical Center of Chinese PLA General Hospital, Beijing, China.
- Chinese PLA Medical School, Beijing, China.
- Academy for Advanced Interdisciplinary Science and Technology, Zhejiang University of Technology, Hangzhou, China.
- Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications, School of Computer Science and Engineering, Southeast University, Nanjing, China.
- Department of Ultrasound, The Fourth Affiliated Hospital, Zhejiang University School of Medicine, Yiwu, China.
- Department of Radiology, The Sixth Medical Center of Chinese PLA General Hospital, Beijing, China.
- Research and Development Department, Zoye Corporation, Qingdao, China.
Abstract
Local tumor progression (LTP) compromises durable local control after radiofrequency ablation (RFA) for stage I non-small cell lung cancer. We developed a three-dimensional minimum ablation margin (3D MAM) quantification workflow and evaluated its integration with post-ablation CT radiomics for individualized LTP prediction. This dual-center retrospective study included 217 patients divided into training, internal validation, and external validation cohorts. Automated segmentation, deformable image registration, and Visualization Toolkit (VTK)-based analysis enabled quantitative 3D margin assessment. An insufficient MAM ( <math xmlns="http://www.w3.org/1998/Math/MathML"><mo><</mo> <mn>5</mn></math> mm) was identified in 147 patients (67.7%). MAM <math xmlns="http://www.w3.org/1998/Math/MathML"><mo>≥</mo> <mn>5</mn></math> mm was independently associated with lower odds of LTP (odds ratio, 0.06; 95% confidence interval, 0.02-0.16). Among nine machine-learning classifiers, the random forest model showed consistent discrimination across cohorts, with areas under the curve of 0.877, 0.851, and 0.889. Combining quantitative margin assessment with CT radiomics may support risk-adapted surveillance after lung RFA.