Integrative modeling of longitudinal cell-free DNA and tumor volume dynamics: a multimodal quantitative prognostic framework

This study proposes a novel prognostic prediction framework that integrates longitudinal changes in cell-free DNA (cfDNA) and CT tumor volume following radiotherapy (RT) in patients with locally advanced NSCLC. The research team divided the cfDNA collection period into early phase and late phase based on day 3, and calculated kinetic parameters such as velocity and acceleration. In this process, Aview (version 1.0.23) software played a role in acquiring the region of interest (ROI) for each target lesion, through which 14 quantitative CT volumetric features were automatically extracted. Analysis results showed that the K-means clustering model integrating cfDNA and CT volume indicators distinguished responders from non-responders in terms of progression-free survival (PFS) much more effectively than the conventional RECIST v1.1-based approach. In conclusion, this model is presented as a precise alternative for monitoring treatment response and assessing prognosis in NSCLC. These findings highlight the value of integrating imaging biomarkers with molecular data, enabling more comprehensive and dynamic assessment of treatment response beyond conventional size-based criteria. In this study, Aview software was utilized to define regions of interest and extract quantitative CT volumetric features, supporting reproducible imaging analysis within multimodal prognostic modeling.