Transl Lung Cancer Res. 2025 Nov 30;14(11):4746-4755. doi: 10.21037/tlcr-2025-940. Epub 2025 Nov 27.

ABSTRACT

BACKGROUND: Liquid biopsy based on cell-free DNA (cfDNA) in oncology has emerged as a promising technique for tracking cancer dynamics, especially for detecting minimal residual disease. To date, most studies have used cfDNA for static evaluations of tumor burden. In this study, we propose a novel approach integrating serial cfDNA and computed tomography (CT) tumor volume to fully reflect the dynamic nature of tumor response after treatment.

METHODS: This prospective study involved 25 patients treated with curative-intent radiotherapy for localized non-small cell lung cancer (NSCLC) between June 2019 and November 2020, with 17 subsequently included in final analysis. Longitudinal blood samples were divided into two phases relative to day 3 after treatment initiation, and kinetic parameters, such as velocity and acceleration of cfDNA levels, were calculated. To complement sparse samplings in later days, volume data from routine CT scans were incorporated. K-means clustering using two different variable sets (cfDNA only and cfDNA with volume parameters) and conventional assessment using Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 were applied to stratify patients, and their performance was compared.

RESULTS: The model incorporating both cfDNA and volume parameters effectively separated responders (mean progression-free survival, 44.2 months) from non-responders [16.6 months, P=0.02; area under the receiver operating characteristic curve (AUC) =0.955], outperforming cfDNA only model (36.0 vs. 14.5 months, P=0.04; AUC =0.848). In contrast, RECIST v1.1-based conventional assessment showed no significant difference (P=0.62, AUC =0.70).

CONCLUSIONS: Therefore, our study demonstrates that integration of longitudinal cfDNA and tumor volume dynamics yielded improved assessment of treatment response and prognosis in NSCLC.

PMID:41367558 | PMC:PMC12683446 | DOI:10.21037/tlcr-2025-940

J Thorac Imaging. 2026 Mar 1;41(2):e0866. doi: 10.1097/RTI.0000000000000866.

ABSTRACT

This review traces the historical path of artificial intelligence (AI) methods that have been applied to medical image interpretation. Early AI approaches, which were based on clinical expertise and domain-specific medical knowledge, established the basis for data-driven methods, initiating the radiomics era and leading to the widespread use of deep learning in medical imaging. More recently, transformer architectures-originally developed for natural language processing-have been adapted for medical image analysis. In the first section, we explore the literature on the use of AI, specifically addressing lung nodules and lung cancer. AI has been effective in detecting lung nodules, evaluating their characteristics, and predicting cancer risk, while also addressing technical issues like kernel conversion. In lung cancer, AI has been applied to various clinical needs, including prognosis evaluation, mutation identification, treatment response analysis, operability prediction, treatment-related pneumonitis, and clinical information extraction. In the following section, we explore foundation models, multimodal AI, and a multiomic approach in the field of lung nodules and lung cancer. Finally, as AI models continue to evolve, so too must the approaches for evaluating their real-world utility; thus, we outline relevant methods for evaluating the performance and application of AI in thoracic radiology.

PMID:41246950 | DOI:10.1097/RTI.0000000000000866