Integrated CT Radiomics and Circulating Tumor Cell Analysis in Predicting Lung Adenocarcinoma Invasion: A Dual-Center...

Area of Science:

• Oncology
• Radiology
• Medical Informatics

Background:

• Lung adenocarcinoma invasiveness assessment is crucial for early-stage diagnosis and treatment planning.
• Distinguishing between minimally invasive adenocarcinoma (MIA) and invasive adenocarcinoma (IAC) remains a clinical challenge.
• Integrating diverse data sources can improve diagnostic accuracy.

Purpose of the Study:

• To develop and validate a risk prediction model for lung adenocarcinoma invasiveness.
• To discriminate between MIA and IAC using radiomics, circulating tumor cells (CTCs), and clinical data.
• To assess the clinical utility of the composite model in early-stage lung cancer diagnosis.

Main Methods:

• Retrospective analysis of 202 lung adenocarcinoma patients from two medical centers.
• Machine learning applied to CT radiomic features and preoperative CTC counts.
• Development of three models: radiomics-only, CTCs-clinical data, and a composite clinical-radiomics-CTCs model.

Main Results:

• The composite model achieved the highest predictive accuracy with an AUC of 0.980 (95% CI: 0.960-1.000).
• The CTCs-clinical model showed superior performance (AUC: 0.960, 95% CI: 0.926-0.994).
• The composite model outperformed single-modality approaches in clinical predictive capability.

Conclusions:

• The radiomics approach effectively discriminates between MIA and IAC in early-stage lung adenocarcinoma.
• The composite clinical-radiomics-CTCs model provides a novel auxiliary diagnostic method.
• This integrated model aids in evaluating invasiveness risk for better treatment strategies.