Assessing Progression of Early-Stage Lung Adenocarcinoma Using Radiomic Trajectory Analysis
Overview
RadioTrace, a deep contrastive learning framework integrating radiomic and pathological data, quantitatively measures early-stage lung adenocarcinoma (esLUAD) progression. It predicts tumor phenotypes such as STAS and lymph node metastasis, serves as an independent prognostic factor, and reveals survival heterogeneity within pathological grades.
Background
Lung adenocarcinoma (LUAD) is the most common subtype of lung cancer and progresses through defined pathological stages. Current histopathological grading, while standard, is invasive, subjective, and may overlook tumor heterogeneity and continuous progression. Computed tomography (CT) imaging combined with radiomics offers a non-invasive approach to characterize tumor progression, but existing methods lack continuous quantification beyond pathological grades. There is a clinical need for tools that integrate radiomic and pathological information to better assess esLUAD progression and improve prognosis accuracy.
Data Highlights
RadioTrace was validated across four multi-institutional cohorts, demonstrating significant prediction of tumor phenotypes including spread through air spaces (STAS) and lymph node metastasis (LNM). Survival analyses showed RadioTrace as an independent prognostic factor with log-rank test p-values < 0.004 across all cohorts. Within the same pathological grade, RadioTrace revealed significant survival heterogeneity (p < 0.02). Radiogenomic analyses linked RadioTrace scores with progression-related gene alterations and expression. Longitudinal CT follow-ups confirmed consistency of RadioTrace with continuous tumor progression.
Key Findings
RadioTrace integrates radiomic and pathological data to generate a continuous pseudo-progression score (PPS) quantifying esLUAD progression.
PPS predicts critical tumor phenotypes such as spread through air spaces (STAS) and lymph node metastasis (LNM).
RadioTrace serves as an independent prognostic factor for patient survival across multiple cohorts (p < 0.004).
Within identical pathological grades, RadioTrace identifies significant survival heterogeneity, highlighting limitations of current grading systems.
Radiogenomic analyses demonstrate associations between RadioTrace scores and tumor progression-related molecular features.
Longitudinal analyses show RadioTrace scores correlate with continuous tumor progression in patients over time.
Clinical Implications
RadioTrace offers a non-invasive, quantitative tool to assess esLUAD progression beyond conventional histopathological grading, enabling more precise risk stratification and prognosis. Its ability to detect heterogeneity within pathological grades can guide personalized treatment decisions. Integration of radiomic trajectory analysis into clinical workflows may improve early detection of aggressive tumor phenotypes and optimize patient management.
Conclusion
RadioTrace provides a novel, interpretable radiomic trajectory that quantitatively captures esLUAD progression, surpassing traditional pathological grading in prognostic value. This approach holds promise for enhancing clinical decision-making and tailoring treatment strategies in early-stage lung adenocarcinoma.
References
Original Study -- Assessing Progression of Early-Stage Lung Adenocarcinoma Using a Radiomic Trajectory Analysis
