Volume-based quantitative assessment of respiratory-induced organ motion in lung cancers

To comprehensively evaluate respiratory-induced organ motion and deformation using volume-based analysis, overcoming the limitations of conventional point-based methods in radiation therapy for lung malignancies.

The Internal target volume (ITV) and organs at risk (OARs) were delineated on averaged-phase images from 4D-CT datasets of 11 lung cancer patients. Deformable image registration using an enhanced non-parametric algorithm generated deformation vector fields (DVFs) between each respiratory phase and a reference phase. Voxel-level displacements within each structure were statistically analyzed, quantifying ITV and OAR motion via geometric centroid displacement, mean voxel displacement, and 80th/90th percentile displacements. Pearson correlation analysis assessed the relationship between ITV motion and volumetric deformation, while one-way repeated-measures ANOVA evaluated directional consistency of OAR centroid and voxel-average displacements across LR, AP, and SI axes (p ≤0.05).

Superior-inferior (SI) displacement of ITVs exceeded lateral and AP motion by 2-6 times. Significant centroid-voxel average motion discrepancies in the left lower lobe, right upper lobe, and diaphragm (p = 0.05) indicate limitations of point-based measures. The 80th percentile voxel motion showed potential as a centroid surrogate. Diaphragm and mediastinum deformation reached 60-80 %, with ITV deformation spanning 5-70 %. Motion metrics correlated strongly within structures but weakly with deformation.

This approach enables comprehensive quantification of respiratory motion, supporting optimized motion management in lung cancer radiotherapy. As conventional metrics fail to capture organ deformation, integrating both motion and deformation assessments is essential for accurately characterizing respiratory-induced target variations.