PCA-based respiratory motion modeling for individualized PTV margin optimization in NSCLC radiotherapy
Huy Dang Quang, Cong Nguyen Thanh, Trang Hoang Thi Kieu, Tao Chau Van
Journal of applied clinical Medical Physics, Vol 27, Issue 6(2026)
Abstract:
Background: Respiratory motion remains a major challenge in radiotherapy for non-small cell lung cancer (NSCLC), often requiring a balance between target coverage and normal tissue sparing.
Introduction: This study aimed to develop and validate a clinically imple-mentable, individualized PTV margin strategy by integrating principal component analysis (PCA) with key clinical predictors.
Methods: A cohort of 61 NSCLC patients (Stage III) underwent 4DCT simulation. 4DCT datasets were used both for motion modeling and as a reference planning approach (ITV-based). The proposed method was compared against conventional 3DCT and 4DCT-based plans. PCA was applied to the deformable vector fields to extract dominant motion modes and quantify directional amplitudes (Amean). Clinical correlations with tumor motion were assessed. An individualized margin recipe, Md =√(2.5 Σ)2 + (0.7𝜎)2 + (𝛼 ∗ Fpos ∗ FT ∗ Amean)2 was implemented, incorporating a Location Factor Fpos and a T stage Factor (FT ) derived from cohort-specific data. Dosimetric plans (3DCT, Individualized-3DCT and 4DCT) were compared using QUANTEC constraints.
Results: Tumor location (p = 0.007) and T stage (p = 0.04) were identified as significant predictors of S–I motion amplitude. Middle/lower lobe tumors and early T1–T2 stages exhibited higher mobility (≥ 15 mm). By applying adaptive weighting factors (Fpos= 1.2 for lower lobes; FT = 0.9 for T3–T4), the individualzed 3D plans achieved a significant reduction in PTV volume compared to 3D plans (410.0 ± 170.0 cm3 vs. 460.2 ± 179.1 cm3, p = 0.002). The individualized margin approach reduced lung dose compared with conventional 3DCT planning (V20: 30.0% vs. 40.5%), while remaining comparable to the 4DCT-based reference plan (28.9%). Other OARs remained within tolerance limits.
Conclusions: Integrating PCA-derived motion with clinical weighting provides a practical framework for individualized margin design, improving lung sparing while maintaining target coverage.
More detail >>
