Segmented jaw-locking in IMRT for upper thoracic esophageal cancer: A plan complexity-driven approach to lung dose...

Extended-field intensity-modulated radiotherapy (EF-IMRT) used for elective nodal irradiation (ENI) in upper thoracic esophageal cancer frequently results in excessive intermediate-to-low-dose pulmonary irradiation. This study presented a practical and reproducible planning strategy (segmented Jaw-locking IMRT, SJL-IMRT) for optimizing dose distributions to ENI target volumes, and assessed its impact on plan complexity and lung parenchyma sparing.

In a paired planning study (n = 40), EF-IMRT and SJL-IMRT plans were generated per patient under identical target coverage objectives. SJL-IMRT partitioned the longitudinal ENI volume into cervical-supraclavicular and upper-mediastinal segments via coordinated orthogonal collimators, and multi-leaf collimator (MLC)-defined apertures locked within each segment. The evaluation was based on multifaceted criteria, including metrics for: (i) plan complexity: the aperture-based edge-area metric (EAM) and the sequence-level modulation complexity score (MCS); (ii) dosimetry: conformity index (CI), homogeneity index (HI), gradient measure (GM), pulmonary parameters (mean lung dose [MLD], V₅∼V₃₀), and spinal cord maximum dose (D_max); (iii) radiobiological effects: tumor control probability (TCP) and normal tissue complication probability (NTCP). The delivery accuracy of SJL-IMRT and EF-IMRT was validated with a PTW OCTAVIUS 729 2D ionization chamber array and RW3 phantom, using γ analysis (3.0%/3.0 mm criterion, global normalization, 10% dose threshold). Statistical analysis was performed using Wilcoxon signed-rank test.

Both SJL-IMRT and EF-IMRT satisfied prescription dose objectives for planning target volume (PTV). No significant differences were observed in CI and HI (p = 0.347 and p = 0.173, respectively). SJL-IMRT demonstrated lower geometric complexity and simpler sequencing: EAM 8.610 ± 2.951 vs. 20.824 ± 4.944 (paired Δ = -12.214; p = 0.00195) and MCS 0.243 ± 0.015 vs. 0.203 ± 0.036 (paired Δ = +0.040; p = 0.0193), respectively. Compared with EF-IMRT, SJL-IMRT (i) reduced gradient measures (GM) by 0.215 cm (p < 0.01), indicating a steeper dose fall-off; (ii) decreased MLD by 1.62 Gy (left) and 2.64 Gy (right); (iii) lowered left lung V₅ and V₂₀ by 19.96%, 3.63%, right lung V₅ and V₂₀ by 25.27%, 3.05%, respectively (all p < 0.01); (iv) exhibited a marginally higher mean γ passing rate (99.4 ± 0.72% vs. 98.8 ± 0.75%, p = 0.048), no dose cold/hot spots in the 2-cm feathered overlap, and a more compact high-dose area; (v) demonstrated a longer mean delivery time (10.9 vs. 8.8 min, p = 0.021) and higher total monitor units (MUs: 2893 vs. 2066; p = 0.026). (vi) translated these dosimetric gains into significant NTCP reductions for both lungs (left -1.30%, right -1.29%; p < 0.01). TCP values were numerically higher for SJL-IMRT, but these differences were not statistically significant; D_max to the spinal cord followed the same pattern.

SJL-IMRT reduces plan complexity (EAM↓, MCS↑) and lowers intermediate-to-low lung dose compared with EF-IMRT for ENI in upper thoracic esophageal cancer, without compromising target coverage, supporting SJL-IMRT as a pragmatic approach to improving dosimetric quality and delivery simplicity. Confirmation in larger cohorts is warranted.