Session Item

The risk of Radiation-Induced Lung Injuries (RILI) is directly related to the volume of functional lung receiving at least 20 Gy (V_20Gy) and to its mean dose (D_mean). Reducing these doses by implementing functional lung imaging in radiotherapy (RT) treatment planning could reduce the risks of RILI. The aim of this study is to retrospectively investigate the feasibility of 4DCT-based functional RT treatment planning for thoracic tumors using available clinical tools.

20 selected patients underwent a 4DCT (Aquilion LB, Canon Medical Systems Europe B.V., Zoetermeer, The Netherlands) for a thoracic RT treatment between July and October 2020. Peak Inhalation (PI) and Peak Exhalation (PE) 4DCT images were used to calculate ventilation via the Jacobian method using the default intensity-based free-form deformable image registration on MIM Maestro (v7.0.6, MIM Software Inc., Cleveland, USA). Functional lung contours FL75, FL60, FL50 and FL25 were segmented by thresholding the ventilation distribution at the 25^th, 40^th, 50^th and 75^th percentiles, respectively, then transferred on the initial planning CT with a rigid registration of the PI image. Two Monaco (v5.51, Elekta AB., Stockholm, Sweden) VMAT plans were optimized for each patient with a 33x2 Gy prescription: one anatomical plan using classical constraints and one functional plan with ALARA constraints on V_20Gy and D_mean for FL60 and FL25. Intra-method distinctness of functional volumes was assessed using a Kruskal-Wallis test. Dose improvements were assessed using a paired Wilcoxon double-tailed signed-rank test. Functional volumes and dose differences were compared with our previous work based on SPECT. A p<0.05 significance level was used for each test.

The Jacobian method provided 4 distinct (p<0.001) volumes (Figure 1) larger than their perfusion SPECT counterparts (p<0.00001).

DVHs showed a reduction in almost all doses for functional lung contours, anatomical lung and the body, with no substantial change in the dose to the target and organs at risk. Results (Figure 2) show a reduction of median V_5Gy, V_20Gy and D_mean in all lung volumes (p<0.01). However, median V_40Gyincreased in FL75, anatomical lungs and body (p<0.01). Small changes were found for the PTV dose coverage, although all constraints were met. No significant changes were observed in doses to the heart, esophagus, and spinal canal. Good agreement was found with our SPECT-based study regarding the impact of functional planning.

The 4DCT-based Jacobian ventilation method was successfully implemented using clinical tools while delivering promising results regarding functional lung avoidance to mitigate RILI. All results were consistent with the literature. Future work should concentrate on improving the implementation of 4DCT ventilation in the RT planning workflow and investigating functional planning at a larger scale. This opens the way for prospective functional lung avoidance studies.