Combining Virtual Bronchoscopy with 4DCT-Based Ventilation Mapping to Preserve Serial- and Parallel-Functioning Lung Elements in Lung Stereotactic Ablative Radiotherapy (SABR)

Lung functional-avoidance radiotherapy (LFRT) aims to mitigate post-RT loss of respiratory function by spatially mapping regional lung function and preferentially avoiding dose to higher-functioning regions. A common limitation of current LFRT approaches is that they ignore peripheral airways (generation 3+), as these are not spatially resolved in most planning CTs and can end up in high-dose regions. Damage to these branching serial structures can result in significant functional loss downstream, thus negating the original intent of LFRT. Here, we develop a novel LFRT framework which combines virtual bronchoscopy (VB)-derived airway maps with 4DCT-derived ventilation maps to preserve both serial and parallel functioning structures. Under IRB approval, a breath-hold CT (BHCT) was acquired from a lung SABR patient and 238 airways were auto-segmented using a research VB system (11 generations; 18-3 mm diameter, 119 terminal airways). Concurrently, a 4DCT-based ventilation map was generated using the highest-scoring of 37 independent techniques from the recent multi-institutional VAMPIRE study; a biomechanical model-based deformable image registration between peak inhale/exhale phases to estimate the initial map, followed by corrections for tissue shrinkage/expansion. The map was deformably registered to the BHCT and the sub-lobar lung volume supported by each terminal airway was estimated. A ventilation score (summed voxel intensities within the volume) was calculated. Upstream airways were scored using the cumulative ventilation of connected downstream airways. Using previously estimated airway radiosensitivities, dose constraints were determined for each airway corresponding to a 95% probability of no damage i.e., airway remains open. Airway dose constraints, sub-volume ventilation scores, and clinical dose constraints (PTV, OARs) were input to a swarm optimization engine to create a conformal SABbR plan. The proposed plan was compared to the clinical plan in terms of total airflow preserved, i.e., number of individual open pathways, and total ventilation preserved, i.e., cumulative ventilation score from the sub-lobar volumes of open pathways. For the proposed plan, reduction in dose and in airway collapse probability were both highly correlated with airway function-based weighting factors (f in table); with p < 0.01 for Spearman as well as two-tailed tests. Our proposed plan preserved 86% of the original airflow and 83% of the original ventilation compared to 66% airflow and 63% ventilation for the clinical plan. These initial results suggest that it is important to consider the serial as well as the parallel nature of the lung. LFRT studies, including ongoing clinical trials, that ignore airway damage may result in increased post-RT functional loss.Abstract 54; Table 1Proposed planClinical planDose reduction (Gy) [min/max/avg]f ≥0.5-0.04/23/7---f <0.5-8/26/3---Airflow at terminals (%)8666Ventilation preservation (%)8363 Open table in a new tab