Program Information
Quantification of Dosimetric Error in Real-Time MLC Tracking for Lung SBRT
A Modiri1*, A Sawant1 , (1) University of Maryland School of Medicine, Baltimore, MD
Presentations
MO-RAM-GePD-J(A)-3 (Monday, July 31, 2017) 9:30 AM - 10:00 AM Room: Joint Imaging-Therapy ePoster Lounge - A
Purpose: Real-time dynamic MLC (DMLC) tracking is an attractive motion management strategy for treating moving thoracic and abdominal tumors. An often overlooked limitation of tracking is that the MLC aperture responds in real-time to the positional change of a single point (typically, an external/internal surrogate for the tumor centroid) and does not account for organs-at-risk (OARs) that may come in the beamline during respiration. This study quantifies such dosimetric errors.
Methods: We retrospectively analyzed three lung SBRT patients with tumor motions >10mm: P1: centrally-located, right-lobe, P2: peripheral, left-lower-lobe and P3: peripheral, right-lower-lobe, with left lung excision. For each patient, a 2-minute abdominal belt recording of the respiratory pattern was obtained. Individual-phase treatment plans were created for ten respiratory phases. To simulate single-point DMLC tracking, end-of-exhale was chosen as the reference phase and the PTV was copied onto the other nine phases. These copied PTVs were translated three-dimensionally according to tumor motion at each phase, measured as the 3D shift between the GTV centroids of that phase and the reference phase. MLC apertures were conformed to the 3D-translated PTV. Dose matrices per respiratory phase were exported, weighted by the corresponding respiratory phase probability, and summed using deformable image registration (NiftyReg). The final PTV and OAR doses were compared with those from the reference phase plan.
Results: We evaluated max and mean doses to esophagus, heart and spinal cord. Mean dose estimation error for our peripheral (P2) case was minimal (<0.2Gy). However, the central (P1) and special (P3) cases showed dosimetric errors as large as 1.2Gy (spinal cord Dmax) and 1.8Gy (heart Dmax), respectively.
Conclusion: While DMLC tracking based on single-point real-time monitoring may provide a dosimetric advantage for peripheral tumors, it may be contraindicated in patients with centrally-located tumors exhibiting significant motion due to increased risk to OARs.
Funding Support, Disclosures, and Conflict of Interest: This study was supported by an NIH grant: R01CA169102
Contact Email: