TH‐C‐BRD‐09: Successes and Limitations of Online Range Adaptive Spot Scanning Proton Therapy for NSCLC

• Published in: Medical physics (Lancaster) 2014-06, Vol.41 (6Part32), p.552-552
• Main Authors: Cheung, JP, Dong, L, Park, P, Zhu, XR, Kudchadker, RJ, Frank, SJ, Court, LE
• Format: Article
• Language: English
• Subjects: 60 APPLIED LIFE SCIENCES; Adaptive methods; ALGORITHMS; ANATOMY; Cancer; CAT SCANNING; Computed tomography; DEPTH DOSE DISTRIBUTIONS; LUNGS; NEOPLASMS; PATIENTS; PHANTOMS; PLANNING; PROTON BEAMS; Proton therapy; Protons; RADIATION DOSES; RADIOTHERAPY; SPINAL CORD; Tissues
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1118/1.4889607

Purpose: To determine the ability to adapt discrete spot‐scanning proton therapy (SSPT) plans based on geometric changes of anatomy to minimize normal tissue dose and maintain target coverage. Methods: We developed and tested a range‐correction algorithm to compensate for anatomy changes in SSPT with correction factors for target lateral size changes and energy scaling. This algorithm adjusts the energy of each spot from the original optimized treatment plan to match the new daily anatomy based on water equivalent path‐length. To correct for the lateral target size changes, the peripheral spots were scaled based on phantom studies with variable target size. For energy change corrections, alternative cumulative scaling factor lookup tables were generated based on calculated central‐axis and integral depth dose calculations for different energies. These various adaptive algorithms were performed on 7 lung cancer patients that were previously treated with proton therapy and who required at least one adaptive intervention. Single‐field optimized SSPT plans were generated for these patients with clinical beam angles. Dose‐volume histogram metrics were obtained for these patients for both the non‐adaptive and the different adaptive plans applied to the last available weekly CT scan. Results: The doses to normal tissue were largely reduced for the spinal cord (Dmax), total lung (V20Gy), and contralateral lung (V20Gy) for all different methods of adaptive planning. With both corrections applied, the average changes for these metrics were −6.2Gy, −2.7%, and −4.9%, respectively. The same method generated unacceptably high target hot spots with average target V110% increase of 12.3%. Conclusion: Adaptive methods based on direct adjustments to proton range can reduce normal tissue doses under large anatomical changes but are insufficient in achieving clinically acceptable target doses and generate unacceptably sizeable hot spots. Adaptive planning methods for proton therapy will likely necessitate some aspect of reoptimization for acceptable clinical utility.