Machine-Specific Delivery Sequence Model of Compact Superconducting Synchrocyclotron Proton Therapy Systems – A Multi-Institutional Investigation

Abstract

<h3>Purpose/Objective(s)</h3> Proton pencil beam scanning delivery sequence varies among institutions due to the vendor, machine type, and configuration. However, having a reliable model of delivery time and sequence is critical to maximizing throughput when proton beam time is limited. In this study, we used an experimental approach to build a precise beam delivery time (BDT) and sequence prediction model for a new compact superconducting synchrocyclotron proton system (PTC#1) and compare it with a different institution using an older version of the same system (PTC#2). We investigated treatment efficiency and the interplay effect for the treatment of mobile tumors. <h3>Materials/Methods</h3> The machine delivery log files from PTC#1 were retrospectively analyzed to model beam delivery parameters. We compared the PTC#1 delivery sequence with PTC#2's published machine-specific model using the same type of proton system. A total of twenty clinical beams from head neck, lung, breast, and esophagus treatment sites were used for model comparison. Also, we used a single clinical treatment day to compare treatment efficiency. The interplay effect is evaluated and compared based on these two different delivery sequence models. A digital thoracic 4DCT phantom image set with a rigid target in the right lung was used for the study. Different delivery techniques such as standard delivery (N_vol=1) and volumetric repainting delivery (N_vol=2-5) were simulated. <h3>Results</h3> The results showed PTC#1 has a much faster spot scanning speed than PTC#2 thanks to a new generation of scanning magnets. The spot switching time (SSWT) is 0s within 11.3mm in x-direction and 8 mm in y-direction. Compared to PTC#2, the average SSWT reduces by 94.37%±3.82% (50.76s±35.10s) in PTC#1. PTC#1 also demonstrated improved burst switching transmission efficiency, which reduced burst switching time (BST) by 39.84%±3.64% (6.00s±2.21s) on average. The energy layer switching time (ELST) dominates the PTC#1 delivery time instead of SSWT for PTC#2 (Table). The total BDT was reduced by 54.61%±12.12% (53.45s±34.47s) on average. Based on the single day patient mix, it was estimated that 67 min irradiation time could be saved by upgrading the scanning magnets in PTC#2 (10% daily patient treatment throughput improvement). The results also indicated the interplay effect would be reduced as volumetric repainting number increased. The system in PTC#2 had better target coverage and homogeneity than PTC#1 for=1-3. Their target coverage converges at=5. <h3>Conclusion</h3> A precise machine-specific delivery sequence model is highly recommended in the interplay effect evaluation and clinical decision. It provided the first quantitative analysis of the impact of the new scanning magnets regarding the improvement in the treatment delivery efficiency.