Treating moving targets with scanned proton therapy: Is 5 mm initial tumour motion a safe threshold?

Purpose/Objective Pencil beam scanning (PBS) is a highly conformal technology to treat cancer. The time structure of PBS makes the treatment of moving tumours challenging due to the interplay effect. According to literature, PBS in combination with rescanning can be safely applied without motion mitigation strategies to lung tumours that move 5 mm or less. However, the question is whether the motion measured during treatment simulation will remain below 5 mm during the course of treatment. We investigated the inter-fractional lung tumour motion variation in a unique data set providing five repeated 4DCTs per patient, to evaluate if a 5 mm threshold is a reliable indicator for considering PBS treatments. Material/Methods For 19 NSCLC patients (11 male, 8 female, age: 47-89, stage: II-IV) weekly 4DCT imaging was performed during treatment simulation before and during the treatment course to monitor the anatomical changes and differences in motion. Gross tumour volumes (GTV) were delineated on the maximum inspiration and expiration phases of the planning 4DCT and on the weekly repeat 4DCTs. GTV volume changes were acquired and the weekly inter-fraction motion variation was evaluated by measuring the GTV centroid translations in all three directions. Results The patients showed a median initial tumour motion of 1.3 mm (range: 0.0 – 5.1 mm) for a median initial GTV volume of 28.7 cm3 (range: 1.9 – 430.0 cm3). Figure 1 shows the measured 3D-vector motion for week 0 (before start treatment) and week 1-5 (treatment course). The maximum deviation from the initial measured motion for the patients was on average 2.1 mm (range: 0.3 – 7.9 mm). Centroid displacements remained under 5 mm for 16 out of 19 patients over the entire course of treatment. For patients number 3, 4 and 11, an increase in motion up to a maximum of respectively 9.1, 11.2, and 6.5 mm was observed (Figure 2). GTV volumes for 15 out of 19 patients shrank during treatment with a median decrease of 35.4% (range: 10.8% - 63.7%), and a median absolute volume change of 8.3 cm3 (range: 0.5 - 105.9 cm3). Conclusion Even if the lung tumour motion during planning simulation is under 5 mm, variations in motion amplitude and larger motions can still occur during the course of treatment. This indicates that the use of daily tumour motion evaluation is required for PBS treatments, especially for hypo-fractionated treatments. Future work will include a risk-stratification for motion variations in lung patients based on initial tumour motion, tumour position and tumour volume for different PBS spot sizes, time structure and prescribed treatment courses