3127 - Adapting for Accuracy Using MRI-guided Adaptive Radiation Therapy for Single-Isocenter, Multi-Target Liver Treatment
Presenter(s)
A. Shepard1, I. Cook2, M. F. Bassetti1, J. Crosby1, M. M. Basree3, G. C. Blitzer1, A. R. Burr1, and C. Glide-Hurst1; 1Department of Human Oncology, University of Wisconsin-Madison, Madison, WI, 2Department of Medical Physics, University of Wisconsin-Madison, Madison, WI, 3Department of Human Oncology, University of Wisconsin Hospitals and Clinics, Madison, WI
Purpose/Objective(s): Single-isocenter, multi-target (SIMT) liver treatments can improve delivery efficiency and simplify assessment of composite normal-tissue dose compared to multiple independent plans. However, SIMT in the liver is highly sensitive to geometric uncertainties due to inter-fraction variations in respiration and deformation, leading to displacement between targets and increased sensitivity to rotational setup errors that may compromise tumor coverage. MRI-guided adaptive radiotherapy (MRgART) is uniquely positioned to mitigate these issues. We describe our experience and quantify inter-fraction variability and target changes in this complex setting.
Materials/Methods: Sixteen patients with 4-11 lesions (93 total) were treated over 5 fractions with a single isocenter using an MRgART workflow. The need for target modification was assessed by tracking the physician-contoured GTV volumes for each fraction compared to the volume in the base plan for that fraction. Inter-fractional positional changes within the liver were assessed by determining the change in lesion-to-lesion distance for each unique lesion pair in the treatment course (256 lesion pairs for 5 fractions each). To isolate the impact of deformation in the liver and remove the impact of daily contouring variability or tumor volume changes, the initial reference plan target contours were propagated to each fraction’s image based on per-lesion registrations. The change in the centroids for each lesion-pair relative to the distance in the initial planning scan was quantified.
Results: Target volumes were modified during daily adaptation for 81% (375/465) of all lesions, with at least one lesion being modified in 85% (68/80) of fractions. The average target volume change relative to the volume in each fraction’s base plan was 1.1 cm3 (range: -5.2 to +22.2 cm3). Of the lesions modified, 72% (269/375) had an increased volume relative to the volume in the base plan for each fraction. Overall, the treatment plan was adapted in 86% (69/80) of all fractions investigated. The mean absolute change in centroid difference for all propagated lesion pairs, as well as the percentage of lesion-pairs with a change of less than 5 and 10 mm is presented in Table 1. All lesion pairs within a given fraction had differences of less than 5 mm and 10 mm in 10% and 54% of fractions, respectively.
Conclusion: Our findings demonstrate substantial inter-fraction variability in target volumes and relative lesion positions, underscoring the geometric complexity of SIMT liver treatments. MRgART enables the flexibility to maintain precise treatment without the need for increased margins for highly variable multi-target geometries.
Table1. Lesion-Pair Centroid Distance Changes.| Mean [mm] | Max [mm] | < 5 mm | All Directions < 5 mm | < 10 mm | All Directions < 10 mm | |
| Left/Right | 3.9 | 31.6 | 73% | 47% | 93% | 88% |
| Ant/Post | 3.0 | 21.9 | 82% | 98% | ||
| Sup/Inf | 3.8 | 21.0 | 68% | 95% |