Abstract Book

S1170

ESTRO 37

Purpose or Objective 3D volumetric MRI in abdomen for radiotherapy (RT) planning is challenged by the pronounced respiratory motion that compromises image quality and thus tissue delineation. Stereotactic body radiotherapy (SBRT) via abdominal compression (AC) could improve treatment delivery accuracy by restricting the respiratory motion. In this study, we aimed to evaluate whether AC could improve volumetric 3D MRI image quality and reduce diaphragm motion for SBRT. Material and Methods 8 healthy volunteers were recruited. Free breathing (FB) and 3 levels of AC (strong, medium and low) were applied using an adjustable screw. For each setting, a single slice sagittal cine image (TRUFI, TR/TE=355.7/1.6ms, voxel size=2.09x2.09x4mm 3 , scan time=16s) of medial right lung and T1w volumetric MR images (3D VIBE, TR/TE=6.68/2.39ms, voxel size=1.6x1.6x1.6mm3) covering thoracic and upper abdominal area were acquired 3 times on a 1.5T MR-simulator. The maximum positional shift of the dome of the diaphragm along SI and AP direction at each AC level was evaluated on the cine images using a custom matlab program based on cross-correlation . Organ edge sharpness, aliasing artifact and motion artifact were individually rated using a 5- point scale (point 1-5:worst-best). The image quality score at each AC level was compared using rank-sum test. Results The lowest image quality scores were associated with FB when compared with other levels of AC (organ edge sharpness:2.24±0.93 (FB), 2.30±0.85 (low), 2.50±0.81 (medium), 2.51±0.77 (strong); aliasing artifact:2.93±1.03 (FB), 2.94±0.90 (low), 3.18±0.88 (medium), 3.11±0.94 (strong); motion artifact:2.40±0.96 (FB), 2.56±0.85 (low), 2.52±0.80 (medium), 2.73±0.87 (strong)). Significantly lower score was only observed on motion artifact under FB (2.40±0.96 ) when compared with strong AC (2.73±0.87, p< 0.05). In terms of maximum positional shift, significantly smaller shift along anterior-posterior (AP) direction were observed under strong AC (1.25±0.99mm) when compared to FB (2.79±0.99mm, p<0.01) and other AC (low:2.56±1.43mm, p<0.01; medium: 2.04±0.06mm, p<0.01). When compared to medium (9.90±2.40mm, p<0.01) and strong (9.63±2.76mm, p<0.01) AC levels, significantly larger shift along superior-anterior(SI) direction were observed under FB (13.24±3.49mm, strong: p<0.01; medium: p<0.01) and low AC (13.21±4.70mm, strong: p<0.01; medium: p<0.05) Conclusion The abdominal image quality was generally improved under SBRT compression. In particular, motion artifact was substantially reduced under strong compression compared to FB. Respiratory motion induced positional shift was also significantly reduced under strong and medium AC compared to FB. EP-2126 4DMRI amplitude binning: better estimation of reconstructed motion at no cost in 4DMRI quality Z. Van Kesteren 1 , A. Van der Horst 1 , N. Onkenhout 2 , A. Bel 1 1 Academic Medical Center, Department of Radiotherapy, Amsterdam, The Netherlands 2 Vrije Universiteit, Faculty of Exact Sciences- Medical Natural Sciences, Amsterdam, The Netherlands Purpose or Objective 4DMRI is gaining popularity in radiotherapy of the upper abdomen, as it provides good soft tissue contrast and depicts the target motion during free breathing. The choice of applied binning strategy may influence imaging quality and potentially underestimate motion amplitude. Here, we aimed to validate and compare 2 amplitude binning strategies, assessing the accuracy of reconstructed motion amplitude and 4DMRI quality.

Material and Methods For each 4DMRI scan, 11 2D coronal slices were acquired repetitively (60 times, 6 minutes total), using a T2W TSE sequence (resolution: 1.3x1.6x5.0 mm 3 ). Prior to each slice acquisition, the position of the diaphragm was acquired using a 1D navigator. The 2D slices were sorted into 10 amplitude bins. For each bin/slice combination, the image with the median diaphragm position was selec- ted for 4DMRI reconstruction, introducing an under- estimation in motion amplitude depending on bin size. We applied 2 amplitude binning strategies: (A) half a bin size for the end-inhale and end-exhale bins and (B) equal- sized amplitude bins (Figure 1), which is current clinical practice. Amplitude determination was assessed with dynamic phantom measurements, quantifying the reconstructed motion amplitude of a water vial moving with sin and cos 6 respiratory patterns with amplitudes of 1.5, 2 and 2.5 cm.

4DMRI quality was evaluated by in vivo measurements with 12 healthy volunteers and 2 abdominal cancer patients. To account for outliers, the images associated with the 5% most extreme diaphragm positions were discarded prior to binning. Binning strategies were compared based on the parameters: • * Reconstruction completeness (RC); fraction of the 110 (11 slices x 10 bins) bin/slice combinations with at least 1 image.

•

* Intra-bin variation (IBV); interquartile range of the diaphragm positions within a bin/slice combination, averaged over 3 central slices. * Image smoothness (S); assessed by quantifying how well a parabola fits the diaphragm shape in a sagittal plane of the reconstructed 4DMRI, per bin (S = R 2 adj averaged over all bins). S ranges from 0 (discontinuous diaphragm shape; artefacts) to 1 (smooth shape; no artefacts).

•

A low IBV indicates high precision of reconstructed diaphragm position. Low RC and high IBV result in image artefacts, e.g. discontinuities between reconstructed slices, quantified by a low S. We tested for differences in