ESTRO 2020 Abstract book

S1015 ESTRO 2020

slightly compromise through-plane resolution. Three healthy volunteers were then recruited for sequence comparison and registration consistency test. They were set up in the treatment position immobilized using a vacuum cushion. Each volunteer was repositioned for each scan (including both default and optimized sequences) and totally received 4 scans. Images acquired at the first scan were used as a reference. The rest scans were rigidly registered to the reference using on-line MONACO (v5.4). Translations in LR, AP and SI were calculated and compared between the two sequences using t-test. Image signal to noise ratio (SNR) was measured in the bladder and iliac bone regions and compared using paired t-test. Two medical physicists blindly evaluated image quality on a 5- point scale (1 poor, 2 fair, 3 moderate, 4 good, 5 excellent), based on SNR, image contrast, sharpness, tissue differentiation, and artifacts. Wilcoxon test and intra-class correlation coefficient (ICC) was used to assess the rating result with a significant level of p=0.05.

quality included relatively low spatial resolution and limited image contrast due to the extremely short TE/TR. The translational displacements of liver dome and kidneys (to the first frame) with respiration in a subject were also illustrated (Fig. 2). Future work of reconstruction latency reduction, motion prediction and ghosting artifact removal using deep learning is on-going.

Conclusion CAIPRINHA-TWIST-VIBE was developed and achieved 3fps frame rate to capture volumetric motion information in the abdomen. It has potentials for motion monitoring, gating and tracking of future MRgRT. PO-1737 Pulse sequence optimization in pelvis on a 1.5T MR-Linac Y. Zhou 1 , J. Yuan 1 , Y.W. Ho 1 , B. Yang 1 , L.C. Ho 1 , K.Y. Cheung 1 , S.K. Yu 1 1 Hong Kong Sanatorium & Hospital, Medical Physics and Research Department, Hong Kong, Hong Kong SAR China Purpose or Objective To optimize the MR pulse sequence for pelvis on a 1.5T MR- Linac to shorten the acquisition time while preserving image quality. Material and Methods The default MR-LINAC pelvis T2 TSE sequence (TR/TE = 1535/278ms, acquired voxel size = 1.5 x 1.5 x 2 mm 3 , SENSE factor=3.6(LR) x 1(SI), half scan = 0.62(PE) x 1(SE), TSE factor=114, RBW = 740.3Hz/pixel, scan time = 117s) was modified to achieve the aim of shorter scan time and comparable or better image quality. After a series of phantom and volunteer imaging, the optimized sequence was finally determined (TR/TE = 1380/267ms, acquired voxel size = 1.5 x 1.5 x 2 mm 3 , SENSE factor 3.5(LR) x 1.1(SI), half scan = 0.6(PE) x 0.8(SE), TSE factor = 95, receiver bandwidth RBW = 496.2Hz/pixel, scan time = 95s). TR was slightly reduced. A smaller TSE factor helped to shorten echo train duration and thus reduce image blurring. Smaller receiver bandwidth improved SNR. The geometric distortion increase was too minor to be measurable as indicated by phantom imaging. SENSE factor was slightly increased in SI without extra observable artifacts. Half scan factor in SI was reduced, which might

Results The acquisition parameters of the default and the optimized sequences were presented in Table 1. The scan time of the optimized sequence was ~19% shorter than the default sequence. The mean registration results were insignificant between the two sequences. The registration differences were -0.07±0.03, -0.04±0.01 and 0.01±0.01 mm in LR (p=0.126), AP (p=0.278) and SI (p=0.588) (Table. 2). SNR in the bladder and iliac bone regions using the optimized sequence were 25.94±0.68 dB and 16.36±0.86 dB, ~4.6% and ~14.2% significantly higher (p<0.01) than that using default sequence (24.79±1.05 dB and 14.33±1.72 dB), respectively. Observers showed excellent agreements (ICC = 0.84) on favoring the optimized images over the default ones (p<0.0001). Conclusion The optimized sequence achieved shorter acquisition time and superior image quality than the default sequence. It might be useful in the future pelvis MRgRT on MR-LINAC. PO-1738 Dosimetric evaluation of deep learning generated synthetic CT data for H&N MRI only radiotherapy E. Palmér 1 , A. Karlsson 1 , F. Nordström 1 , C. Siversson 2 , K. Petruson 3 , M. Ljungberg 1 , M. Sohlin 1 1 University of Gothenburg, Department of Radiation Physics- Institute of Clinical Sciences- Sahlgrenska Academy, Gothenburg, Sweden ; 3 University of Gothenburg, Department of Oncology and Radiotherapy- Institute of Clinical Sciences- Sahlgrenska Academy, Gothenburg, Sweden