ESTRO 2022 - Abstract Book

S278

Abstract book

ESTRO 2022

image-registration options and post-processing steps. Results further improved by using AOUC atlas, created and optimized in-house: the fraction of contours scored 0 or 1 reaches 92%, comparable to DL-generated contours. Conclusion DL-based algorithms represent a turning point in the field of auto-contouring and produce high quality contours. Even if some corrections are needed before clinical use, in clinical practice, important time-savings may be obtained, if no or minor corrections are needed. To date, evaluated deep-learning algorithms are capable to produce high quality contours, in most cases clinically acceptable or prone to be quickly edited with minor revision.

[1] Casati et Al. DOI: 10.1002/acm2.13093

PD-0316 Contouring lumbosacral plexus based on MR neurography and MR/CT deformable registration technique

X. Cao 1 , X. Gao 1 , W. Li 2 , P. Liu 3 , S. Qin 1 , H. Li 1 , D. Wang 4

1 Peking University First Hospital, Radiation oncology, Beijing, China; 2 Peking University First Hospital, Radiation department, Beijing, China; 3 Peking University First Hospital, Radiation oncology, Beijing , China; 4 Rush University Medical Center, Radiation oncology, Chicago, USA Purpose or Objective Traditional image techniques such as nonspecialized MR (layer thickness of 5mm and above) do not perform well in assisting nerve contouring on simulation CT, especially the nerves in the lumbosacral plexus, which have complicated anatomy and are greatly affected by spinal curve changes. We investigated the feasibility of using specialized MR neurography (MRN) and MR/CT deformable registration technique to develop a method for contouring the lumbosacral plexus (LSP). Materials and Methods The LSP structure in this study was defined as a combination of the T12 nerve root, L1 nerve root, L2-L4 nerve roots plus femoral nerve (FN), and the L5-S3 nerve roots plus sciatic nerve (SN). Specialized MRN sequences were reviewed, and the NerveVIEW sequence with layer thickness of 2 mm was selected in this study. NerveVIEW sequence were prospectively collected in 18 prostate cancer patients who planned to receive high dose irradiation and one volunteer who completed the carbon ion therapy of her extremity sarcoma in 2017 and subsequently developed radiation-induced lumbosacral plexopathy (RILSP) one year later. Deformable registration was applied to the NerveVIEW sequence and simulation CT images to get fused MR/CT images, on which the LSP was delineated. A dosimetric analysis was done on the enrolled cases with their original treatment plan. Results The LSP structures were successfully created in all enrolled patients. In the patient cohort, the mean volume of LSP, FN and SN were 52.7 ± 6.5 cm 3 (range, 41.44 – 68.2 cm 3 ), 17.5 ± 3.2 cm 3 (range, 13.2 - 23.5 cm 3 ) and 33.4 ± 3.7 cm 3 (range, 27.0 – 42.0 cm 3 ), respectively. The median maximal dose to LSP was 52.72 Gy. Three patients received dose excessing 60 Gy to the LSP and two patients received dose excessing 66 Gy. As for the volunteer, the relative biological equivalent maximal dose to her femoral nerve was 67.5 GyE, with V60 at 50.4% and V66 at 34.3%. This high dose to her femoral nerve corresponded to her femoral nerve injury.