ESTRO 2023 - Abstract Book

S2110

Digital Posters

ESTRO 2023

Conclusion: Despite high cumulative doses. re-irradiation at curative intend with PT or IMRT seems safe and efficient for local control for patients with recurrent HNACC.

PO-2345 Evaluation of manual and DirectOrgans algorithm for delineation of OAR in thorax radiation therapy

M. Sand 1 , L. Ring 1 , P. Qvistgaard 1 , J. Hansen 1

1 Aarhus University Hospital, Department of Oncology, Aarhus, Denmark

Purpose or Objective Delineation of target and organ at Risk (OAR) is one of the most important, time consuming and operator-dependent tasks in radiation therapy. There are different systems/algorithms for automatic OAR contouring. One of them is Siemens DirectOrgans algorithm (Optimized Reconstruction based Generative Adversarial Networks). This study aims to: a) Evaluate manual and DirectOrgans algorithm for delineation of OARs in thorax. b) Form the basis for future work to implement automatic DirectOrgan contouring in our clinic. Materials and Methods Manual and DirectOrgans-segmentation of the lungs, heart and esophagus of 16 patients simulated on Siemens Go Open Pro CT were made. Difference between manually delineated volume and automatic volume ( ∆ V, %), dice similarity coefficient (DSC), sensitivity (Se) and specificity (Incl) were calculated to compare the accuracy of these two methods. The contours were also compared visually slice by slice. Results The average difference between the lungs, heart and esophagus volume delineated manually and with DirectOrgan algorithm were: -2.65% ± 3.58 ( ∆ V ± SD), 9.93 ±11.43 and 0.86 ± 13.27. For esophagus we found ∆ V from about -21.6% to 21.04%. The ∆ V values for esophagus underline the difficulty in the contouring. Regarding the DSC index, the average and SD for delineated organs were: 0.96 ± 0.02 for lungs, 0.92 ± 0.03 and 0.72 ± 0.09 for heart and esophagus respectively. DSC index value approach nearly 1 for lungs and heart, it means that manually and DirectOrgans outcomes are significantly similar. For esophagus, differ outcomes significantly. The Se and Incl for lungs were nearly the same: 0.94 ± 0.03 and 0.97±0.01 (Se ± SD and Incl ± SD). For the heart the difference between Se and Incl were larger and were: 0.96 ± 0.02 and 0.88±0.07 respectively. For esophagus both Se and Incl were the same but low: 0.73 ± 0.10. Conclusion Result of this study demonstrate that DirectOrgan algorithm can be used for the delineation of OARs in the thorax region. The greatest difference between volumes contoured using the two methods is seen in those patients who had an advanced disease in these organs of interest. This is a pilot study and future work with more patient data is needed to test the contouring of other organs from DirectOrgan's catalog before implementing in clinical use. 1 The Hong Kong Polytechnic University, Department of Health Technology and Informatics, Hong Kong, China; 2 Hong Kong Sanatorium & Hospital, Department of Radiotherapy, Hong Kong, China; 3 The Chinese University of Hong Kong, Department of Otorhinolaryngology, Head and Neck Surgery, Hong Kong, China Purpose or Objective External auditory canal (EAC) squamous cell carcinoma (SCC) is a rare tumour, accounting for less than 0.2% of all head and neck (H&N) cancers. Nasopharyngeal carcinoma (NPC) survivors who had previous radiotherapy (RT) have a 1000-fold risk of developing second primary EACSCC. This could be attributed to previous high dose irradiation to the H&N region. In this study, we first investigated the radiation doses to the EACs in those NPC survivors who eventually developed EACSCC. With the recent advancement of RT techniques, our second objective was to compare the doses to the EACs between conventional 2-dimensional RT (2DRT) and advanced intensity (IMRT)/ volumetric-modulated (VMRT) techniques in NPC RT. Materials and Methods For part one, five cases with EACSCC who previously underwent 2DRT for NPC (prescribed dose >60 Gy) were retrospectively reviewed using planning images to estimate the radiation doses to the EACs. The mean and maximum dose to the EACs were compared between the affected and unaffected sides. In part two, twenty IMRT/VMAT and nineteen 2DRT plans treated between 1988-2020 were randomly extracted from a tertiary hospital for estimating the dose to the EACs. Similar to part one, the dose range (difference between the maximum and minimum dose points within the EAC), mean and maximum dose to the EACs between the two techniques were compared. Results For the NPC survivors with EACSCC treated with 2DRT, there were no statistical differences in the mean and maximum doses of EACs between the affected and unaffected sides (Table 1). Comparing between 2DRT and IMRT/VMAT techniques, the mean dose to the EACs using IMRT/VMAT (32.94Gy) was drastically lower than using 2DRT (62.79Gy), which would minimise acute ototoxicity. The dose range in the EAC region using IMRT/VMAT (64.57Gy) was significantly larger than 2DRT (13.53Gy) (p< .001), indicating that the EAC immersed in a relatively lower and more heterogeneous radiation bath, compared with 2DRT (Table 2). This may paradoxically increase the risk of developing second primary cancers from PO-2346 Dosimetric study of radiation dose to external auditory canal among nasopharyngeal cancer survivors S.Y.S. Ho 1,2 , W.L. Yeung 1 , C.H. Mak 1 , H.T. Ng 1 , K. Tsang 1 , W.S.V. Leung 1 , C.P.L. Chan 3 , S.W.Y. Lee 1