Physica Medica 80 (2020) 308–316 1120-1797/© 2020 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved. Original paper Evaluation of a cycle-generative adversarial network-based cone-beam CT to synthetic CT conversion algorithm for adaptive radiation therapy Miriam Eckl a, 1 , Lea Hoppen a, *, 1 , Gustavo R. Sarria b , Judit Boda-Heggemann a , Anna Simeonova-Chergou a , Volker Steil a , Frank A. Giordano b , Jens Fleckenstein a a Department of Radiation Oncology, University Medical Center Mannheim, University of Heidelberg, Germany b Department of Radiology and Radiation Oncology, University Hospital Bonn, Germany A R T I C L E INFO Keywords: Synthetic CT Cone-beam CT Cycle-generative adversarial network-based image correction Adaptive radiation therapy ABSTRACT Purpose: Image-guided radiation therapy could benefit from implementing adaptive radiation therapy (ART) techniques. A cycle-generative adversarial network (cycle-GAN)-based cone-beam computed tomography (CBCT)-to-synthetic CT (sCT) conversion algorithm was evaluated regarding image quality, image segmentation and dosimetric accuracy for head and neck (H&N), thoracic and pelvic body regions. Methods: Using a cycle-GAN, three body site-specific models were priorly trained with independent paired CT and CBCT datasets of a kV imaging system (XVI, Elekta). sCT were generated based on first-fraction CBCT for 15 patients of each body region. Mean errors (ME) and mean absolute errors (MAE) were analyzed for the sCT. On the sCT, manually delineated structures were compared to deformed structures from the planning CT (pCT) and evaluated with standard segmentation metrics. Treatment plans were recalculated on sCT. A comparison of clinically relevant dose-volume parameters (D 98 , D 50 and D 2 of the target volume) and 3D-gamma (3%/3mm) analysis were performed. Results: The mean ME and MAE were 1.4, 29.6, 5.4 Hounsfield units (HU) and 77.2, 94.2, 41.8 HU for H&N, thoracic and pelvic region, respectively. Dice similarity coefficients varied between 66.7 ± 8.3% (seminal ves- icles) and 94.9 ± 2.0% (lungs). Maximum mean surface distances were 6.3 mm (heart), followed by 3.5 mm (brainstem). The mean dosimetric differences of the target volumes did not exceed 1.7%. Mean 3D gamma pass rates greater than 97.8% were achieved in all cases. Conclusions: The presented method generates sCT images with a quality close to pCT and yielded clinically acceptable dosimetric deviations. Thus, an important prerequisite towards clinical implementation of CBCT- based ART is fulfilled. 1. Introduction Adaptive radiation therapy (ART) is one of the latest improvements in radiation therapy treatment quality. Currently, the conventional linac-based radiation therapy workflow neglects morphological changes during fractionated treatment courses. By applying ART techniques, daily imaging could not only be used for translational and rotational patient positioning but also for modifying the initial treatment plan to compensate for a deformed patient anatomy [1–4]. Interfractional treatment adaptations following CBCT based image guidance are com- mon in the head and neck (H&N), thoracic and pelvic region [5] due to random organ motion or deformations (weight loss, tumor shrinkage, etc.) [6]. For conventional linear accelerator (linac)-based radiation therapy a major challenge on the way towards CBCT-based ART is insufficient cone-beam computed tomography (CBCT) image quality [7–9]. A CBCT possesses image artifacts due to detector scatter, patient specific scatter, image lag and beam hardening, making dose calculations prone to er- rors. Many calibration approaches for accurate dose calculations on CBCT currently exist [10]: Patient- or population-specific CT number to electron density calibration (CT-ED-calibration) [11], bulk density override [12], image processing algorithms that further improve the * Corresponding author at: Department of Radiation Oncology, University Medical Center Mannheim, University of Heidelberg, Theodor-Kutzer Ufer 1-3, 68167 Mannheim, Germany. E-mail address: Lea.Hoppen@umm.de (L. Hoppen). 1 The first and second author contributed equally to this work. Contents lists available at ScienceDirect Physica Medica journal homepage: www.elsevier.com/locate/ejmp https://doi.org/10.1016/j.ejmp.2020.11.007 Received 10 July 2020; Received in revised form 29 October 2020; Accepted 5 November 2020