ESTRO 2022 - Abstract Book

S1436

Abstract book

ESTRO 2022

Conclusion 3D printing applications in radiotherapy or by specialized companies will continue to increase providing an essential tool for affordable innovation and customization (e.g. patient or application-specific phantoms). Nonetheless, in-house developed prototypes need medical certification through regulations, risk analysis, documentation and other requirements for implementation for which hospitals may not be prepared.

PO-1641 Virtual brain MRI missing sequences creation using machine learning generative adversarial networks

E. Giacomello 1 , D. Dei 2,3 , N. Lambri 2 , P. Mancosu 4 , D. Loiacono 1

1 Politecnico di Milano, Dipartimento di Elettronica, Informazione e Bioingegneria, Milano, Italy; 2 Humanitas University, Department of Biomedical Sciences, Pieve Emanuele, Milano, Italy; 3 IRCCS Humanitas Research Hospital, Department of Radiotherapy and Radiosurgery, Rozzano-Milano, Italy; 4 IRCCS Humanitas Research Hospital, Medical Physics Unit, Radiotherapy and Radiosurgery Department, Rozzano-Milano, Italy Purpose or Objective Deep learning (DL) has been successfully applied to segment regions of interest (ROI) from magnetic resonance imaging (MRI). These segmentation models are often designed to work with multi-modal images as input to exploit the most from the different characteristics among the various sequences and generate better segmentations. However, when applying a segmentation model to a new patient, all the sequences required by the model might not be available. This can happen,