• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.59.2023.tde-03102023-153312
Document
Author
Full name
Rafael Oddone Scatena
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
Ribeirão Preto, 2023
Supervisor
Committee
Murta Junior, Luiz Otavio (President)
Marques, Paulo Mazzoncini de Azevedo
Felipe, Joaquim Cezar
Title in Portuguese
Construção de modelos geométricos quadráticos específicos a partir de imagens de tomografia computadorizada para aplicações em simulações Monte Carlo visando o planejamento radioterápico
Keywords in Portuguese
Imagens tomográficas
Modelos geométricos específicos
Radioterapia
Segmentação automática
Simulação Monte Carlo
Abstract in Portuguese
A simulação de Monte Carlo é o algoritmo computacional de referência para o transporte de radiação no contexto da radioterapia. Embora o algoritmo de Monte Carlo seja muito lento para ser usado no ambiente clínico, nós o aceleramos usando um código híbrido personalizado de CPU e GPU chamado PaRtIcle-matter SiMulATIon Cuda (PRISMMATIC) baseado no PENELOPE. O objetivo principal desta pesquisa foi construir uma geometria quadrática torácica de voxels específica para pacientes com câncer de pulmão construída utilizando segmentação automática a partir de imagens de tomografia computadorizada. Métodos simples de processamento de imagem foram utilizados para segmentação. As regiões delimitadas por voxels foram simplificadas e convertidas automaticamente em geometrias quadráticas compatíveis com os softwares Monte Carlo PENELOPE e PRISMATIC. Foram feitas simulações de tratamentos de câncer de pulmão com o software PENELOPE e pelo PRISMATIC. Foram calculadas as doses integrais nos órgãos de risco e as restrições de doses sobre volume foram computadas pelo software SLIICER-RT. Os resultados mostraram que é possível obter mapas de doses precisamente compatíveis entre CPU e GPU com aceleração de 3,86 x para o cenário de radiocirurgia e 4,11 x no cenário do acelerador True Beam. Os resultados obtidos permitem prever a viabilidade da simulação Monte Carlo no cenário clínico usando modelos geométricos específicos para o paciente em tempo clinicamente aceitável.
Title in English
Construction of specific quadratic geometric models from computed tomography images for applications in Monte Carlo simulations aiming at radiotherapy planning
Keywords in English
Automatic segmentation
Monte Carlo simulation
Radiotherapy
Specific geometric model
Tomographic images
Abstract in English
Monte Carlo simulation is the reference computational algorithm for radiation transport in the radiotherapy context. Although the Monte Carlo algorithm is too slow to be used in the clinical setting, we have accelerated it using a custom CPU-GPU hybrid code called PaRtIcle-matter SiMulATIon Cuda (PRISMMATIC) based on PENELOPE. The main objective of this research was to build a quadratic thoracic geometry specific for patients with lung cancer built using automatic segmentation from computed tomography images. Simple image processing methods were used for segmentation. The regions delimited by threshold voxels of tissues and tumors were simplified and automatically converted into a quadratic phantom geometry compatible with the PENELOPE and PRISMATIC software. Simulations of lung cancer treatments were performed using the PENELOPE software and PRISMATIC. Full doses to organs at risk were calculated and dose restrictions on volume were computed using the SLIICER-RT software. The results confirmed that it is possible to obtain precisely matched dose maps between CPU and GPU with speed up of 3.86x for the radiosurgery scenario and 4.11x for conventional radiotherapy. The obtained results allow us to predict the feasibility of the Monte Carlo simulation in the clinical setting using patient-specific geometric models in clinically acceptable time.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2023-12-21
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.