Deformable image registration for dose mapping between external beam radiotherapy and brachytherapy images of cervical cancer.

Phys Med Biol

Univ Rennes, CLCC Eugène Marquis, Inserm, LTSI-UMR 1099, F-35000 Rennes, France. Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, United States of America. Author to whom any correspondence should be addressed.

Published: May 2019

For locally advanced cervical cancer (LACC), anatomy correspondence with and without BT applicator needs to be quantified to merge the delivered doses of external beam radiation therapy (EBRT) and brachytherapy (BT). This study proposed and evaluated different deformable image registration (DIR) methods for this application. Twenty patients who underwent EBRT and BT for LACC were retrospectively analyzed. Each patient had a pre-BT CT at EBRT boost (without applicator) and a CT and MRI at BT (with applicator). The evaluated DIR methods were the diffeomorphic Demons, commercial intensity and hybrid methods, and three different biomechanical models. The biomechanical models considered different boundary conditions (BCs). The impact of the BT devices insertion on the anatomy was quantified. DIR method performances were quantified using geometric criteria between the original and deformed contours. The BT dose was deformed toward the pre-CT BT by each DIR method. The impact of boundary conditions to drive the biomechanical model was evaluated based on the deformation vector field and dose differences. The GEC-ESTRO guideline dose indices were reported. Large organ displacements, deformations, and volume variations were observed between the pre-BT and BT anatomies. Rigid registration and intensity-based DIR resulted in poor geometric accuracy with mean Dice similarity coefficient (DSC) inferior to 0.57, 0.63, 0.42, 0.32, and 0.43 for the rectum, bladder, vagina, cervix and uterus, respectively. Biomechanical models provided a mean DSC of 0.96 for all the organs. By considering the cervix-uterus as one single structure, biomechanical models provided a mean DSC of 0.88 and 0.94 for the cervix and uterus, respectively. The deformed doses were represented for each DIR method. Caution should be used when performing DIR for this application as standard techniques may have unacceptable results. The biomechanical model with the cervix-uterus as one structure provided the most realistic deformations to propagate the BT dose toward the EBRT boost anatomy.

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ab1378DOI Listing

Publication Analysis

Top Keywords

biomechanical models
16
dir method
12
deformable image
8
image registration
8
external beam
8
cervical cancer
8
dir methods
8
ebrt boost
8
boundary conditions
8
biomechanical model
8

Similar Publications

State-of-the-art controllers for active back exosuits rely on body kinematics and state machines. These controllers do not continuously target the lumbosacral compression forces or adapt to unknown external loads. The use of additional contact or load detection could make such controllers more adaptive; however, it can be impractical for daily use.

View Article and Find Full Text PDF

Predictive control of musculotendon loads across fast and slow-twitch muscles in a simulated system with parallel actuation.

Wearable Technol

February 2025

Neuromuscular Robotics Laboratory, Department of Biomechanical Engineering, University of Twente, Enschede, the Netherlands.

Research in lower limb wearable robotic control has largely focused on reducing the metabolic cost of walking or compensating for a portion of the biological joint torque, for example, by applying support proportional to estimated biological joint torques. However, due to different musculotendon unit (MTU) contractile speed properties, less attention has been given to the development of wearable robotic controllers that can steer MTU dynamics directly. Therefore, closed-loop control of MTU dynamics needs to be robust across fiber phenotypes, that is ranging from slow type I to fast type IIx in humans.

View Article and Find Full Text PDF

Purpose: This study aims to investigate the stress distribution in bone tissue, implant, abutment, screw, and bridge restoration when the mesial implant is placed axially and the distal implant is inserted at varying angles in the posterior maxillary region with free-end partial dentition defects, using three-dimensional finite element analysis.

Materials And Methods: Cone-beam computed-tomography were utilized to create 3D reconstruction models of the maxilla. Stereolithography data of dental implants and accessories were used to design a three-unit full zirconia bridge for the maxillary model.

View Article and Find Full Text PDF

Soft micropillar arrays enable detailed studies of cellular mechanotransduction and biomechanics using traditional beam-bending models. However, they often rely on simplified assumptions, leading to significant errors in force estimation. We present MechanoBioCAD (MBC), a finite element method (FEM)-based tool designed specifically for micropillar research and error estimation.

View Article and Find Full Text PDF

Background: Grasping and manipulating objects requires humans to adapt both grip and manipulation forces. When handling an object with both hands, the additional degrees of freedom introduce more levels to the redundancy of the object manipulation since we can distribute the contribution of the grip and manipulation forces between hands.

Methods: In this study, we investigated the forces produced by both hands during coupled bimanual manipulation of a needle object in a virtual environment.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!