Deep boosted regression for mr to ct synthesis

TY - JOUR

T1 - Deep boosted regression for mr to ct synthesis

AU - Kläser, Kerstin

AU - Markiewicz, Pawel

AU - Ranzini, Marta

AU - Li, Wenqi

AU - Modat, Marc

AU - Hutton, Brian F.

AU - Atkinson, David

AU - Thielemans, Kris

AU - Cardoso, M. Jorge

AU - Ourselin, Sébastien

PY - 2018/9/12

Y1 - 2018/9/12

N2 - Attenuation correction is an essential requirement of positron emission tomography (PET) image reconstruction to allow for accurate quantification. However, attenuation correction is particularly challenging for PET-MRI as neither PET nor magnetic resonance imaging (MRI) can directly image tissue attenuation properties. MRI-based computed tomography (CT) synthesis has been proposed as an alternative to physics based and segmentation-based approaches that assign a population-based tissue density value in order to generate an attenuation map. We propose a novel deep fully convolutional neural network that generates synthetic CTs in a recursive manner by gradually reducing the residuals of the previous network, increasing the overall accuracy and generalisability, while keeping the number of trainable parameters within reasonable limits. The model is trained on a database of 20 pre-acquired MRI/CT pairs and a four-fold random bootstrapped validation with a 80:20 split is performed. Quantitative results show that the proposed framework outperforms a state-of-the-art atlas-based approach decreasing the Mean Absolute Error (MAE) from 131HU to 68HU for the synthetic CTs and reducing the PET reconstruction error from 14.3% to 7.2%.

AB - Attenuation correction is an essential requirement of positron emission tomography (PET) image reconstruction to allow for accurate quantification. However, attenuation correction is particularly challenging for PET-MRI as neither PET nor magnetic resonance imaging (MRI) can directly image tissue attenuation properties. MRI-based computed tomography (CT) synthesis has been proposed as an alternative to physics based and segmentation-based approaches that assign a population-based tissue density value in order to generate an attenuation map. We propose a novel deep fully convolutional neural network that generates synthetic CTs in a recursive manner by gradually reducing the residuals of the previous network, increasing the overall accuracy and generalisability, while keeping the number of trainable parameters within reasonable limits. The model is trained on a database of 20 pre-acquired MRI/CT pairs and a four-fold random bootstrapped validation with a 80:20 split is performed. Quantitative results show that the proposed framework outperforms a state-of-the-art atlas-based approach decreasing the Mean Absolute Error (MAE) from 131HU to 68HU for the synthetic CTs and reducing the PET reconstruction error from 14.3% to 7.2%.

UR - http://www.scopus.com/inward/record.url?scp=85053905341&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-00536-8_7

DO - 10.1007/978-3-030-00536-8_7

M3 - Conference paper

AN - SCOPUS:85053905341

SN - 0302-9743

VL - 11037

SP - 61

EP - 70

JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

T2 - 3rd International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2018 Held in Conjunction with 21st International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2018

Y2 - 16 September 2018 through 16 September 2018

ER -