Optimizing Attenuation Correction in <sup>68</sup>Ga-PSMA PET Imaging Using Deep Learning and Artifact-Free Dataset Refinement

Optimizing Attenuation Correction in <sup>68</sup>Ga-PSMA PET Imaging Using Deep Learning and Artifact-Free Dataset Refinement

<b>Background/Objectives:</b> Attenuation correction (AC) is essential for achieving quantitatively accurate PET imaging. In <sup>68</sup>Ga-PSMA PET, however, artifacts such as respiratory motion, halo effects, and truncation errors in CT-based AC (CT-AC) images compromise i...

Full description

Saved in:
Bibliographic Details
Main Authors: Masoumeh Dorri Giv, Guluzar Ozbolat, Hossein Arabi, Somayeh Malmir, Shahrokh Naseri, Vahid Roshan Ravan, Hossein Akbari-Lalimi, Raheleh Tabari Juybari, Ghasem Ali Divband, Nasrin Raeisi, Vahid Reza Dabbagh Kakhki, Emran Askari, Sara Harsini
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Diagnostics
Subjects:
attenuation correction
positron emission tomography computed tomography
deep learning
image artifacts
neural networks
Online Access:https://www.mdpi.com/2075-4418/15/11/1400
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:<b>Background/Objectives:</b> Attenuation correction (AC) is essential for achieving quantitatively accurate PET imaging. In <sup>68</sup>Ga-PSMA PET, however, artifacts such as respiratory motion, halo effects, and truncation errors in CT-based AC (CT-AC) images compromise image quality and impair model training for deep learning-based AC. This study proposes a novel artifact-refinement framework that filters out corrupted PET-CT images to create a clean dataset for training an image-domain AC model, eliminating the need for anatomical reference scans. <b>Methods:</b> A residual neural network (ResNet) was trained using paired PET non-AC and PET CT-AC images from a dataset of 828 whole-body <sup>68</sup>Ga-PSMA PET-CT scans. An initial model was trained using all data and employed to identify artifact-affected samples via voxel-level error metrics. These outliers were excluded, and the refined dataset was used to retrain the model with an L2 loss function. Performance was evaluated using metrics including mean error (ME), mean absolute error (MAE), relative error (RE%), RMSE, and SSIM on both internal and external test datasets. <b>Results:</b> The model trained with the artifact-free dataset demonstrated significantly improved performance: ME = −0.009 ± 0.43 SUV, MAE = 0.09 ± 0.41 SUV, and SSIM = 0.96 ± 0.03. Compared to the model trained on unfiltered data, the purified data model showed enhanced quantitative accuracy and robustness in external validation. <b>Conclusions:</b> The proposed data purification framework significantly enhances the performance of deep learning-based AC for <sup>68</sup>Ga-PSMA PET by mitigating artifact-induced errors. This approach facilitates reliable PET imaging in the absence of anatomical references, advancing clinical applicability and image fidelity.
ISSN:2075-4418

Similar Items