New Trends in Single Photon Emission Computed Tomography (SPECT)

21K
views
54
authors
10
articles
Editors
2
Impact
Loading...
1,729 views
3 citations
Joint histogram and linear regression analysis of fast SPECT and different denoised images in the 50 testing datasets for five shorter acquisition times. The filtered FT images are used as the reference.
Original Research
03 February 2023
Fast myocardial perfusion SPECT denoising using an attention-guided generative adversarial network
Jingzhang Sun
5 more and 
Greta S. P. Mok

Purpose: Deep learning-based denoising is promising for myocardial perfusion (MP) SPECT. However, conventional convolutional neural network (CNN)-based methods use fixed-sized convolutional kernels to convolute one region within the receptive field at a time, which would be ineffective for learning the feature dependencies across large regions. The attention mechanism (Att) is able to learn the relationships between the local receptive field and other voxels in the image. In this study, we propose a 3D attention-guided generative adversarial network (AttGAN) for denoising fast MP-SPECT images.

Methods: Fifty patients who underwent 1184 MBq 99mTc-sestamibi stress SPECT/CT scan were retrospectively recruited. Sixty projections were acquired over 180° and the acquisition time was 10 s/view for the full time (FT) mode. Fast MP-SPECT projection images (1 s to 7 s) were generated from the FT list mode data. We further incorporated binary patient defect information (0 = without defect, 1 = with defect) into AttGAN (AttGAN-def). AttGAN, AttGAN-def, cGAN, and Unet were implemented using Tensorflow with the Adam optimizer running up to 400 epochs. FT and fast MP-SPECT projection pairs of 35 patients were used for training the networks for each acquisition time, while 5 and 10 patients were applied for validation and testing. Five-fold cross-validation was performed and data for all 50 patients were tested. Voxel-based error indices, joint histogram, linear regression, and perfusion defect size (PDS) were analyzed.

Results: All quantitative indices of AttGAN-based networks are superior to cGAN and Unet on all acquisition time images. AttGAN-def further improves AttGAN performance. The mean absolute error of PDS by AttcGAN-def was 1.60 on acquisition time of 1 s/prj, as compared to 2.36, 2.76, and 3.02 by AttGAN, cGAN, and Unet.

Conclusion: Denoising based on AttGAN is superior to conventional CNN-based networks for MP-SPECT.

2,477 views
4 citations
Original Research
02 February 2023

Purpose: To assess the utility of skeletal standardized uptake values (SUVs) obtained using quantitative single-photon emission computed tomography/computed tomography (SPECT/CT) in differentiating bone metastases from benign lesions, particularly in patients with lung adenocarcinoma.

Methods: Patients with lung adenocarcinoma who had undergone whole-body Tc-99m methyl-diphosphonate (99mTc-MDP) bone scans and received late phase SPECT/CT were retrospectively analyzed in this study. The maximum SUV (SUVmax); Hounsfield units (HUs); and volumes of osteoblastic, osteolytic, mixed, CT-negative metastatic and benign bone lesions, and normal vertebrae were compared. Receiver operating characteristic curves were used to determine the optimal cutoff SUVmax between metastatic and benign lesions as well as the cutoff SUVmax between CT-negative metastatic lesions and normal vertebrae. The linear correlation between SUVmax and HUs of metastatic lesions as well as that between SUVmax and the volume of all bone lesions were investigated.

Results: A total of 252 bone metastatic lesions, 140 benign bone lesions, and 199 normal vertebrae from 115 patients with lung adenocarcinoma were studied (48 males, 67 females, median age: 59 years). Metastatic lesions had a significantly higher SUVmax (23.85 ± 14.34) than benign lesions (9.67 ± 7.47) and normal vertebrae (6.19 ± 1.46; P < 0.0001). The SPECT/CT hotspot of patients with bone metastases could be distinguished from benign lesions using a cutoff SUVmax of 11.10, with a sensitivity of 87.70% and a specificity of 80.71%. The SUVmax of osteoblastic (29.16 ± 16.63) and mixed (26.62 ± 14.97) lesions was significantly greater than that of osteolytic (15.79 ± 5.57) and CT-negative (16.51 ± 6.93) lesions (P < 0.0001, P = 0.0003, and 0.002). SUVmax at the cutoff value of 8.135 could distinguish CT-negative bone metastases from normal vertebrae, with a sensitivity of 100.00% and a specificity of 91.96%. SUVmax showed a weak positive linear correlation with HUs in all bone metastases and the volume of all bone lesions.

Conclusion: SUVmax of quantitative SPECT/CT is a useful index for distinguishing benign bone lesions from bone metastases in patients with lung adenocarcinoma, particularly in the diagnosis of CT-negative bone metastases, but other factors that may affect SUVmax should be considered.

2,726 views
5 citations
2,174 views
2 citations
Open for submission
Frontiers Logo

Frontiers in Medicine

Advances in PET-CT Imaging
Edited by Nataliya Lutay, Francesco Dondi
Deadline
31 July 2024
Submit a paper
Recommended Research Topics
Frontiers Logo

Frontiers in Medicine

Perspectives in Small Animal Radionuclide Imaging
Edited by Francesco Cicone, Gaurav Malviya, Gianmario Sambuceti
33.4K
views
6
articles
Frontiers Logo

Frontiers in Medicine

Case reports in PET Imaging
Edited by Silvia Taralli, Natale Quartuccio, Gaurav Malviya
36.8K
views
71
authors
14
articles
Frontiers Logo

Frontiers in Medicine

New Diagnostic Perspectives in Urogenital Radiology
Edited by Mohamed Shehata, Mostafa Elhosseini
10.2K
views
36
authors
5
articles
Frontiers Logo

Frontiers in Medicine

Case Reports in PET Imaging 2023
Edited by Carmelo Caldarella, Ramin Sadeghi, Matteo Bauckneht
Deadline
24 Nov 2023
Submit
Frontiers Logo

Frontiers in Nuclear Medicine

Recent advances in radiotheranostics
Edited by Kondapa Bobba, Chuangyan Zhai
Deadline
02 Jul 2024
Submit