The promise of artificial intelligence and deep learning in PET and SPECT imaging

Physica Medica, Год журнала: 2021, Номер 83, С. 122 - 137

Опубликована: Март 1, 2021

This review sets out to discuss the foremost applications of artificial intelligence (AI), particularly deep learning (DL) algorithms, in single-photon emission computed tomography (SPECT) and positron (PET) imaging. To this end, underlying limitations/challenges these imaging modalities are briefly discussed followed by a description AI-based solutions proposed address challenges. will focus on mainstream generic fields, including instrumentation, image acquisition/formation, reconstruction low-dose/fast scanning, quantitative imaging, interpretation (computer-aided detection/diagnosis/prognosis), as well internal radiation dosimetry. A brief algorithms fundamental architectures used for is also provided. Finally, challenges, opportunities, barriers full-scale validation adoption improvement quality accuracy PET SPECT images clinic discussed.

Язык: Английский

---

Simultaneous quantitative imaging of two PET radiotracers via the detection of positron–electron annihilation and prompt gamma emissions

Nature Biomedical Engineering, Год журнала: 2023, Номер 7(8), С. 1028 - 1039

Опубликована: Июль 3, 2023

Язык: Английский

---

Deep learning-based PET image denoising and reconstruction: a review

Radiological Physics and Technology, Год журнала: 2024, Номер 17(1), С. 24 - 46

Опубликована: Фев. 6, 2024

This review focuses on positron emission tomography (PET) imaging algorithms and traces the evolution of PET image reconstruction methods. First, we provide an overview conventional methods from filtered backprojection through to recent iterative algorithms, then deep learning for data up latest innovations within three main categories. The first category involves post-processing denoising. second comprises direct that learn mappings sinograms reconstructed images in end-to-end manner. third combine with neural-network enhancement. We discuss future perspectives technology.

Язык: Английский

---

Artificial intelligence-driven assessment of radiological images for COVID-19

Computers in Biology and Medicine, Год журнала: 2021, Номер 136, С. 104665 - 104665

Опубликована: Июль 20, 2021

Язык: Английский

---

Artificial Intelligence-Based Image Enhancement in PET Imaging

PET Clinics, Год журнала: 2021, Номер 16(4), С. 553 - 576

Опубликована: Сен. 15, 2021

Язык: Английский

---

Parametric image generation with the uEXPLORER total-body PET/CT system through deep learning

European Journal of Nuclear Medicine and Molecular Imaging, Год журнала: 2022, Номер 49(8), С. 2482 - 2492

Опубликована: Март 21, 2022

Язык: Английский

---

A review on AI in PET imaging

Annals of Nuclear Medicine, Год журнала: 2022, Номер 36(2), С. 133 - 143

Опубликована: Янв. 14, 2022

Язык: Английский

---

PET image denoising based on denoising diffusion probabilistic model

European Journal of Nuclear Medicine and Molecular Imaging, Год журнала: 2023, Номер 51(2), С. 358 - 368

Опубликована: Окт. 3, 2023

Язык: Английский

---

Decentralized Distributed Multi-institutional PET Image Segmentation Using a Federated Deep Learning Framework

Clinical Nuclear Medicine, Год журнала: 2022, Номер 47(7), С. 606 - 617

Опубликована: Апрель 20, 2022

Purpose The generalizability and trustworthiness of deep learning (DL)–based algorithms depend on the size heterogeneity training datasets. However, because patient privacy concerns ethical legal issues, sharing medical images between different centers is restricted. Our objective to build a federated DL-based framework for PET image segmentation utilizing multicentric dataset compare its performance with centralized DL approach. Methods from 405 head neck cancer patients 9 formed basis this study. All tumors were segmented manually. converted SUV maps resampled isotropic voxels (3 × 3 mm ) then normalized. subvolumes (12 12 cm consisting whole background analyzed. Data each center divided into train/validation (80% patients) test sets (20% patients). modified R2U-Net was used as core model. A parallel model developed compared approach where data are pooled one server. Segmentation metrics, including Dice similarity Jaccard coefficients, percent relative errors (RE%) peak , mean median max metabolic tumor volume, total lesion glycolysis computed manual delineations. Results versus methods nearly identical metrics: (0.84 ± 0.06 vs 0.84 0.05) (0.73 0.08 0.73 0.07). For quantitative parameters, we obtained comparable RE% (6.43% 4.72% 6.61% 5.42%), volume (12.2% 16.2% 12.1% 15.89%), (6.93% 9.6% 7.07% 9.85%) negligible . No significant differences in ( P > 2 frameworks (centralized federated) observed. Conclusion achieved respect Federated models could provide robust generalizable segmentation, while addressing issues clinical sharing.

Язык: Английский

---

AIGAN: Attention–encoding Integrated Generative Adversarial Network for the reconstruction of low-dose CT and low-dose PET images

Medical Image Analysis, Год журнала: 2023, Номер 86, С. 102787 - 102787

Опубликована: Фев. 28, 2023

Язык: Английский

---