Analysis of image quality by regulating beta function of BSREM reconstruction algorithm and comparison with conventional reconstructions in carcinoma breast studies of PET CT with BGO detector

Dwivedi, Pooja; Sawant, Viraj; Vajarkar, Vishal; Vatsa, Rakhee; Choudhury, Sayak; Jha, Ashish Kumar; Rangarajan, Venkatesh

doi:10.1097/mnm.0000000000001631

Cited by 2 publications

(1 citation statement)

References 26 publications

(54 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, when N subsets are used, the convergence rate is approximately N times faster. Despite its benefits, the OSEM algorithm suffers from a significant limitation in that it does not converge to the minimum when the cost function incorporates statistical noise from the projection data, which has prompted the development of alternative methods such as relaxation techniques (Browne and De Pierro 1996) and the block sequential regularized expectation maximization (BSREM) algorithm (Dwivedi et al 2022, Guo et al 2022.…”

Section: Introductionmentioning

confidence: 99%

An iterative reconstruction algorithm for unsupervised PET image

Wang,

Liu,

Xie

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Objective. In recent years, convolutional neural networks (CNNs) have shown great potential in PET image reconstruction. However, most of them rely on many low quality and high-quality reference PET image pairs for training, which are not always feasible in clinical practice. On the other hand, many works improve the quality of PET image reconstruction by adding explicit regularization or optimizing the network structure, which may lead to complex optimization problems. Approach. In this paper, we develop a novel iterative reconstruction algorithm by integrating the deep image prior (DIP) framework, which only needs the prior information (e.g., MRI) and sinogram data of patients. To be specific, we construct the objective function as a constrained optimization problem and utilize the existing PET image reconstruction packages to streamline calculations. Moreover, to further improve both the reconstruction quality and speed, we introduce the Nesterov’s acceleration part and the restart mechanism in each iteration. Main results. 2D experiments on PET data sets based on computer simulations and real patients demonstrate that our proposed algorithm can outperform existing MLEM-GF, KEM and DIPRecon methods. Significance. Unlike traditional CNN methods, the proposed algorithm does not rely on large data sets, but only leverages inter-patient information. Furthermore, we enhance reconstruction performance by optimizing the iterative algorithm. Notably, the proposed method does not require much modification of the basic algorithm, allowing for easy integration into standard implementations.

show abstract

Section: Introductionmentioning

confidence: 99%

An iterative reconstruction algorithm for unsupervised PET image

Wang,

Liu,

Xie

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

show abstract

Reduction of [68Ga]Ga-DOTA-TATE injected activity for digital PET/MR in comparison with analogue PET/CT

Cox,

Brabander,

Vegt

et al. 2024

EJNMMI Phys

View full text Add to dashboard Cite

Background New digital detectors and block-sequential regularized expectation maximization (BSREM) reconstruction algorithm improve positron emission tomography (PET)/magnetic resonance (MR) image quality. The impact on image quality may differ from analogue PET/computed tomography (CT) protocol. The aim of this study is to determine the potential reduction of injected [68Ga]Ga-DOTA-TATE activity for digital PET/MR with BSREM reconstruction while maintaining at least equal image quality compared to the current analogue PET/CT protocol. Methods NEMA IQ phantom data and 25 patients scheduled for a diagnostic PET/MR were included. According to our current protocol, 1.5 MBq [68Ga]Ga-DOTA-TATE per kilogram (kg) was injected. After 60 min, scans were acquired with 3 (≤ 70 kg) or 4 (> 70 kg) minutes per bedposition. PET/MR scans were reconstructed using BSREM and factors β 150, 300, 450 and 600. List mode data with reduced counts were reconstructed to simulate scans with 17%, 33%, 50% and 67% activity reduction. Image quality was measured quantitatively for PET/CT and PET/MR phantom and patient data. Experienced nuclear medicine physicians performed visual image quality scoring and lesion counting in the PET/MR patient data. Results Phantom analysis resulted in a possible injected activity reduction of 50% with factor β = 600. Quantitative analysis of patient images revealed a possible injected activity reduction of 67% with factor β = 600. Both with equal or improved image quality as compared to PET/CT. However, based on visual scoring a maximum activity reduction of 33% with factor β = 450 was acceptable, which was further limited by lesion detectability analysis to an injected activity reduction of 17% with factor β = 450. Conclusion A digital [68Ga]Ga-DOTA-TATE PET/MR together with BSREM using factor β = 450 result in 17% injected activity reduction with quantitative values at least similar to analogue PET/CT, without compromising on PET/MR visual image quality and lesion detectability.

show abstract

Analysis of image quality by regulating beta function of BSREM reconstruction algorithm and comparison with conventional reconstructions in carcinoma breast studies of PET CT with BGO detector

Cited by 2 publications

References 26 publications

An iterative reconstruction algorithm for unsupervised PET image

An iterative reconstruction algorithm for unsupervised PET image

Reduction of [68Ga]Ga-DOTA-TATE injected activity for digital PET/MR in comparison with analogue PET/CT

Contact Info

Product

Resources

About