Image reconstruction using UNET-transformer network for fast and low-dose PET scans.

INTRODUCTION: Low-dose and fast PET imaging (low-count PET) play a significant role in enhancing patient safety, healthcare efficiency, and patient comfort during medical imaging procedures. To achieve high-quality images with low-count PET scans, effective reconstruction models are crucial for denoising and enhancing image quality. The main goal of this paper is to develop an effective and accurate deep learning-based method for reconstructing low-count PET images, which is a challenging problem due to the limited amount of available data and the high level of noise in the acquired images. The proposed method aims to improve the quality of reconstructed PET images while preserving important features, such as edges and small details, by combining the strengths of UNET and Transformer networks. MATERIAL AND METHODS: The proposed TrUNET-MAPEM model integrates a residual UNET-transformer regularizer into the unrolled maximum a posteriori expectation maximization (MAPEM) algorithm for PET image reconstruction. A loss function based on a combination of structural similarity index (SSIM) and mean squared error (MSE) is utilized to evaluate the accuracy of the reconstructed images. The simulated dataset was generated using the Brainweb phantom, while the real patient dataset was acquired using a Siemens Biograph mMR PET scanner. We also implemented state-of-the-art methods for comparison purposes: OSEM, MAPOSEM, and supervised learning using 3D-UNET network. The reconstructed images are compared to ground truth images using metrics such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and relative root mean square error (rRMSE) to quantitatively evaluate the accuracy of the reconstructed images. RESULTS: Our proposed TrUNET-MAPEM approach was evaluated using both simulated and real patient data. For the patient data, our model achieved an average PSNR of 33.72 dB, an average SSIM of 0.955, and an average rRMSE of 0.39. These results outperformed other methods which had average PSNRs of 36.89 dB, 34.12 dB, and 33.52 db, average SSIMs of 0.944, 0.947, and 0.951, and average rRMSEs of 0.59, 0.49, and 0.42. For the simulated data, our model achieved an average PSNR of 31.23 dB, an average SSIM of 0.95, and an average rRMSE of 0.55. These results also outperformed other state-of-the-art methods, such as OSEM, MAPOSEM, and 3DUNET-MAPEM. The model demonstrates the potential for clinical use by successfully reconstructing smooth images while preserving edges. The comparison with other methods demonstrates the superiority of our approach, as it outperforms all other methods for all three metrics. CONCLUSION: The proposed TrUNET-MAPEM model presents a significant advancement in the field of low-count PET image reconstruction. The results demonstrate the potential for clinical use, as the model can produce images with reduced noise levels and better edge preservation compared to other reconstruction and post-processing algorithms. The proposed approach may have important clinical applications in the early detection and diagnosis of various diseases.

Standardizing SPECT/CT dosimetry following radioembolization with yttrium-90 microspheres.

BACKGROUND: Multiple post-treatment dosimetry methods are currently under investigation for Yttrium-90 ([Formula: see text]) radioembolization. Within each methodology, a variety of dosimetric inputs exists that affect the final dose estimates. Understanding their effects is essential to facilitating proper dose analysis and crucial in the eventual standardization of radioembolization dosimetry. The purpose of this study is to investigate the dose differences due to different self-calibrations and mass density assignments in the non-compartmental and local deposition methods. A practical mean correction method was introduced that permits dosimetry in images where the quality is compromised by patient motion and partial volume effects. METHODS: Twenty-one patients underwent [Formula: see text] radioembolization and were imaged with SPECT/CT. Five different self-calibrations (FOV, Body, OAR, Liverlung, and Liver) were implemented and dosimetrically compared. The non-compartmental and local deposition method were used to perform dosimetry based on either nominal- or CT calibration-based mass densities. A mean correction method was derived assuming homogeneous densities. Cumulative dose volume histograms, linear regressions, boxplots, and Bland Altman plots were utilized for analysis. RESULTS: Up to 270% weighted dose difference was found between self-calibrations with mean dose differences up to 50 Gy in the liver and 23 Gy in the lungs. Between the local deposition and non-compartmental methods, the liver and lung had dose differences within 0.71 Gy and 20 Gy, respectively. The local deposition method's nominal and CT calibration-based mass density implementations dosimetric metrics were within 1.4% in the liver and 24% in the lungs. The mean lung doses calculated with the CT method were shown to be inflated. The mean correction method demonstrated that the corrected mean doses were greater by up to [Formula: see text] Gy in the liver and lower by up to [Formula: see text] Gy in the lungs. CONCLUSIONS: The OAR calibration may be utilized as a potentially more accurate and precise self-calibration. The non-compartmental method was found more comparable to the local deposition method in organs that were more homogeneous in mass densities. Due to the potential for inflated lung mean doses, the non-compartmental and local deposition method implemented with nominal mass densities is recommended for more consistent dosimetric results. If patient motion and partial volume effects are present in the liver, our practical correction method will calculate more representative doses in images suboptimal for dosimetry.

Practical considerations for establishing dead-time corrections in quantitative SPECT imaging.

Quantitative SPECT studies require specific information about the equipment being used. Particularly in the context of therapeutic studies, the effect of dead-time can be significant and must be quantified. We explored different techniques for measuring the dead-time constant and applying dead-time corrections to the data. METHOD: The dead-time constant was measured on four similar SPECT/CT systems by following the response of the system to a uniform phantom initially containing 17 GBq of Lu-177 over a period of 23 days. It was then calculated using the two-source method with 1 332 MBq of Tc-99 m. The dead-time constant found was used to correct SPECT/CT phantom images either applying the correction by projection or globally on the image. RESULTS: Both methods of calculating the dead-time constant produced equivalent results. However, the dead-time constant varied by as much as 8% between machines of the same model and manufacturer. Correcting for dead-time by projection rather than globally produced slightly more precise results (0.94% error rather than 2.59% error). The benefit of this correction technique will be dependent on the level of asymmetry in the patient as well as the magnitude of the dead-time correction effect. CONCLUSION: quantification of the dead-time of a system can be performed quickly using the two-source method and any radioisotope. However, it is important to perform this measurement on every system being used. In vastly asymmetric images with high dead-time correction, correcting for dead-time by projection can be pertinent, increasing the precision of dosimetry calculations by several percent. However this additional gain may be within the error of SUV measurements for many clinical acquisitions.

A guide to 90Y radioembolization and its dosimetry.

Radioembolization gains continuous traction as a primarily palliative radiation treatment for hepatic tumours. A form of nuclear medicine therapy, Yttrium-90 containing microspheres are catheter guided and injected into the right, left, or a specifically selected hepatic artery. A multitude of comprehensive planning steps exist to ensure a thorough and successful treatment. Clear clinical and physiological guidelines have been established and nuclear imaging is used to plan and verify dose distributions. Radioembolization's treatment rationale is based on tumour and blood vessel dynamics that allow a targeted treatment approach. However, radioembolization's dosimetry is grossly oversimplified. In fact, the currently utilized clinical dosimetric standards (e.g. partition method) have persisted since the 1990s. Moreover, the multitude of radioembolization's intertwining components lies disjointed within the literature. Particularly relevant to new readers, this review provides a methodical guide that presents the treatment rationale behind every clinical step. The emerging dosimetry methods and its factors are further discussed to provide a comprehensive review on an essential research direction.