Search

Deep learning-based convolutional neural networks (CNNs) have been proposed for enhancing the quality of digital tomosynthesis (DTS) images. However, the direct applications of the conventional CNNs for low-dose DTS imaging are limited to provide acceptable image quality due to the inaccurate recognition of complex texture patterns. In this study, a deep residual learning network combined with the Anscombe transformation was proposed for simplifying the complex texture and restoring the low-dose DTS image quality. The proposed network consisted of convolution layers, max-pooling layers, up-sampling layers, and skip connections. The network training was performed to learn the residual images between the ground-truth and low-dose projections, which were converted using the Anscombe transformation. As a result, the proposed network enhanced the quantitative accuracy and noise characteristic of DTS images by 1.01–1.27 and 1.14–1.71 times, respectively, in comparison to low-dose DTS images and other deep learning networks. The spatial resolution of the DTS image restored using the proposed network was 1.12 times higher than that obtained using a deep image learning network. In conclusion, the proposed network can restore the low-dose DTS image quality and provide an optimal model for low-dose DTS imaging.

This study aimed to analyze the applicability of the improved median filter in positron emission tomography (PET)/magnetic resonance (MR) fusion images based on parallel imaging using generalized autocalibrating partially parallel acquisition (GRAPPA). In this study, a PET/MR fusion imaging system based on a 3.0T magnetic field and 18F radioisotope were used. An improved median filter that can set a mask of the median value more efficiently than before was modeled and applied to the acquired image. As quantitative evaluation parameters of the noise level, the contrast to noise ratio (CNR) and coefficient of variation (COV) were calculated. Additionally, no-reference-based evaluation parameters were used to analyze the overall image quality. We confirmed that the CNR and COV values of the PET/MR fusion images to which the improved median filter was applied improved by approximately 3.32 and 2.19 times on average, respectively, compared to the noisy image. In addition, the no-reference-based evaluation results showed a similar trend for the noise-level results. In conclusion, we demonstrated that it can be supplemented by using an improved median filter, which suggests the problem of image quality degradation of PET/MR fusion images that shortens scan time using GRAPPA.

Among the detector materials available at room temperature, thallium bromide (TlBr), which has a relatively high atomic number and density, is widely used for gamma camera imaging. This study aimed to verify the usefulness of TlBr through quantitative evaluation by modeling detectors of various compound types using Monte Carlo simulations. The Geant4 application for tomographic emission was used for simulation, and detectors based on cadmium zinc telluride and cadmium telluride materials were selected as a comparison group. A pixel-matched parallel-hole collimator with proven excellent performance was modeled, and phantoms used for quality control in nuclear medicine were used. The signal-to-noise ratio (SNR), contrast to noise ratio (CNR), sensitivity, and full width at half maximum (FWHM) were used for quantitative analysis to evaluate the image quality. The SNR, CNR, sensitivity, and FWHM for the TlBr detector material were approximately 1.05, 1.04, 1.41, and 1.02 times, respectively, higher than those of the other detector materials. The SNR, CNR and sensitivity increased with increasing detector thickness, but the spatial resolution in terms of FWHM decreased. Thus, we demonstrated the feasibility and possibility of using the TlBr detector material in comparison with commercial detector materials.

In integrated positron emission tomography (PET)/magnetic resonance imaging (MRI) systems, the field of view (FOV) in MRI is smaller than in PET, the difference of FOV between PET and MRI can cause geometric distortion and truncation artifacts at off-center positions. Especially, breast PET/MRI can easily lead to off-center positions around the back of the body by prone patient on the breast radiofrequency (RF) coil, it leads to truncation artifact in MRI, the PET imaging requires the attenuation correction within PET FOV using an attenuation map by acquiring MRI data. However, MRI-based attenuation maps containing truncation artifacts result in errors in visibility and quantification for PET imaging due to the limitations of MRI FOV. The maximum likelihood estimation of activity and attenuation (MLAA)-based algorithm can estimate the truncated part of an attenuation map from PET emission data. This study aimed to reduce truncation artifacts using MLAA-based attenuation correction (AC) in breast PET/MRI. A uniform phantom and national electrical manufacturers’ association international electro-technical commission body PET phantom, simulated using a Siemens Biograph PET/MRI scanner with a breast RF coil, was evaluated. PET images were evaluated between the conventional Dixon volume interpolated breath-hold examination (VIBE) of the MRI-based attenuation map and MLAA-based attenuation map by predicting the missing body contour. In this study, the MLAA algorithm was joint estimation based attenuation map composed of Dixon VIBE with a PET emission data. PET images with Dixon VIBE AC showed truncation artifacts along the horizontal axis of patient at the off-center of the FOV. The contours of the phantoms were recovered using the MLAA algorithm. The PET activity concentration with Dixon VIBE in truncated regions was 15–17% of the non-truncated regions; however, this increased to 91–97% when using the MLAA algorithm. The MLAA algorithm in breast PET/MRI reduces truncation artifacts and improves the visibility and quantification.