Deep Learning Reconstruction Research Articles

Deep Learning-accelerated MRI in Body and Chest.

Deep learning reconstruction (DLR) provides an elegant solution for MR acceleration while preserving image quality. This advancement is crucial for body imaging, which is frequently marred by the increased likelihood of motion-related artifacts. Multiple vendor-specific models focusing on T2, T1, and diffusion-weighted imaging have been developed for the abdomen, pelvis, and chest, with the liver and prostate being the most well-studied organ systems. Variational networks with supervised DL models, including data consistency layers and regularizers, are the most common DLR methods. The common theme for all single-center studies on this subject has been noninferior or superior image quality metrics and lesion conspicuity to conventional sequences despite significant acquisition time reduction. DLR also provides a potential for denoising, artifact reduction, increased resolution, and increased signal-noise ratio (SNR) and contrast-to-noise ratio (CNR) that can be balanced with acceleration benefits depending on the imaged organ system. Some specific challenges faced by DLR include slightly reduced lesion detection, cardiac motion-related signal loss, regional SNR variations, and variabilities in ADC measurements as reported in different organ systems. Continued investigations with large-scale multicenter prospective clinical validation of DLR to document generalizability and demonstrate noninferior diagnostic accuracy with histopathologic correlation are the need of the hour. The creation of vendor-neutral solutions, open data sharing, and diversifying training data sets are also critical to strengthening model robustness.

Application of a pulmonary nodule detection program using AI technology to ultra-low-dose CT: differences in detection ability among various image reconstruction methods.

This study aimed to investigate the performance of an artificial intelligence (AI)-based lung nodule detection program in ultra-low-dose CT (ULDCT) imaging, with a focus on the influence of various image reconstruction methods on detection accuracy. A chest phantom embedded with artificial lung nodules (solid and ground-glass nodules [GGNs]; diameters: 12mm, 8mm, 5mm, and 3mm) was scanned using six combinations of tube currents (160mA, 80mA, and 10mA) and voltages (120kV and 80kV) on a Canon Aquilion One CT scanner. Images were reconstructed using filtered back projection (FBP), hybrid iterative reconstruction (HIR), model-based iterative reconstruction (MBIR), and deep learning reconstruction (DLR). Nodule detection was performed using an AI-based lung nodule detection program, and performance metrics were analyzed across different reconstruction methods and radiation dose protocols. At the lowest dose protocol (80kV, 10mA), FBP showed a 0% detection rate for all nodule sizes. HIR and DLR consistently achieved 100% detection rates for solid nodules ≥ 5mm and GGNs ≥ 8mm. No method detected 3mm GGNs under any protocol. DLR demonstrated the highest detection rates, even under ultra-low-dose settings, while maintaining high image quality. AI-based lung nodule detection in ULDCT is strongly dependent on the choice of image reconstruction method.

Enhancing 18F-FDG PET image quality and lesion diagnostic performance across different body mass index using the deep progressive learning reconstruction algorithm

BackgroundAs body mass index (BMI) increases, the quality of 2-deoxy-2-[fluorine-18]fluoro-D-glucose (18F-FDG) positron emission tomography (PET) images reconstructed with ordered subset expectation maximization (OSEM) declines, negatively impacting lesion diagnostics. It is crucial to identify methods that ensure consistent diagnostic accuracy and maintain image quality. Deep progressive learning (DPL) algorithm, an Artificial Intelligence(AI)-based PET reconstruction technique, offers a promising solution.Methods150 patients underwent 18F-FDG PET/CT scans and were categorized by BMI into underweight, normal, and overweight groups. PET images were reconstructed using both OSEM and DPL and their image quality was assessed both visually and quantitatively. Visual assessment employed a 5-point Likert scale to evaluate overall score, image sharpness, image noise, and diagnostic confidence. Quantitative assessment parameters included the background liver image-uniformity-index () and signal-to-noise ratio (). Additionally, 466 identifiable lesions were categorized by size: sub-centimeter and larger. We compared maximum standard uptake value (), signal-to-background ratio (), , contrast-to-background ratio (), and contrast-to-noise ratio () of these lesions to evaluate the diagnostic performance of the DPL and OSEM algorithms across different lesion sizes and BMI categories.ResultsDPL produced superior PET image quality compared to OSEM across all BMI groups. The visual quality of DPL showed a slight decline with increasing BMI, while OSEM exhibited a more significant decline. DPL maintained a stable across BMI increases, whereas OSEM exhibited increased noise. In the DPL group, quantitative image quality for overweight patients matched that of normal patients with minimal variance from underweight patients. In contrast, OSEM demonstrated significant declines in quantitative image quality with rising BMI. DPL yielded significantly higher contrast (, ,) and than OSEM for all lesions across all BMI categories.ConclusionDPL consistently provided superior image quality and lesion diagnostic performance compared to OSEM across all BMI categories in 18F-FDG PET/CT scans. Therefore, we recommend using the DPL algorithm for 18F-FDG PET/CT image reconstruction in all BMI patients.

Free-breathing pediatric cardiac dark-blood imaging with reverse double inversion-recovery and single-shot deep learning reconstruction

Free-breathing pediatric cardiac dark-blood imaging with reverse double inversion-recovery and single-shot deep learning reconstruction

Super-Resolution Deep Learning Reconstruction for T2*-Weighted Images: Improvement in Microbleed Lesion Detection and Image Quality.

Super-resolution deep learning reconstruction (SR-DLR) is a promising tool for improving image quality by enhancing spatial resolution compared to conventional deep learning reconstruction (DLR). This study aimed to evaluate whether SR-DLR improves microbleed detection and visualization in brain magnetic resonance imaging (MRI) compared to DLR. This retrospective study included 69 patients (66.2 ± 13.8years; 44 females) who underwent 3T brain MRI with T2*-weighted 2D gradient echo and 3D flow-sensitive black blood imaging (reference standard) between June and August 2024. T2*-weighted images were reconstructed using SR-DLR and DLR. Three blinded readers detected microbleeds and assessed image quality, including microbleed and normal structure visibility, sharpness, noise, artifacts, and overall quality. Quantitative analysis involved measuring signal intensity along the septum pellucidum. Microbleed detection performance was analyzed using jackknife alternative free-response receiver operating characteristic analysis, while image quality was analyzed using the Wilcoxon signed-rank test and paired t-test. SR-DLR significantly outperformed DLR in microbleed detection (figure of merit: 0.690 vs. 0.645, p < 0.001). SR-DLR also demonstrated higher sensitivity for microbleed detection. Qualitative analysis showed better microbleed visualization for two readers (p < 0.001) and improved image sharpness for all readers (p ≤ 0.008). Quantitative analysis revealed enhanced sharpness, especially in full width at half maximum and edge rise slope (p < 0.001). SR-DLR improved image sharpness and quality, leading to better microbleed detection and visualization in brain MRI compared to DLR.

Application of deep learning reconstruction combined with time-resolved post-processing method to improve image quality in CTA derived from low-dose cerebral CT perfusion data

BackgroundTo assess the effect of the combination of deep learning reconstruction (DLR) and time-resolved maximum intensity projection (tMIP) or time-resolved average (tAve) post-processing method on image quality of CTA derived from low-dose cerebral CTP.MethodsThirty patients underwent regular dose CTP (Group A) and other thirty with low-dose (Group B) were retrospectively enrolled. Group A were reconstructed with hybrid iterative reconstruction (R-HIR). In Group B, four image datasets of CTA were gained: L-HIR, L-DLR, L-DLRtMIP and L-DLRtAve. The CT attenuation, image noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR) and subjective images quality were calculated and compared. The Intraclass Correlation (ICC) between CTA and MRA of two subgroups were calculated.ResultsThe low-dose group achieved reduction of radiation dose by 33% in single peak arterial phase and 18% in total compared to the regular dose group (single phase: 0.12 mSv vs 0.18 mSv; total: 1.91mSv vs 2.33mSv). The L-DLRtMIP demonstrated higher CT values in vessels compared to R-HIR (all P < 0.05). The CNR of vessels in L-HIR were statistically inferior to R-HIR (all P < 0.001). There was no significant different in image noise and CNR of vessels between L-DLR and R-HIR (all P > 0.05, except P = 0.05 for CNR of ICAs, 77.19 ± 21.64 vs 73.54 ± 37.03). However, the L-DLRtMIP and L-DLRtAve presented lower image noise, higher CNR (all P < 0.05) and subjective scores (all P < 0.001) in vessels than R-HIR. The diagnostic accuracy in Group B was excellent (ICC = 0.944).ConclusionCombining DLR with tMIP or tAve allows for reduction in radiation dose by about 33% in single peak arterial phase and 18% in total in CTP scanning, while further improving image quality of CTA derived from CTP data when compared to HIR.

Improvement of image quality of diffusion-weighted imaging (DWI) with deep learning reconstruction of the pancreas: comparison with respiratory-gated conventional DWI.

This study aimed to evaluate the efficacy of deep learning-based reconstruction(DLR) in improving pancreatic diffusion-weighted imaging (DWI) quality. In total, 117 patients (mean age of 68.0 ± 12.9years) suspected of pancreatic diseases underwent magnetic resonance imaging (MRI) between July and December 2023. MRI sequences includedrespiratory-gated conventional diffusion-weighted images (RGC-DWIs), respiratory-gated diffusion-weighted images with deep learning-based reconstruction (DLR) (RGDLR-DWIs), and breath-hold diffusion-weighted images with DLR (BHDLR-DWIs) (short TE and long TE equal to other DWIs) at a 3T MR system. Among these patients, 27 had solid lesions. Two radiologists qualitatively assessed pancreatic shape, main pancreatic duct (MPD) visualization, and solid lesion conspicuity using a 5-point scale. Quantitative analysis included apparent diffusion coefficient (ADC)values for pancreatic parenchyma and solid lesions, signal-to-noise ratio (SNR), pancreas-to-muscle signal-intensity ratio (PM-SIR) and lesion-to-pancreas signal-intensity ratio (LP-SIR). Differences among DWI sequences were analyzed using Friedman's and Bonferroni's tests. Qualitatively, BHDLR-DWIs (short TE) had the highest scores for pancreatic shape and MPD but lowest for solid lesions visibility, whereas RGDLR-DWIs had the highest score for solid lesions. Quantitatively, BHDLR-DWIs (short TE) had the lowest ADC values for pancreatic parenchyma and solid lesions, with the highest PM-SIR. There was no significant difference between BHDLR-DWIs (short TE) and RGDLR-DWIs for solid lesion ADC values. RGC-DWIs had the highest SNR, though differences from RGDLR-DWIs and BHDLR-DWIs (short TE) were not significant. Although LP-SIR in RGDLR-DWIs were the lowest, the difference was not significant. BHDLR-DWIs (short TE) provided the best pancreatic morphology image quality, whereas RGDLR-DWIs were superior for solid lesion detection.

Super-resolution deep learning reconstruction to evaluate lumbar spinal stenosis status on magnetic resonance myelography.

To investigate whether super-resolution deep learning reconstruction (SR-DLR) of MR myelography-aided evaluations of lumbar spinal stenosis. In this retrospective study, lumbar MR myelography of 40 patients (16 males and 24 females; mean age, 59.4 ± 31.8years) were analyzed. Using the MR imaging data, MR myelography was separately reconstructed via SR-DLR, deep learning reconstruction (DLR), and conventional zero-filling interpolation (ZIP). Three radiologists, blinded to patient background data and MR reconstruction information, independently evaluated the image sets in terms of the following items: the numbers of levels affected by lumbar spinal stenosis; and cauda equina depiction, sharpness, noise, artifacts, and overall image quality. The median interobserver agreement in terms of the numbers of lumbar spinal stenosis levels were 0.819, 0.735, and 0.729 for SR-DLR, DLR, and ZIP images, respectively. The imaging quality of the cauda equina, and image sharpness, noise, and overall quality on SR-DLR images were significantly better than those on DLR and ZIP images, as rated by all readers (p < 0.001, Wilcoxon signed-rank test). No significant differences were observed for artifacts on SR-DLR against DLR and ZIP. SR-DLR improved the image quality of lumbar MR myelographs compared to DLR and ZIP, and was associated with better interobserver agreement during assessment of lumbar spinal stenosis status.

Deep learning reconstruction for detection of liver lesions at standard-dose and reduced-dose abdominal CT.

Deep learning reconstruction (DLR) has shown promising image denoising ability, but its radiation dose reduction potential remains unknown. The objective of this study was to investigate the diagnostic performance of DLR compared to iterative reconstruction (IR) in the detection of liver lesions at standard-dose and reduced-dose CT. Participants with known liver metastases from gastrointestinal and pancreatic adenocarcinoma were prospectively included from routine follow-up (October 2020 to March 2022). Participants received standard-dose CT and two additional reduced-dose scans during the same contrast administration, each reconstructed with IR and high-strength DLR. Two radiologists evaluated images for the presence of liver lesions, and a third established a reference standard. Diagnostic performance was compared using McNemar's test and mixed effects logistic regression. Forty-four participants (mean age 66 years ± 11 [standard deviation], 28 men) were evaluated with 348 included liver lesions ≤ 20 mm (297 metastases, 51 benign; mean size 9.1 ± 4.3 mm). Mean volume CT dose index was 14.2, 7.8 mGy, and 5.1 mGy. Between algorithms, no significant difference in lesion detection was observed within dose levels. Detection of 233 lesions ≤ 10 mm was deteriorated with lower dose levels despite DLR denoising, with 185 detected at standard-dose IR (79.4%; 95% CI: 73.5-84.3) vs 128 at medium-dose DLR (54.9%; 95% CI: 48.3-61.4; p < 0.001) and 105 at low-dose DLR (45.1%; 95% CI: 38.6-51.7; p < 0.001). Diagnostic performance for liver lesion detection was comparable between algorithms. When the detection of smaller lesions is important, DLR did not facilitate substantial dose reduction. Question Methods to reduce CT radiation dose are desirable in clinical practice, and DLR has shown promising image denoising capabilities. Findings Liver lesion detection was comparable for DLR and IR across dose levels, but detection of smaller lesions deteriorated with lower dose levels. Clinical relevance Although potent in image noise reduction, the diagnostic performance of DLR is comparable to IR at standard-dose and reduced-dose CT. Care must be taken in pursuit of dose reduction when the detection and characterization of smaller liver lesions are of clinical importance.

Deep learning reconstruction of diffusion-weighted imaging with single-shot echo-planar imaging in endometrial cancer: a comparison with multi-shot echo-planar imaging.

To evaluate the efficacy of deep learning reconstruction (DLR) in diffusion-weighted imaging (DWI) with single-shot echo-planar imaging (SSEPI) for endometrial cancer, compared to multiplexed sensitivity-encoding (MUSE) DWI. We retrospectively reviewed 31 women with surgically confirmed endometrial cancer who underwent preoperative pelvic magnetic resonance imaging (MRI) including DWI. Qualitative analysis including overall image quality, susceptibility artifacts, sharpness of the uterine edge, and lesion conspicuity were compared among conventional SSEPI (SSEPI-C), SSEPI with DLR (SSEPI-DL), and MUSE using the Friedman's test. Quantitative analysis including the apparent diffusion coefficient (ADC) values, noise, signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were also compared among three DWI sequences using the Friedman's test. In addition, the diagnostic accuracy for deep myometrial invasion was compared to three DWI sequences using Cochran's Q test. The scores of overall image quality, sharpness of the uterine edge, and lesion conspicuity in SSEPI-DL were higher than SSEPI-C (p < 0.001) with no significant difference compared to MUSE (p > 0.05). Noise in SSEPI-DL was lower than SSEPI-C (p < 0.001), with no significant difference compared to MUSE (p > 0.05). SNR and CNR in SSEPI-DL were also superior to SSEPI-C (p < 0.001), and comparable to MUSE (p > 0.05). The diagnostic accuracy for detecting deep myometrial invasion showed no significant difference among SSEPI-C, SSEPI-DL and MUSE (p > 0.05). DLR improves the image quality of DWI in endometrial cancer, demonstrating image quality equivalent to that of SSEPI-DL and MUSE. SSEPI-DL can be an alternative to MUSE in female pelvic MRI, with the benefit of significantly shortened scan time.

DWI of the rectum with deep learning reconstruction: comparison of PROPELLER, reduced FOV, and conventional DWI.

To compare the image quality and diagnostic performance of periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER), reduced field-of-view (rFOV), and conventional diffusion-weighted imaging (cDWI) combined with deep learning reconstruction (DLR) for evaluating rectal tumors. This prospective study included 42 MRI examinations of 38 patients with rectal tumors who underwent initial staging and/or restaging MRI. PROPELLER-DWI, rFOV-DWI, and cDWI obtained with DLR were reviewed by two radiologists and compared for image quality and diagnostic performance for local tumor extent at staging and restaging and response to chemoradiotherapy at restaging. PROPELLER-DWI had significantly the least artifacts and distortions, but the worst perceptive noise, while rFOV-DWI had significantly the best sharpness for both readers (P < 0.01). For overall image quality and rectal/tumor conspicuity, PROPELLER-DWI and rFOV-DWI were significantly superior to cDWI in both readers (P < 0.01). The incidence of suboptimal image quality was significantly lower with PROPELLER-DWI and rFOV-DWI than with cDWI (5 and 1 patients with PROPELLER-DWI, 14 and 6 with rFOV-DWI, and 29 and 25 with cDWI by the 2 readers, P < 0.01). Although there were no significant differences in the accuracy of staging and restaging among the 3 types of DWI, inter-reader agreement was highest for PROPELLER-DWI (weighted kappa, 0.62-0.71) compared with cDWI (weighted kappa, 0.38-0.52) and rFOV-DWI (weighted kappa, 0.47-0.61). PROPELLER-DWI and rFOV-DWI with DLR may improve the image quality of rectal DWI by reducing artifacts and distortions or increasing sharpness, although the impact on diagnostic accuracy was not significant.

Physics-Informed Deep Learning Reconstruction for Ultrafast Clinical 3D Fluid-Attenuated Inversion Recovery Brain MRI

Abstract Background Physics-informed deep learning (DL) reconstructions show promise in accelerating MRI yet have not been extensively validated, particularly for 3D fluid-attenuated inversion recovery (FLAIR) sequence. Purpose To evaluate the diagnostic quality and interchangeability of DL-based 3D FLAIR with a state-of-the-art acceleration technique (wave-controlled aliasing in parallel imaging [Wave-CAIPI] FLAIR) in a clinical setting with 3T brain MRI. Materials and Methods Participants undergoing evaluation for demyelinating disease between October and December of 2023 were prospectively enrolled at a single center. For each participant, state-of-the-art Wave-CAIPI FLAIR and a resolution-matched 6-fold-under-sampled Cartesian FLAIR acquisition with DL reconstruction were performed at 3-T system (MAGNETOM Vida, Siemens Healthineers, Erlangen, Germany). Four neuroradiologists evaluated overall image quality, anatomic conspicuity, lesion conspicuity, and imaging artifacts. Lesion count, volume, and regional brain volume were compared between imaging methods. Inter-reader agreement was assessed using quadratic weighted Cohen’s Kappa and Kendall's correlation coefficient. Agreement of continuous metrics were evaluated using intraclass correlation coefficients (ICC), linear regression, and Bland-Altman analysis. Interchangeability regarding the quantitative metrics was evaluated with individual equivalence index (IEI). Results 88 participants (61 women [69%], 47±13 years) were evaluated. DL-FLAIR reduced scan time (1:53 vs. 2:50) and showed higher overall image quality, anatomic conspicuity, lesion conspicuity, imaging artifacts compared with state-of-the-art technique (all Ps &amp;lt; 0.001). DL-FLAIR also demonstrated higher signal-to-noise ratio and contrast-to-noise ratio compared to Wave-CAIPI-FLAIR, with high agreement in lesion and regional brain volumes between both methods (ICC(2,k) range, 0.91 to 0.99). DL-FLAIR proved interchangeable with Wave-CAIPI-FLAIR for lesion count (IEI: 0.10, acceptable proportion: 0.977, 95% CI: [0.943, 1.000]) and for lesion volume (IEI: 0.32, acceptable proportion: 0.966, 95% CI: [0.930, 1.000]). Conclusion Deep learning reconstruction of 3D-FLAIR provides higher image quality compared to a state-of-the-art technique with 30% less scan time, while maintaining excellent agreement and interchangeability in quantitative evaluation.

Deep Learning-Driven Abbreviated Shoulder MRI Protocols: Diagnostic Accuracy in Clinical Practice.

Deep learning (DL) reconstruction techniques have shown promise in reducing MRI acquisition times while maintaining image quality. However, the impact of different acceleration factors on diagnostic accuracy in shoulder MRI remains unexplored in clinical practice. The purpose of this study was to evaluate the diagnostic accuracy of 2-fold and 4-fold DL-accelerated shoulder MRI protocols compared to standard protocols in clinical practice. In this prospective single-center study, 88 consecutive patients (49 males, 39 females; mean age, 51 years) underwent shoulder MRI examinations using standard, 2-fold (DL2), and 4-fold (DL4) accelerated protocols between June 2023 and January 2024. Four independent radiologists (experience range: 4-25 years) evaluated the presence of bone marrow edema (BME), rotator cuff tears, and labral lesions. The sensitivity, specificity, and interobserver agreement were calculated. Diagnostic confidence was assessed using a 4-point scale. The impact of reader experience was analyzed by stratifying the radiologists into ≤10 and >10 years of experience. Both accelerated protocols demonstrated high diagnostic accuracy. For BME detection, DL2 and DL4 achieved 100% sensitivity and specificity. In rotator cuff evaluation, DL2 showed a sensitivity of 98-100% and specificity of 99-100%, while DL4 maintained a sensitivity of 95-98% and specificity of 99-100%. Labral tear detection showed perfect sensitivity (100%) with DL2 and slightly lower sensitivity (89-100%) with DL4. Interobserver agreement was excellent across the protocols (Kendall's W = 0.92-0.98). Reader experience did not significantly impact diagnostic performance. The area under the ROC curve was 0.94 for DL2 and 0.90 for DL4 (p = 0.32). The implementation of DL-accelerated protocols, particularly DL2, could improve workflow efficiency by reducing acquisition times by 50% while maintaining diagnostic reliability. This could increase patient throughput and accessibility to MRI examinations without compromising diagnostic quality. DL-accelerated shoulder MRI protocols demonstrate high diagnostic accuracy, with DL2 showing performance nearly identical to that of the standard protocol. While DL4 maintains acceptable diagnostic accuracy, it shows a slight sensitivity reduction for subtle pathologies, particularly among less experienced readers. The DL2 protocol represents an optimal balance between acquisition time reduction and diagnostic confidence.

Comparison of 18F-FDG PET image quality and quantitative parameters between DPR and OSEM reconstruction algorithm in patients with lung cancer

ObjectivesThe present study aimed to investigate the influence of the deep progressive learning reconstruction (DPR) algorithm on the 18F-FDG PET image quality and quantitative parameters.MethodsIn this retrospective study, data were collected from 55 healthy individuals and 184 patients with primary malignant pulmonary tumors who underwent 18F-FDG PET/CT examinations. PET data were reconstructed using the ordered subset expectation maximization (OSEM) and DPR algorithms. The influence of DPR algorithm on quantitative parameters was explored, including the SUVmax, SUVmean, standard deviation of SUV (SUVSD), metabolic tumor volume (MTV), total lesion glycolysis (TLG), and tumor-to-background uptake ratio (TBR). Finally, the differences in image quality parameters, including signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR), between the two reconstruction algorithms were evaluated.ResultsDPR algorithm significantly reduced the SUVmax and SUVSD of background tissues (all, P < 0.001) compared to OSEM algorithm, while no statistical difference was observed in SUVmean between the two algorithms (all, P > 0.05). DPR algorithm notably increased the SUVmax, SUVmean, and TBR of lesions (all, P < 0.001) and reduced MTV (P = 0.005), with minimal differences in TLG noted between the reconstruction algorithms (P < 0.001). The percentage differences in SUVmax (P = 0.001), SUVmean (P = 0.005), and TBR (P = 0.001) between the two algorithms were significantly higher in solid nodules than in pure ground glass nodules (pGGNs). The ΔCNR between solid nodules (P = 0.031) and mixed ground glass nodules (P = 0.020) was greater than that between pGGNs. SNR and CNR obtained using the DPR algorithm were markedly improved compared to those determined using the OSEM algorithm (all, P < 0.001).ConclusionUnder identical acquisition conditions, the DPR algorithm enhanced the accuracy of quantitative parameters in pulmonary lesions and potentially improved lesion detectability. The DPR algorithm increased image SNR and CNR compared to those obtained using the OSEM algorithm, significantly optimizing overall image quality. This advancement facilitated precise clinical diagnosis, underpinning its potential to significantly contribute to the field of medical imaging.

Novel Deep Learning Reconstruction to Augment Contrast Enhancement: Initial Evaluation.

To assess image quality between single-energy CT (SECT) and dual-energy CT (DECT) scans compared with a novel deep learning (DL) reconstruction for SECT used to improve contrast enhancement. The raw data from a prior prospective HIPAA-compliant study (March through August 2022) was used to create a novel reconstruction in patients with biopsy-proven colorectal adenocarcinoma and liver metastases. Patients underwent 120kVp SECT and DECT (50keV reconstruction) abdominal scans in the portal venous phase in the same breath hold. Two readers independently assessed the scans. The final study group was 13 men and 2 women with a mean age of 60 years ± 10, a mean height of 171cm ± 8, a mean weight of 87kg ± 23, and a mean body mass index of 30kg/m2 ± 6. Liver, pancreas, spleen, psoas muscle, and aorta HUs were all significantly higher with the virtual DL reconstruction compared with the 120kVp series, but significantly lower than the 50keV series (P<0.05). Readers scored the DL reconstruction to have better contrast enhancement than the standard 120kVp series and improved artifacts, noise texture, and resolution compared with the 50keV series (P<0.05). Contrast enhancement with the new reconstruction is superior compared with the standard 120kVp series approaching that of 50keV DECT, but with improved perception of artifacts, noise texture, and resolution.

Fast and Robust Single-Shot Cine Cardiac MRI Using Deep Learning Super-Resolution Reconstruction

Objective The aim of the study was to compare the diagnostic quality of deep learning (DL) reconstructed balanced steady-state free precession (bSSFP) single-shot (SSH) cine images with standard, multishot (also: segmented) bSSFP cine (standard cine) in cardiac MRI. Methods and Materials This prospective study was performed in a cohort of participants with clinical indication for cardiac MRI. SSH compressed-sensing bSSFP cine and standard multishot cine were acquired with breath-holding and electrocardiogram-gating in short-axis view at 1.5 Tesla. SSH cine images were reconstructed using an industry-developed DL super-resolution algorithm (DL-SSH cine). Two readers evaluated diagnostic quality (endocardial edge definition, blood pool to myocardium contrast and artifact burden) from 1 (nondiagnostic) to 5 (excellent). Functional left ventricular (LV) parameters were assessed in both sequences. Edge rise distance, apparent signal-to-noise ratio (aSNR) and contrast-to-noise ratio were calculated. Statistical analysis for the comparison of DL-SSH cine and standard cine included the Student's t-test, Wilcoxon signed-rank test, Bland-Altman analysis, and Pearson correlation. Results Forty-five participants (mean age: 50 years ±18; 30 men) were included. Mean total scan time was 65% lower for DL-SSH cine compared to standard cine (92 ± 8 s vs 265 ± 33 s; P &lt; 0.0001). DL-SSH cine showed high ratings for subjective image quality (eg, contrast: 5 [interquartile range {IQR}, 5–5] vs 5 [IQR, 5–5], P = 0.01; artifacts: 4.5 [IQR, 4–5] vs 5 [IQR, 4–5], P = 0.26), with superior values for sharpness parameters (endocardial edge definition: 5 [IQR, 5–5] vs 5 [IQR, 4–5], P &lt; 0.0001; edge rise distance: 1.9 [IQR, 1.8–2.3] vs 2.5 [IQR, 2.3–2.6], P &lt; 0.0001) compared to standard cine. No significant differences were found in the comparison of objective metrics between DL-SSH and standard cine (eg, aSNR: 49 [IQR, 38.5–70] vs 52 [IQR, 38–66.5], P = 0.74). Strong correlation was found between DL-SSH cine and standard cine for the assessment of functional LV parameters (eg, ejection fraction: r = 0.95). Subgroup analysis of participants with arrhythmia or unreliable breath-holding (n = 14/45, 31%) showed better image quality ratings for DL-SSH cine compared to standard cine (eg, artifacts: 4 [IQR, 4–5] vs 4 [IQR, 3–5], P = 0.04). Conclusions DL reconstruction of SSH cine sequence in cardiac MRI enabled accelerated acquisition times and noninferior diagnostic quality compared to standard cine imaging, with even superior diagnostic quality in participants with arrhythmia or unreliable breath-holding.

Image quality and diagnostic performance of deep learning reconstruction for diffusion- weighted imaging in 3 T breast MRI.

Image quality and diagnostic performance of deep learning reconstruction for diffusion- weighted imaging in 3 T breast MRI.

SuperMRF: deep robust reconstruction for highly accelerated magnetic resonance fingerprinting.

Magnetic resonance fingerprinting (MRF) is a rapid imaging technique for simultaneous mapping of multiple tissue properties such as T1 and T2 relaxation times. However, conventional pattern matching reconstruction and iterative low rank reconstruction methods may not take full advantage of the spatiotemporal content of MRF data and can require significant computational resources with long reconstruction times. Deep learning reconstruction using a three-dimensional (3D) convolutional neural network (CNN)-based method may enable high-quality, rapid MRF reconstruction. Evaluation of such proposed deep learning reconstruction methods for MRF is needed to clarify whether deep learning techniques adapted from other MR image reconstruction problems will yield benefits when employed in MRF applications. The objective of this study is to design and evaluate a novel deep learning framework (SuperMRF) that directly transforms undersampled parameter-weighted 3D Cartesian MRF data into quantitative T1 and T2 maps, bypassing traditional pattern-matching in MRF. In contrast to conventional MRF where only the temporal evolution of each voxel is used for quantification, SuperMRF exploits both two-dimensional spatial and one-dimensional temporal information with a 3D CNN for reconstruction. Controlled simulation experiments were performed using reference parameter maps from in vivo knee scans of healthy volunteers. To evaluate the robustness to noise, we trained our network using clean data and tested it on simulated noisy data. Conventional inner product-based pattern matching and state-of-the-art iterative low rank reconstruction techniques were used for comparison. The performance of all methods was evaluated with respect to structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and normalized mean squared error (NMSE). Prospective real-world MRF scans were performed in four volunteer subjects using the trained network from simulations and cartilage and muscle T1 and T2 values were compared between conventional pattern matching, low rank reconstruction, and SuperMRF. SuperMRF estimated accurate T1 and T2 mapping in a highly accelerated scan (15× undersampling in k-space with a 20-fold reduction in the number of acquired MRF frames) with low error (NMSE of 5%) and high resemblance (SSIM of 94%) to reference quantitative maps. SuperMRF was observed to be superior to the conventional and low rank MRF reconstruction methods in terms of NMSE, SSIM, and robustness to noise. In prospective real-world data, SuperMRF provided comparable T1 and T2 maps as compared to low rank MRF. The only significantly different cartilage and muscle values in prospective data across the three reconstruction methods were those from conventional MRF T2. Our results demonstrate that the proposed SuperMRF can achieve rapid, robust reconstruction with reduced frames in addition to k-space undersampling, outperforming the conventional and state-of-the-art reconstruction methods in simulation and providing comparable results to low rank reconstruction in prospective real-world subjects.

Combination of deep learning reconstruction and quantification for dynamic contrast-enhanced (DCE) MRI.

Combination of deep learning reconstruction and quantification for dynamic contrast-enhanced (DCE) MRI.

Deep Learning Reconstruction for Prostate MRI: Impact of Field-of-View Selection on Global and Regional Image Quality

Deep Learning Reconstruction for Prostate MRI: Impact of Field-of-View Selection on Global and Regional Image Quality