The objective of this study is to develop a convolutional neural network (CNN) for computed tomography (CT) image super-resolution. The network learns an end-to-end mapping between low (thick-slice ...thickness) and high (thin-slice thickness) resolution images using the modified U-Net. To verify the proposed method, we train and test the CNN using axially averaged data of existing thin-slice CT images as input and their middle slice as the label. Fifty-two CT studies are used as the CNN training set, and 13 CT studies are used as the test set. We perform five-fold cross-validation to confirm the performance consistency. Because all input and output images are used in two-dimensional slice format, the total number of slices for training the CNN is 7670. We assess the performance of the proposed method with respect to the resolution and contrast, as well as the noise properties. The CNN generates output images that are virtually equivalent to the ground truth. The most remarkable image-recovery improvement by the CNN is deblurring of boundaries of bone structures and air cavities. The CNN output yields an approximately 10% higher peak signal-to-noise ratio and lower normalized root mean square error than the input (thicker slices). The CNN output noise level is lower than the ground truth and equivalent to the iterative image reconstruction result. The proposed deep learning method is useful for both super-resolution and de-noising.
In patients with acute myocardial infarction (MI) and multivessel coronary artery disease, percutaneous coronary intervention (PCI) of non-infarct-related artery reduces death or MI. However, whether ...selective PCI guided by fractional flow reserve (FFR) is superior to routine PCI guided by angiography alone is unclear. The current trial sought to compare FFR-guided PCI with angiography-guided PCI for non-infarct-related artery lesions among patients with acute MI and multivessel disease.
Patients with acute MI and multivessel coronary artery disease who had undergone successful PCI of the infarct-related artery were randomly assigned to either FFR-guided PCI (FFR ≤0.80) or angiography-guided PCI (diameter stenosis of >50%) for non-infarct-related artery lesions. The primary end point was a composite of time to death, MI, or repeat revascularization. A total of 562 patients underwent randomization. Among them, 60.0% underwent immediate PCI for non-infarct-related artery lesions and 40.0% were treated by a staged procedure during the same hospitalization. PCI was performed for non-infarct-related artery in 64.1% in the FFR-guided PCI group and 97.1% in the angiography-guided PCI group, and resulted in significantly fewer stent used in the FFR-guided PCI group (2.2 ± 1.1 vs. 2.5 ± 0.9, P < 0.001). At a median follow-up of 3.5 years (interquartile range: 2.7-4.1 years), the primary end point occurred in 18 patients of 284 patients in the FFR-guided PCI group and in 40 of 278 patients in the angiography-guided PCI group (7.4% vs. 19.7%; hazard ratio, 0.43; 95% confidence interval, 0.25-0.75; P = 0.003). The death occurred in five patients (2.1%) in the FFR-guided PCI group and in 16 patients (8.5%) in the angiography-guided PCI group; MI in seven (2.5%) and 21 (8.9%), respectively; and unplanned revascularization in 10 (4.3%) and 16 (9.0%), respectively.
In patients with acute MI and multivessel coronary artery disease, a strategy of selective PCI using FFR-guided decision-making was superior to a strategy of routine PCI based on angiographic diameter stenosis for treatment of non-infarct-related artery lesions regarding the risk of death, MI, or repeat revascularization.
We present the KASI-Yonsei Deep Imaging Survey of Clusters targeting 14 clusters at 0.015 z 0.144 using the Inamori Magellan Areal Camera and Spectrograph on the 6.5 m Magellan Baade telescope and ...the MegaCam on the 3.6 m Canada-France-Hawaii Telescope. We provide a catalog of cluster galaxies that lists magnitudes, redshifts, morphologies, bulge-to-total ratios, and local density. Based on the 1409 spectroscopically confirmed cluster galaxies brighter than −19.8 in the r band, we study galaxy morphology, color, and visual features generated by galaxy mergers. We see a clear trend between morphological content and cluster velocity dispersion, which was not presented by previous studies using local clusters. Passive spirals are preferentially found in a highly dense region (i.e., cluster center), indicating that they have gone through environmental quenching. In deep images ( r′ ∼ 27 ), 20% of our sample shows signatures of recent mergers, which is not expected from theoretical predictions and a low frequency of ongoing mergers in our sample (∼4%). Such a high fraction of recent mergers in the cluster environment supports a scenario that the merger events that made the features have preceded the galaxy accretion into the cluster environment. We conclude that mergers affect a cluster population mainly through the preprocessing of recently accreted galaxies.
Simultaneous reconstruction of activity and attenuation using the maximum-likelihood reconstruction of activity and attenuation (MLAA) augmented by time-of-flight information is a promising method ...for PET attenuation correction. However, it still suffers from several problems, including crosstalk artifacts, slow convergence speed, and noisy attenuation maps (μ-maps). In this work, we developed deep convolutional neural networks (CNNs) to overcome these MLAA limitations, and we verified their feasibility using a clinical brain PET dataset.
We applied the proposed method to one of the most challenging PET cases for simultaneous image reconstruction (
F-fluorinated-
-3-fluoropropyl-2-β-carboxymethoxy-3-β-(4-iodophenyl)nortropane
F-FP-CIT PET scans with highly specific binding to striatum of the brain). Three different CNN architectures (convolutional autoencoder CAE, Unet, and Hybrid of CAE) were designed and trained to learn a CT-derived μ-map (μ-CT) from the MLAA-generated activity distribution and μ-map (μ-MLAA). The PET/CT data of 40 patients with suspected Parkinson disease were used for 5-fold cross-validation. For the training of CNNs, 800,000 transverse PET and CT slices augmented from 32 patient datasets were used. The similarity to μ-CT of the CNN-generated μ-maps (μ-CAE, μ-Unet, and μ-Hybrid) and μ-MLAA was compared using Dice similarity coefficients. In addition, we compared the activity concentration of specific (striatum) and nonspecific (cerebellum and occipital cortex) binding regions and the binding ratios in the striatum in the PET activity images reconstructed using those μ-maps.
The CNNs generated less noisy and more uniform μ-maps than the original μ-MLAA. Moreover, the air cavities and bones were better resolved in the proposed CNN outputs. In addition, the proposed deep learning approach was useful for mitigating the crosstalk problem in the MLAA reconstruction. The Hybrid network of CAE and Unet yielded the most similar μ-maps to μ-CT (Dice similarity coefficient in the whole head = 0.79 in the bone and 0.72 in air cavities), resulting in only about a 5% error in activity and binding ratio quantification.
The proposed deep learning approach is promising for accurate attenuation correction of activity distribution in time-of-flight PET systems.
Although the presence of ischemia is a key prognostic factor in patients with coronary artery disease, the presence of high-risk plaque characteristics (HRPC) is also associated with increased risk ...of cardiovascular events. Limited data exist regarding the prognostic implications of combined information on physiological stenosis severity assessed by fractional flow reserve (FFR) and plaque vulnerability by coronary computed tomography angiography (CTA)–defined HRPC.
The current study aimed to evaluate the: 1) association between physiological stenosis severity and coronary CTA-defined HRPC; and 2) prognostic implications of coronary CTA-defined HRPC according to physiological stenosis severity in patients with coronary artery disease.
A total of 772 vessels (299 patients) evaluated by both coronary CTA and FFR were analyzed. The presence and number of HRPC (minimum lumen area <4 mm2, plaque burden ≥70%, low attenuating plaque, positive remodeling, napkin-ring sign, or spotty calcification) were assessed using coronary CTA images. The risk of vessel-oriented composite outcome (VOCO) (a composite of vessel-related ischemia-driven revascularization, vessel-related myocardial infarction, or cardiac death) at 5 years was compared according to the number of HRPC and FFR categories.
The proportion of lesions with ≥3 HRPC was significantly decreased according to the increase in FFR values (58.6%, 46.5%, 36.8%, 15.7%, and 3.5% for FFR ≤0.60, 0.61 to ≤0.70, 0.71 to ≤0.80, 0.81 to ≤0.90, and >0.90, respectively; overall p value <0.001). Both FFR and number of HRPC showed significant association with the estimated risk of VOCO (p = 0.008 and p = 0.023, respectively). In the FFR >0.80 group, lesions with ≥3 HRPC showed significantly higher risk of VOCO than those with <3 HRPC (15.0% vs. 4.3%; hazard ratio: 3.964; 95% confidence interval: 1.451 to 10.828; p = 0.007). However, there was no significant difference in the risk of VOCO according to HRPC in the FFR ≤0.80 group. By multivariable analysis, the presence of ≥3 HRPC was independently associated with the risk of VOCO in the FFR >0.80 group.
Physiological stenosis severity and the number of HRPC were closely related, and both components had significant association with the risk of clinical events. However, the prognostic implication of HRPC was different according to FFR. Integration of both physiological stenosis severity and plaque vulnerability would provide better prognostic stratification of patients than either individual component alone, especially in patients with FFR >0.80. (Clinical Implication of 3-vessel Fractional Flow Reserve 3V FFR-FRIENDS study; NCT01621438)
Display omitted
Abstract
The muscular discontinuities at the pulmonary vein (PV)–left atrial (LA) junction are known. The high-density mapping may help to find the muscular discontinuity. This study evaluated the ...efficacy of a partial antral ablation for a pulmonary vein (PV) isolation using high density (HD) mapping. A total of 60 drug-refractory atrial fibrillation (AF) patients undergoing catheter ablation were enrolled. The detailed activation mapping of each PV and LA junction was performed using an HD mapping system, and each PV segment’s activation pattern was classified into a “directly-activated from the LA” or “passively-activated from an adjacent PV segment” pattern. The antral ablations were performed at the directly-activated PV segments only when the PV had “passively-activated segments”. If the PV did not contain passively-activated segments, a circumferential antral ablation was performed on those PVs. A “successful partial antral ablation” was designated if the electrical isolation of targeted PV was achieved by ablation at the directly-activated segments only. If the isolation was not achieved even though all directly-activated segments were ablated, a “failed partial antral ablation” was designated, and then a circumferential ablation was performed. Among 240 PVs, passively-activated segments were observed in 140 (58.3%) PVs. Both inferior PVs had more passively-activated segments than superior PVs, and the posteroinferior segments had the highest proportion of passive activation. The overall rate of successful partial antral ablation was 85%. The atrial tachyarrhythmia recurrence was observed in 10 patients (16.7%) at 1-year. HD mapping allowed the evaluation of the detailed activation patterns of the PVs, and passively-activated segments may represent muscular discontinuity. Partial antral ablation of directly-activated antral segments only was feasible and effective for a PVI.
The association of the soluble suppression of tumorigenicity 2 (sST2) and the prognosis of heart failure have been well evaluated. However, little is known about the prediction of sST2 for left ...ventricular (LV) remodeling in acute coronary syndrome (ACS). We investigated the ability of sST2 to predict LV remodeling following the revascularization of ACS. From May 2019 to December 2020, 95 patients with LV ejection fraction (EF) < 50% who underwent coronary revascularization for ACS (unstable angina, non-ST-elevation myocardial infarction, ST-elevation myocardial infarction) were enrolled. Echocardiography and sST2 were performed at baseline and at a 3-month follow-up. The association between LV remodeling, using the end-diastolic volume index, and sST2 at baseline and at the 3-month follow-up, and the difference between each value was explored. During follow-up, 41 patients showed LV adverse remodeling. The baseline sST2 increased in patients without adverse remodeling (32.05 ng/mL vs. 23.5 ng/mL,
p
< 0.001), although clinical characteristics were similar between the two groups. During the mean follow-up of 3 months, a significant correlation was found in the changes between sST2 and LV end-diastolic/systolic volume index (
r
= 0.649;
p
< 0.001,
r
= 0.618;
p
< 0.001, respectively), but not in the changes of LVEF (
r
= − 0.132,
p
= 0.204). The use of angiotensin-converting enzyme 2 inhibitors/receptor blockers was higher (90.7% vs. 53.7%,
p
< 0.001) and sST2 decreased more predominantly in patients without adverse remodeling (23.18 ng/mL vs 26.40 ng/mL,
p
= 0.003). However, the changes in sST2 and LV volume were not different according to the ACS types (
p
> 0.05, for all). Estimates of the odds ratio (OR) for remodeling according to the sST2 difference increased substantially with a negative increase in the sST2 difference. Multivariable analysis found that, the difference between the baseline and 3-month sST2 was the most important determinant of LV remodeling following the revascularization of ACS (OR 1.24; 95% confidence interval: 1.09 to 1.41;
p
= 0.001). In conclusion, an increase in sST2 during follow-up was a useful predictor of LV remodeling.
BACKGROUND: Intravascular imaging–guided percutaneous coronary intervention (PCI) with intravascular ultrasound (IVUS) or optical coherence tomography (OCT) showed superior clinical outcomes compared ...with angiography-guided PCI. However, the comparative effectiveness of OCT-guided and IVUS-guided PCI regarding clinical outcomes is unknown. METHODS: In this prospective, multicenter, open-label, pragmatic trial, we randomly assigned 2008 patients with significant coronary artery lesions undergoing PCI in a 1:1 ratio to undergo either an OCT-guided or IVUS-guided PCI. The primary end point was a composite of death from cardiac causes, target vessel–related myocardial infarction, or ischemia-driven target-vessel revascularization at 1 year, which was powered for noninferiority of the OCT group compared with the IVUS group. Safety outcomes were also assessed. RESULTS: At 1 year, primary end point events occurred in 25 of 1005 patients (Kaplan-Meier estimate, 2.5%) in the OCT group and in 31 of 1003 patients (Kaplan-Meier estimate, 3.1%) in the IVUS group (absolute difference, −0.6 percentage points; upper boundary of one-sided 97.5% CI, 0.97 percentage points; P <0.001 for noninferiority). The incidence of contrast-induced nephropathy was similar (14 patients 1.4% in the OCT group versus 15 patients 1.5% in the IVUS group; P =0.85). The incidence of major procedural complications was lower in the OCT group than in the IVUS group (22 2.2% versus 37 3.7%; P =0.047), although imaging procedure-related complications were not observed. CONCLUSIONS: In patients with significant coronary artery lesions, OCT-guided PCI was noninferior to IVUS-guided PCI with respect to the incidence of a composite of death from cardiac causes, target vessel–related myocardial infarction, or ischemia-driven target-vessel revascularization at 1 year. The selected study population and lower-than-expected event rates should be considered in interpreting the trial. REGISTRATION: URL: https://www.clinicaltrials.gov ; Unique number: NCT03394079.
Background:Total bilirubin (TB) concentration is inversely associated with stable coronary artery disease, but there have been few studies on initial TB in patients with ST-segment elevation ...myocardial infarction (STEMI).Methods and Results:A total of 1,111 consecutive patients with STEMI undergoing primary percutaneous coronary intervention (PCI) with drug-eluting stents (DES) were divided into a high TB group (n=295) and a low TB group (n=816) according to the optimal cut-off 0.79 mg/dl. The high TB group had a higher rate of in-hospital major adverse cardiac events (MACE), a composite of cardiac death, non-fatal MI, and definite/probable stent thrombosis (14.2% vs. 4.2%, P<0.001) and cardiac death (13.9% vs. 3.9%, P<0.001) compared with the low TB group. The 30-day MACE-free survival rate was also significantly different between the groups (P<0.001, log-rank test). On multivariate Cox regression, initial high TB was a significant predictor of in-hospital MACE (HR, 2.69; 95% CI: 1.67–4.34, P=0.010) and of cardiac death (HR 2.72, 95% CI: 1.67–4.44, P=0.012). Adding initial TB to TIMI risk score significantly improved prediction for in-hospital MACE according to net reclassification improvement (NRI=5.2%, P=0.040) and integrated discrimination improvement (IDI=0.027, P=0.006).Conclusions:Initial TB is a powerful prognostic marker, and inclusion of this can improve prediction of in-hospital MACE in patients with STEMI undergoing primary PCI with DES. (Circ J 2016; 80: 1437–1444)
Gadolinium aluminum gallium garnet (GAGG) is a promising scintillator crystal for positron emission tomography (PET) detectors owing to its advantages of energy resolution, light yield, and absence ...of intrinsic radiation. However, a large portion of the incident photons undergoes Compton scattering within GAGG crystal because of its low stopping power compared to that of lutetium-based crystals such as Lu2SiO5 (LSO). Inter-detector scattering (IDS) and inter-crystal scattering (ICS) result in loss of sensitivity and image quality of PET, respectively. We performed a Monte Carlo simulation study to evaluate IDS recovery in our currently developing brain-dedicated PET, and extended the idea to ICS recovery. We also compared the impact of the recoveries on LSO- and GAGG-based PET scanners. We measured the sensitivity and spatial resolution of the brain PET, and analyzed the image quality using a lesion phantom, a hot-rod phantom, and a 2D Hoffman phantom with applying IDS or ICS recovery. IDS recovery increased the PET sensitivity and improved the noise level of the reconstructed images. ICS recovery enhanced the spatial resolution and the contrast of the images was improved. As the occurrence rates of IDS and ICS were higher in GAGG than in LSO, the overall impact of IDS or ICS recovery was significant in GAGG. In conclusion, we showed that the proportional method would be suitable for IDS and ICS recoveries of PET, and emphasized the importance of ICS and IDS recoveries for PET using crystals with low stopping power.
PDF
