Dynamic chest radiography for pulmonary vascular diseases: clinical applications and correlation with other imaging modalities.

Dynamic chest radiography (DCR) is a novel functional radiographic imaging technique that can be used to visualize pulmonary perfusion without using contrast media. Although it has many advantages and clinical utility, most radiologists are unfamiliar with this technique because of its novelty. This review aims to (1) explain the basic principles of lung perfusion assessment using DCR, (2) discuss the advantages of DCR over other imaging modalities, and (3) review multiple specific clinical applications of DCR for pulmonary vascular diseases and compare them with other imaging modalities.

Deep learning approach using SPECT-to-PET translation for attenuation correction in CT-less myocardial perfusion SPECT imaging.

OBJECTIVE: Deep learning approaches have attracted attention for improving the scoring accuracy in computed tomography-less single photon emission computed tomography (SPECT). In this study, we proposed a novel deep learning approach referring to positron emission tomography (PET). The aims of this study were to analyze the agreement of representative voxel values and perfusion scores of SPECT-to-PET translation model-generated SPECT (SPECTSPT) against PET in 17 segments according to the American Heart Association (AHA). METHODS: This retrospective study evaluated the patient-to-patient stress, resting SPECT, and PET datasets of 71 patients. The SPECTSPT generation model was trained (stress: 979 image pairs, rest: 987 image pairs) and validated (stress: 421 image pairs, rest: 425 image pairs) using 31 cases of SPECT and PET image pairs using an image-to-image translation network. Forty of 71 cases of left ventricular base-to-apex short-axis images were translated to SPECTSPT in the stress and resting state (stress: 1830 images, rest: 1856 images). Representative voxel values of SPECT and SPECTSPT in the 17 AHA segments against PET were compared. The stress, resting, and difference scores of 40 cases of SPECT and SPECTSPT were also compared in each of the 17 segments. RESULTS: For AHA 17-segment-wise analysis, stressed SPECT but not SPECTSPT voxel values showed significant error from PET at basal anterior regions (segments #1, #6), and at mid inferoseptal regions (segments #8, #9, and #10). SPECT, but not SPECTSPT, voxel values at resting state showed significant error at basal anterior regions (segments #1, #2, and #6), and at mid inferior regions (segments #8, #9, and #11). Significant SPECT overscoring was observed against PET in basal-to-apical inferior regions (segments #4, #10, and #15) during stress. No significant overscoring was observed in SPECTSPT at stress, and only moderate over and underscoring in the basal inferior region (segment #4) was found in the resting and difference states. CONCLUSIONS: Our PET-supervised deep learning model is a new approach to correct well-known inferior wall attenuation in SPECT myocardial perfusion imaging. As standalone SPECT systems are used worldwide, the SPECTSPT generation model may be applied as a low-cost and practical clinical tool that provides powerful auxiliary information for the diagnosis of myocardial blood flow.

Myocardial Strain Derived from 13N-ammonia Positron Emission Tomography: Detection of Ischemia-Related Wall Motion Abnormality.

Background: Due to the limitation of spatial resolution, cardiac nuclear medicine images have not been applied to feature-tracking method to automatic extraction of myocardial contours. We have successfully applied the feature-tracking method to high-resolution cine 13N-ammonia positron emission tomography (PET) images to calculate the regional myocardial strains. Here, we investigate the potential of 13N-ammonia PET-derived strain to detect ischemia-related wall motion abnormality. Methods: Data of adenosine-stress/rest 13N-ammonia PET for 95 coronary artery disease patients was retrospectively analyzed. Using an original algorithm dedicated to 13N-ammonia PET, the longitudinal strain (LS) corresponding to the three main coronary artery territories [right coronary artery (RCA), left anterior descending artery (LAD), and left circumflex coronary artery (LCX)] was calculated from semi-automatic endocardial contours extraction on cine 13N-ammonia PET images of the left ventricular long-axis. The presence of ischemia in three main territories was determined from rest and stress-perfusion images. Results: In all three coronary territories, LS at stress was significantly smaller at rest in the ischemic region RCA: -19.2±8.0% vs. -22.7±6.1%, LAD: -19.0±6.9% vs. -24.4±6.4%, LCX: -20.5%±7.6% vs. -22.6±6.9%). In contrast, in the non-ischemic region, there was no significant difference between the LS at stress and at rest. Receiver-operating-characteristic analysis revealed that using the optimal cutoff of the LS ratio of stress to rest, ischemia could be diagnosed with area under the curve of 0.82 in the RCA, 0.86 in the LAD, and 0.69 in the LCX. Conclusions: Myocardial strain derived from endocardial feature-tracking of 13N-ammonia PET cine imaging is reduced in the ischemia induced by adenosine-stress. The LS ratio of stress to rest may detect wall motion abnormality related to ischemia.

Differentiation between mild and severe myocarditis using multiparametric cardiac magnetic resonance.

The pathophysiology of myocarditis is associated with mild inflammation and may progress silently, or in severe cases such as fulminant myocarditis, may lead to sudden hemodynamic compromise. An invasive myocardial biopsy is generally required for a definitive myocarditis diagnosis. Alternatively, cardiac magnetic resonance (CMR), which evaluates myocardial characteristics and cardiac function, can be used as a noninvasive tool for diagnosing myocarditis. We describe the cases of a 49-year-old woman with mild acute eosinophilic myocarditis and a 48-year-old man with severe acute lymphocytic myocarditis. CMR was performed during the acute and convalescent phases in both cases. Compared with mild myocarditis, CMR in severe myocarditis showed higher T2 values and decreased left ventricular and atrial volumes and strains; however, the right ventricular strain was preserved. Late gadolinium enhancement showed faint contrast enhancement in the whole and strong enhancement in the local myocardium. Follow-up CMR showed recovery from myocardial inflammation and cardiac function. Some late gadolinium enhancement persisted whereas acute inflammation-associated enhancement disappeared. This case report highlights the differences between the cardiac parameters of patients with mild and severe myocarditis. Severe myocardial inflammation can be caused by severe heart failure owing to the concurrent reduction of cardiac function and compliance. Additionally, preserved right ventricular strain may predict cardiac function recovery in acute myocarditis. Noninvasive and repeatable CMR provides information on myocardial characteristics, cardiac function, and hemodynamics in a single scan at that time, which is useful not only for diagnosis but also for severity assessment and patient management in acute myocarditis.

High-echoic line tracing of transthoracic echocardiography accurately assesses right ventricular enlargement in adult patients with atrial septal defect.

Accurate measurement of right ventricular (RV) size using transthoracic echocardiography (TTE) is important for evaluating the severity of congenital heart diseases. The RV end-diastolic area index (RVEDAi) determined using TTE is used to assess RV dilatation; however, the tracing line of the RVEDAi has not been clearly defined by the guidelines. This study aimed to determine the exact tracing method for RVEDAi using TTE. We retrospectively studied 107 patients with atrial septal defects who underwent cardiac magnetic resonance imaging (CMR) and TTE. We measured the RVEDAi according to isoechoic and high-echoic lines, and compared it with the RVEDAi measured using CMR. The isoechoic line was defined as the isoechoic endocardial border of the RV free wall, whereas the high-echoic line was defined as the high-echoic endocardial border of the RV free wall more outside than the isoechoic line. RVEDAi measured using high-echoic line (high-RVEDAi) was more accurately related to RVEDAi measured using CMR than that measured using isoechoic line (iso-RVEDAi). The difference in the high-RVEDAi was 0.3 cm2/m2, and the limit of agreement (LOA) was - 3.7 to 4.3 cm2/m2. With regard to inter-observer variability, high-RVEDAi was superior to iso-RVEDAi. High-RVEDAi had greater agreement with CMR-RVEDAi than with iso-RVEDAi. High-RVEDAi can become the standard measurement of RV size using two-dimensional TTE.

A deep learning model based on fusion images of chest radiography and X-ray sponge images supports human visual characteristics of retained surgical items detection.

PURPOSE: Although a novel deep learning software was proposed using post-processed images obtained by the fusion between X-ray images of normal post-operative radiography and surgical sponge, the association of the retained surgical item detectability with human visual evaluation has not been sufficiently examined. In this study, we investigated the association of retained surgical item detectability between deep learning and human subjective evaluation. METHODS: A deep learning model was constructed from 2987 training images and 1298 validation images, which were obtained from post-processing of the image fusion between X-ray images of normal post-operative radiography and surgical sponge. Then, another 800 images were used, i.e., 400 with and 400 without surgical sponge. The detection characteristics of retained sponges between the model and a general observer with 10-year clinical experience were analyzed using the receiver operator characteristics. RESULTS: The following values from the deep learning model and observer were, respectively, derived: Cutoff values of probability were 0.37 and 0.45; areas under the curves were 0.87 and 0.76; sensitivity values were 85% and 61%; and specificity values were 73% and 92%. CONCLUSION: For the detection of surgical sponges, we concluded that the deep learning model has higher sensitivity, while the human observer has higher specificity. These characteristics indicate that the deep learning system that is complementary to humans could support the clinical workflow in operation rooms for prevention of retained surgical items.

Prediction of cardiovascular events using myocardial strain ratio derived from 13N-ammonia positron emission tomography.

OBJECTIVES: Myocardial flow reserve (MFR), derived from ammonia N-13 positron emission tomography (NH3-PET), can predict the prognosis of patients with various heart diseases. We aimed to investigate whether myocardial strain ratio (MSR) was useful in predicting MACE and allowed for further risk stratification of cardiovascular events in patients with ischemic heart disease (IHD) in addition to MFR. METHODS: Ninety-five patients underwent NH3-PET because of IHD. MFR was determined as the ratio of hyperemic to resting myocardial blood flow (MBF). MSR was defined as the ratio of strains at stress and rest. The endpoint was major adverse cardiac events (MACE), including all-cause death, acute coronary syndrome, heart failure hospitalization, and revascularization. The ability to predict MACE was assessed using receiver operating characteristic (ROC) analysis, and the predictability of ME was analyzed using Kaplan-Meier analysis. The Cox proportional hazards regression model was used to calculate the hazard ratio (HR) with 95% confidence interval (CI). RESULTS: The ROC curve analysis demonstrated a cutoff of 0.93 for MACE with MSR (sensitivity and specificity of 77% and 71%, respectively). Patients with MSR < 0.93 displayed a significantly higher MACE rate than those with MSR ≥ 0.93 (p = 0.0036). The Cox proportional hazards regression analysis indicated that MSR was an independent marker that could predict MACE in imaging and clinical parameters (HR, 7.32; 95% CI: 1.59-33.7, p = 0.011). CONCLUSIONS: MSR was an independent predictor of MACE and was useful for further risk stratification in IHD. MSR has the potential for a new indicator of revascularization in patients with IHD. KEY POINTS: • We hypothesized that combining myocardial flow reserve (MFR) with the myocardial strain ratio (MSR) obtained by applying the feature-tracking technique to ammonia N-13 PET would make it predictive of major adverse cardiac events (MACE) compared to MFR alone. • MSR was an independent predictor of MACE, allowing for further risk stratification in addition to MFR in patients with ischemic heart disease. • MSR is a potential new indicator of revascularization.

Cardiac Disease Classification Using Two-Dimensional Thickness and Few-Shot Learning Based on Magnetic Resonance Imaging Image Segmentation.

Cardiac cine magnetic resonance imaging (MRI) is a widely used technique for the noninvasive assessment of cardiac functions. Deep neural networks have achieved considerable progress in overcoming various challenges in cine MRI analysis. However, deep learning models cannot be used for classification because limited cine MRI data are available. To overcome this problem, features from cine image settings are derived by handcrafting and addition of other clinical features to the classical machine learning approach for ensuring the model fits the MRI device settings and image parameters required in the analysis. In this study, a novel method was proposed for classifying heart disease (cardiomyopathy patient groups) using only segmented output maps. In the encoder-decoder network, the fully convolutional EfficientNetB5-UNet was modified to perform the semantic segmentation of the MRI image slice. A two-dimensional thickness algorithm was used to combine the segmentation outputs for the 2D representation of images of the end-diastole (ED) and end-systole (ES) cardiac volumes. The thickness images were subsequently used for classification by using a few-shot model with an adaptive subspace classifier. Model performance was verified by applying the model to the 2017 MICCAI Medical Image Computing and Computer-Assisted Intervention dataset. High segmentation performance was achieved as follows: the average Dice coefficients of segmentation were 96.24% (ED) and 89.92% (ES) for the left ventricle (LV); the values for the right ventricle (RV) were 92.90% (ED) and 86.92% (ES). The values for myocardium were 88.90% (ED) and 90.48% (ES). An accuracy score of 92% was achieved in the classification of various cardiomyopathy groups without clinical features. A novel rapid analysis approach was proposed for heart disease diagnosis, especially for cardiomyopathy conditions using cine MRI based on segmented output maps.

Right ventricular strain and volume analyses through deep learning-based fully automatic segmentation based on radial long-axis reconstruction of short-axis cine magnetic resonance images.

OBJECTIVE: We propose a deep learning-based fully automatic right ventricle (RV) segmentation technique that targets radially reconstructed long-axis (RLA) images of the center of the RV region in routine short axis (SA) cardiovascular magnetic resonance (CMR) images. Accordingly, the purpose of this study is to compare the accuracy of deep learning-based fully automatic segmentation of RLA images with the accuracy of conventional deep learning-based segmentation in SA orientation in terms of the measurements of RV strain parameters. MATERIALS AND METHODS: We compared the accuracies of the above-mentioned methods in RV segmentations and in measuring RV strain parameters by Dice similarity coefficients (DSCs) and correlation coefficients. RESULTS: DSC of RV segmentation of the RLA method exhibited a higher value than those of the conventional SA methods (0.84 vs. 0.61). Correlation coefficient with respect to manual RV strain measurements in the fully automatic RLA were superior to those in SA measurements (0.5-0.7 vs. 0.1-0.2). DISCUSSION: Our proposed RLA realizes accurate fully automatic extraction of the entire RV region from an available CMR cine image without any additional imaging. Our findings overcome the complexity of image analysis in CMR without the limitations of the RV visualization in echocardiography.

Assessment of Bi-Ventricular and Bi-Atrial Areas Using Four-Chamber Cine Cardiovascular Magnetic Resonance Imaging: Fully Automated Segmentation with a U-Net Convolutional Neural Network.

Four-chamber (4CH) cine cardiovascular magnetic resonance imaging (CMR) facilitates simultaneous evaluation of cardiac chambers; however, manual segmentation is time-consuming and subjective in practice. We evaluated deep learning based on a U-Net convolutional neural network (CNN) for fully automated segmentation of the four cardiac chambers using 4CH cine CMR. Cine CMR datasets from patients were randomly assigned for training (1400 images from 70 patients), validation (600 images from 30 patients), and testing (1000 images from 50 patients). We validated manual and automated segmentation based on the U-Net CNN using the dice similarity coefficient (DSC) and Spearman's rank correlation coefficient (ρ); p < 0.05 was statistically significant. The overall median DSC showed high similarity (0.89). Automated segmentation correlated strongly with manual segmentation in all chambers-the left and right ventricles, and the left and right atria (end-diastolic area: ρ = 0.88, 0.76, 0.92, and 0.87; end-systolic area: ρ = 0.81, 0.81, 0.92, and 0.83, respectively; p < 0.01). The area under the curve for the left ventricle, left atrium, right ventricle, and right atrium showed high scores (0.96, 0.99, 0.88, and 0.96, respectively). Fully automated segmentation could facilitate simultaneous evaluation and detection of enlargement of the four cardiac chambers without any time-consuming analysis.

13N-ammonia positron emission tomography-derived endocardial strain for the assessment of ischemia using feature-tracking in high-resolution cine imaging.

BACKGROUND: Assessing endocardial strain using a single 13N-ammonia positron emission tomography (PET) scan would be clinically useful, given the association between ischemia and myocardial deformation. However, no software has been developed for strain analysis using PET. We evaluated the clinical potential of feature tracking-derived strain values measured using PET, based on associations with the myocardial flow reserve (MFR). METHODS AND RESULTS: This retrospective study included 95 coronary artery disease patients who underwent myocardial 13N-ammonia PET. Semi-automatic measurements were made using a feature-tracking technique during myocardial cine imaging, and values were calculated using a 16-segment model. Adenosine-stressed global circumferential strain (CS) and global longitudinal strain (LS) values were compared with global MFR values. Stressed and resting global strain values were also compared. Global strain values were significantly lower in 39 patients with abnormal MFRs [< 2.0] than in 56 patients with normal MFRs [≥ 2.0]. The global CS values in the stressed state were significantly decreased than the resting state values in patients with abnormal MFRs. CONCLUSIONS: This study applied endocardial feature-tracking to 13N-ammonia PET, and the results suggested that blood flow and myocardial motility could be clinically assessed in ischemic patients using a single PET scan.

13 N-ammonia PET-derived right ventricular longitudinal strain and myocardial flow reserve in right coronary artery disease.

PURPOSE: We developed a feature-tracking algorithm for use with electrocardiography-gated high-resolution 13 N-ammonia positron emission tomography (PET) imaging, and we hypothesized it could be used to clarify the association between right ventricular (RV) longitudinal strain (LS) and right coronary artery (RCA) ischemia. The aim of this study was to investigate the association between the reduction of regional myocardial flow reserve (MFR) in RCA territories and PET-derived LS of the RV free wall. METHODS: Ninety-three patients with coronary artery stenosis > 50%, diagnosed by coronary computed tomography angiography, and 10 controls were retrospectively analyzed. RV-LS in the free wall was measured by a feature-tracking technique on the resting and stressed 13 N-ammonia PET images of horizontal long axis slices. The patients were sub-grouped according to regional MFR values at the territories of RCA, left anterior descending artery (LAD), and left circumflex coronary artery (LCx): RCA-MFR < 2.0 [n = 34], RCA-MFR ≥ 2.0 but MFR < 2.0 at LAD or LCx territories [n = 11], and MFR ≥ 2.0 for all territories [n = 48]. Stress and resting RV-LS were compared in each of the four groups. Multiple comparisons of RV-LS among the four groups were performed in the stress and resting state. RESULTS: Decreased stress RV-LS in patients with an RCA-MFR < 2.0 was observed. In the patients with MFR ≥ 2.0 for all territories, the stressed RV-LS was significantly increased compared to that in the resting state. Significantly decreased RV free wall LS during adenosine stress in patients with RCA-MFR < 2.0 was observed in the other three groups. CONCLUSIONS: We measured RV myocardial LS using feature tracking in cine imaging of 13 N-ammonia PET. The results of this study suggest that PET-derived stressed RV-LS is useful for detecting reduced RV myocardial motion due to ischemia in the RCA territory.

13N-ammonia positron emission tomography-derived left-ventricular strain in patients after heart transplantation validated using cardiovascular magnetic resonance feature tracking as reference.

OBJECTIVE: Heart transplant rejection leads to cardiac allograft vasculopathy (CAV). 13N-ammonia positron emission tomography (PET) can be useful in detecting CAV, as it can evaluate both epicardial vessels and microvasculature. In this study, we evaluated the regional wall motion in heart transplant patients using our PET-specific feature-tracking (FT) algorithm for myocardial strain calculation and validated it using a cardiovascular magnetic resonance (CMR) FT strain as a reference. METHODS: A total of 15 heart transplant patients who underwent both 13N-ammonia PET and CMR within 3 months were retrospectively enrolled. The same slice position of short-axis cine images of the middle slice of left ventricle (LV) and the same slice position of horizontal long-axis cine images were selected for the two modalities to measure the circumferential strain (CS) and longitudinal strain (LS), respectively. Based on the FT technique, time-strain curves were calculated by semi-automatic tracking of the endocardial contour on cine images throughout a cardiac cycle. The peak value in the time-strain curve was defined as the representative value. Correlations of CS and LS between PET and CMR were analyzed using Pearson correlation coefficients. The inter-modality error of strain measurements was evaluated using intraclass correlation coefficients (ICCs) with two-way random single measures. RESULTS: Excellent correlations of CS and LS between PET and CMR were observed (CS: r = 0.80; p < 0.01; LS: r = 0.87; p < 0.01). Excellent ICCs were observed (0.89 and 0.85) in CS and LS derived from PET. CONCLUSIONS: We propose the first PET strain showing an excellent agreement with the CMR strain and high reproducibility in measurement.

Unchanged right ventricular strain in repaired tetralogy of Fallot after pulmonary valve replacement with radial long-axis cine magnetic resonance images.

We measured right ventricular (RV) strain by applying a novel postprocessing technique to conventional short-axis cine magnetic resonance imaging in the repaired tetralogy of Fallot (TOF) and investigated whether pulmonary valve replacement (PVR) changes the RV strain. Twenty-four patients with repaired TOF who underwent PVR and 16 healthy controls were enrolled. Global maximum and minimum principal strains (GPSmax, GPSmin) and global circumferential and longitudinal strains (GCS, GLS) were measured from short-axis cine images reconstructed radially along the long axis. Strain parameters before and after PVR were compared using paired t-tests. One-way ANOVA with Tukey post-hoc analysis was used for comparisons between the before and after PVR groups and the control group. There were no differences in strain parameters before and after PVR. The GPSmax before PVR was lower than that in the control group (P = 0.002). Before and after PVR, GCSs were higher and GLSs were lower than those in the control group (before and after GCSs: P = 0.002 for both, before and after GLSs: P < 0.0001 and P = 0.0003). RV strains from radially reconstructed short-axis cine images revealed unchanged myocardial motion after PVR. When compared to the control group, changes in GCS and GLS in TOF patients before and after PVR might be due to RV remodeling.

Influence of beam hardening in dual-energy CT imaging: phantom study for iodine mapping, virtual monoenergetic imaging, and virtual non-contrast imaging.

In this study, we investigated the influence of beam hardening on the dual-energy computed tomography (DECT) values of iodine maps, virtual monoenergetic (VME) images, and virtual non-contrast (VNC) images. 320-row DECT imaging was performed by changing the x-ray tube energy for the first and second rotations. DECT values of 5 mg/mL iodine of the multi-energy CT phantom were compared with and without a 2-mm-thick attenuation rubber layer (~700 HU) wound around the phantom. It was found that the CT density values UH, with/without the rubber layer had statistical differences in the iodine map (184 ± 0.7 versus 186 ± 1.8), VME images (125 ± 0.3 versus 110 ± 0.4), and VNC images (-58 ± 0.7 versus -76 ± 1.7) (p < 0.010 for all). This suggests that iodine mapping may be underestimated by DECT and overestimated by VME imaging because of x-ray beam hardening. The use of VNC images instead of plain CT images requires further investigation because of underestimation.

New transluminal attenuation gradient derived from dynamic coronary CT angiography: diagnostic ability of ischemia detected by 13N-ammonia PET.

Coronary computed tomography angiography (CCTA) has low specificity for detecting significant functional coronary stenosis. We developed a new transluminal attenuation gradient (TAG)-derived dynamic CCTA with dose modulation, and we investigated its diagnostic performance for myocardial ischemia depicted by 13N-ammonia positron emission tomography (PET). Data from 48 consecutive patients who had undergone both dynamic CCTA and 13N-ammonia PET were retrospectively analyzed. Dynamic CCTA was continuously performed in mid-diastole for five cardiac cycles with prospective electrocardiography gating after a 10-s contrast medium injection. One scan of the dynamic CCTA was performed as a boost scan for conventional CCTA at the peak phase of the ascending aorta. Absolute TAG values at five phases around the boost scan were calculated. The dynamic TAG index (DTI) was defined as the ratio of the maximum absolute TAG to the standard deviation of five TAG values. We categorized the coronary territories as non-ischemia or ischemia based on the 13N-ammonia PET results. A receiver operating characteristic (ROC) analysis was performed to determine the optimal cutoff of the DTI for identifying ischemia. The DTI was significantly higher for ischemia compared to non-ischemia (8.8 ± 3.9 vs. 4.6 ± 2.0, p < 0.01). The ROC analysis revealed 5.60 as the optimal DTI cutoff to detect ischemia, with an area under the curve of 0.87, 85.7% sensitivity, and 76.2% specificity. TAG provided no additional diagnostic value for the detection of ischemia. We propose the DTI derived from dynamic CCTA as a novel coronary flow index. The DTI is a valid technique for detecting functional coronary stenosis.