Journal of Thoracic Imaging最新文献

Purpose: To study the performance of artificial intelligence (AI) for detecting pleural pathology on chest radiographs (CXRs) using computed tomography as ground truth.

Patients and methods: Retrospective study of subjects undergoing CXR in various clinical settings. Computed tomography obtained within 24 hours of the CXR was used to volumetrically quantify pleural effusions (PEfs) and pneumothoraxes (Ptxs). CXR was evaluated by AI software (INSIGHT CXR; Lunit) and by 3 second-year radiology residents, followed by AI-assisted reassessment after a 3-month washout period. We used the area under the receiver operating characteristics curve (AUROC) to assess AI versus residents' performance and mixed-model analyses to investigate differences in reading time and interreader concordance.

Results: There were 96 control subjects, 165 with PEf, and 101 with Ptx. AI-AUROC was noninferior to aggregate resident-AUROC for PEf (0.82 vs 0.86, P < 0.001) and Ptx (0.80 vs 0.84, P = 0.001) detection. AI-assisted resident-AUROC was higher but not significantly different from the baseline. AI-assisted reading time was reduced by 49% (157 vs 80 s per case, P = 0.009), and Fleiss kappa for Ptx detection increased from 0.70 to 0.78 ( P = 0.003). AI decreased detection error for PEf (odds ratio = 0.74, P = 0.024) and Ptx (odds ratio = 0.39, P < 0.001).

Conclusion: Current AI technology for the detection of PEf and Ptx on CXR was noninferior to second-year resident performance and could help decrease reading time and detection error.

Purpose: Pleural plaques (PPs) are morphologic manifestations of long-term asbestos exposure. The relationship between PP and lung function is not well understood, whereas the time-consuming nature of PP delineation to obtain volume impedes research. To automate the laborious task of delineation, we aimed to develop automatic artificial intelligence (AI)-driven segmentation of PP. Moreover, we aimed to explore the relationship between pleural plaque volume (PPV) and pulmonary function tests.

Materials and methods: Radiologists manually delineated PPs retrospectively in computed tomography (CT) images of patients with occupational exposure to asbestos (May 2014 to November 2019). We trained an AI model with a no-new-UNet architecture. The Dice Similarity Coefficient quantified the overlap between AI and radiologists. The Spearman correlation coefficient ( r ) was used for the correlation between PPV and pulmonary function test metrics. When recorded, these were vital capacity (VC), forced vital capacity (FVC), and diffusing capacity for carbon monoxide (DLCO).

Results: We trained the AI system on 422 CT scans in 5 folds, each time with a different fold (n = 84 to 85) as a test set. On these independent test sets combined, the correlation between the predicted volumes and the ground truth was r = 0.90, and the median overlap was 0.71 Dice Similarity Coefficient. We found weak to moderate correlations with PPV for VC (n = 80, r = -0.40) and FVC (n = 82, r = -0.38), but no correlation for DLCO (n = 84, r = -0.09). When the cohort was split on the median PPV, we observed statistically significantly lower VC ( P = 0.001) and FVC ( P = 0.04) values for the higher PPV patients, but not for DLCO ( P = 0.19).

Conclusion: We successfully developed an AI algorithm to automatically segment PP in CT images to enable fast volume extraction. Moreover, we have observed that PPV is associated with loss in VC and FVC.

Purpose: Shortened injection durations are not recommended in step-and-shoot coronary computed tomography angiography (CCTA). We aimed to evaluate the image quality of CCTA performed using bodyweight-adjusted iodinated contrast media (ICM) with different injection durations to generate an optimized ICM administration protocol to acquire convincible image quality in step-and-shoot CCTA.

Materials and methods: A total of 200 consecutive patients with suspected coronary artery disease (CAD) were enrolled in group A (N=50, 350 mgI/mL, bodyweight×0.8 mL/kg with a 13-s injection duration), group B (N=50, 350 mgI/mL, bodyweight×0.9 mL/kg with a 13-s injection duration), group C (N=50, 350 mgI/mL, bodyweight×0.8 mL/kg with a 12-s injection duration), and group D (N=50, 320 mgI/mL, bodyweight×0.8 mL/kg with a 13-s injection duration). Patient characteristics, ICM administration protocols, quantitative computed tomography (CT) value measurements, and qualitative image scores were analyzed and compared among the groups.

Results: Groups A and D achieved the lowest ICM volume, saline volume, injection flow rate, and total iodine and iodine injection rates among the groups. All the CT values of the coronary arteries in all groups were >300 HU. All the observers' average scores exceeded three points. In group A, the CT values showed significant positive correlation with the iodine injection rate ( r =0.226, P <0.001), whereas the signal-to-noise ratio ( r =-0.004, P =0.927) and contrast-to-noise ratio ( r =-0.006, P =0.893) values were not.

Conclusions: Bodyweight×0.8 mL/kg with a 13-second injection duration is a comprehensive option for step-and-shoot CCTA with improved image quality, and a 350 mgI/mL iodine concentration is preferred.

Purpose: Spectral computed tomography (CT) enables improved tissue characterization, although virtually all research has focused on contrast-enhanced examinations. We hypothesized that changes in myocardial tissue related to acute myocardial infarction (AMI) might potentially be identified without the need for contrast administration using electron density (ED) imaging.

Patients and methods: This retrospective observational study involved a small series (n = 15) of patients admitted to our institution with a first AMI without signs of hemodynamic instability and identification of a culprit vessel with invasive coronary angiography during the same admission, who also underwent a noncontrast, low-dose chest CT using a dual-layer spectral CT scanner. Images were assessed in search of dark areas with low density on ED imaging, and the mean percentage ED relative to water (%EDW) was calculated.

Results: Using a qualitative approach, ED assessment enabled the identification of 11/15 (73%) affected coronary territories, with a sensitivity of 73% (95% CI: 45; 92%) and a specificity of 87% (95% CI: 69; 96%). AMI segments showed significantly lower ED values than the remote myocardium (103.8 ± 0.8 vs 104.3 ± 0.6 %EDW, P < 0.0001), and a threshold below 103.9 %EDW had a sensitivity of 66% and specificity of 79% for the identification of AMI. In a control group of patients without a history of cardiovascular disease, none had areas with focal reduction of ED following the shape of the myocardial wall.

Conclusions: In our preliminary series, ED imaging showed the potential to enable the identification of myocardial tissue changes related to AMI without iodinated contrast requirement.

Objective: Noninvasive measurement of myocardial work (MW) incorporates left ventricular (LV) pressure, and, therefore, allows correction of global longitudinal strain for changing afterload conditions. We sought to investigate MW as a tool to detect early signs of LV dysfunction in primary systemic hypertension patients, particularly with different predictive indices.

Methods and results: None left ventricular hypertrophy (NLVH) and left ventricular hypertrophy (LVH) patients established were all primary systemic hypertension with preserved ejection fraction. Forty in NLVH and forty in LVH according to left ventricular end-diastolic mass index (LVEDmassI) were prospectively enrolled. The following indices of MW were assessed: global work index, global constructive work, global wasted work (GWW), and global work efficiency (GWE). Both global work index ( P =0.348) and global constructive work ( P =0.225) were increased in NLVH and decreased in LVH, and GWW ( P <0.001) was increased significantly in NLVH and increased more in LVH, while GWE ( P <0.001) was decreased significantly in NLVH and decreased more in LVH. The clinical utility of GWW (95% CI: 0.802-0.951) and GWE (95% CI: 0.811-0.950) were verified by receiver-operating characteristic curve analysis showing larger net benefits as evaluated with LVH and control comparisons. In multivariate linear regression analysis, 4-dimenaional LVEDmassI was independently associated with GWE ( P =0.018) in systemic hypertension patients. Assessment of intraobserver and interobserver variability in the MW echocardiographic data documented good interclass correlation coefficients (all >0.85).

Conclusion: GWW and GWE derived from MW are more accurate, sensitive, and reproducible predictors to detect early LV dysfunction in primary systemic hypertension patients, especially in distinguishing the potential functional abnormality of NLVH and LVH, even though the ejection fraction is preserved.

Purpose: Pericardial fat (PF) and epicardial adipose tissue (EAT) may enhance the proinflammatory response in corona virus-19 (COVID-19) patients. Higher PF and EAT volumes might result in multiorgan failure and explain unfavorable trajectories.The aim of this study was to examine the association between the volume of PF and EAT and multiorgan failure over time.

Materials and methods: All mechanically ventilated COVID-19 patients with an available chest computed tomography were prospectively included (March-June 2020). PF and EAT volumes were quantified using chest computed tomography scans. Patients were categorized into sex-specific PF and EAT tertiles. Variables to calculate Sequential Organ Failure Assessment (SOFA) scores were collected daily to indicate multiorgan failure. Linear mixed-effects regression was used to investigate the association between tertiles for PF and EAT volumes separately and serial SOFA scores over time. All models were adjusted.

Results: Sixty-three patients were divided into PF and EAT tertiles, with median PF volumes of 131.4 mL (IQR [interquartile range]: 115.7, 143.2 mL), 199.8 mL (IQR: 175.9, 221.6 mL), and 318.8 mL (IQR: 281.9, 376.8 mL) and median EAT volumes of 69.6 mL (IQR: 57.0, 79.4 mL), 107.9 mL (IQR: 104.6, 115.1 mL), and 163.8 mL (IQR: 146.5, 203.1 mL). Patients in the highest PF tertile had a statistically significantly lower SOFA score over time (1.3 [-2.5, -0.1], P =0.033) compared with the lowest PF tertile. EAT tertiles were not significantly associated with SOFA scores over time.

Conclusion: A higher PF volume is associated with less multiorgan failure in mechanically ventilated COVID-19 patients. EAT volumes were not associated with multiorgan failure.

Purpose: To evaluate the left atrial (LA) function in participants with apical hypertrophic cardiomyopathy (AHCM) by cardiovascular magnetic resonance feature tracking (CMR-FT).

Materials and methods: Thirty typical AHCM (TAHCM) patients, 23 subclinical AHCM (SAHCM) patients and 32 normal healthy volunteers who underwent CMR exam were retrospectively analyzed. LA reservoir, conduit, and contractile function were quantified by volumetric and CMR-FT derived strain and strain rate (SR) parameters from 2-chamber and 4-chamber cine imaging.

Results: Compared with healthy participants, both TAHCM and SAHCM patients had impaired LA reservoir function (total strain [%]: TAHCM 31.3±12.2, SAHCM 31.8±12.3, controls 40.4±10.7, P <0.01; total SR [/s]: TAHCM 1.1±0.4, SAHCM 1.1±0.5, controls 1.4 ± 0.4, P <0.01) and conduit function (passive strain [%]: TAHCM 14.4±7.6, SAHCM 16.4±8.8, controls 23.3±8.1, P <0.01; passive SR [/s]: TAHCM -0.5±0.3, SAHCM -0.6±0.3, controls -1.0±0.4, P <0.01). Regarding contraction function, although TAHCM and SAHCM patients had preserved active emptying fraction and strain (all P >0.05), patients with TAHCM had the lowest active SR value among the 3 groups ( P= 0.03). LA reservoir and conduit strain were both significantly associated with left ventricular mass index and maximal wall thickness (all P <0.05). A moderate correlation between LA passive SR and left ventricular cardiac index ( P <0.01).

Conclusions: The LA reservoir and conduit function are predominately impaired and appeared in both SAHCM and TAHCM patients.

Purpose: To compare dual-energy computed tomography (DECT) based qualitative and quantitative parameters in chronic thromboembolic pulmonary hypertension with various postoperative primary and secondary endpoints.

Materials and methods: This was a retrospective analysis of 64 patients with chronic thromboembolic pulmonary hypertension who underwent DECT. First, a clot score was calculated by assigning the following score: pulmonary trunk-5, each main pulmonary artery-4, each lobar-3, each segmental-2, and subsegmental-1 per lobe; the sum total was then calculated. The perfusion defect (PD) score was calculated by assigning 1 point to each segmental PD. The combined score was calculated by adding clot and PD scores. For quantitative evaluation, we calculated perfused blood volume (PBV) (%) of each lung and the sum of both lungs. Primary endpoints included testing association between combined score and total PBV with change in mean pulmonary arterial pressure ([mPAP], change calculated as preop minus postop values). Secondary endpoints included explorative analysis of the correlation between combined score and PBV with change in preoperative and postoperative pulmonary vascular resistance, change in preoperative 6-minute walk distance (6MWD), and immediate postoperative complications such as reperfusion edema, ECMO placement, stroke, death and mechanical ventilation for more than 48 hours, all within 1 month of surgery.

Results: Higher combined scores were associated with larger decreases in mPAP ( =0.27, P =0.036). On average, the decrease in mPAP (pre mPAP-post mPAP) increased by 2.2 mm Hg (95% CI: -0.6, 5.0) with each 10 unit increase in combined score. The correlation between total PBV and change in mPAP was small and not statistically significant. During an exploratory analysis, higher combined scores were associated with larger increases in 6MWD at 6 months postprocedure ( =0.55, P =0.002).

Conclusion: Calculation of DECT-based combined score offers potential in the evaluation of hemodynamic response to surgery. This response can also be objectively quantified.

Purpose: Previous studies demonstrated the impact of sex on left ventricular (LV) strain in patients with essential hypertension. However, little is known about the effect of sex on left atrial (LA) strain in patients with hypertension. This study aimed to explore the sex-related differences of LA strain by using cardiac magnetic resonance feature tracking in patients with hypertension and preserved LV ejection fraction.

Materials and methods: One hundred and fifty hypertensive patients (100 men and 50 women) and 105 age-matched and sex-matched normotensive controls (70 men and 35 women) were retrospectively enrolled and underwent cardiac magnetic resonance examination. LA strain parameters included LA reservoir strain (εs), conduit strain (εe), pump strain (εa), and their corresponding strain rate (SRs, SRe, and SRa).

Results: Men had significantly higher LV mass index, lower εs and εe than women in both patients and controls (all P <0.05). LA strain and strain rate were significantly reduced in hypertensive patients compared with controls, both in men and women (all P <0.05). In men, hypertension and its interaction were associated with increased LV mass index and decreased εs and εe. In multivariable analysis, men, LV ejection fraction, and LA minimum volume index remained independent determinants of εs and εe in all hypertensive patients (all P <0.05).

Conclusion: LA strain was significantly impaired in hypertensive patients, and men had more impaired LA strain than women. These findings further emphasize the sex-related differences in the response of LA strain to hypertension in the early stage.