Human AI Teaming for Coronary CT Angiography Assessment: Impact on Imaging Workflow and Diagnostic Accuracy.

As the number of coronary computed tomography angiography (CTA) examinations is expected to increase, technologies to optimize the imaging workflow are of great interest. The aim of this study was to investigate the potential of artificial intelligence (AI) to improve clinical workflow and diagnostic accuracy in high-volume cardiac imaging centers. A total of 120 patients (79 men; 62.4 (55.0-72.7) years; 26.7 (24.9-30.3) kg/m2) undergoing coronary CTA were randomly assigned to a standard or an AI-based (human AI) coronary analysis group. Severity of coronary artery disease was graded according to CAD-RADS. Initial reports were reviewed and changes were classified. Both groups were similar with regard to age, sex, body mass index, heart rate, Agatston score, and CAD-RADS. The time for coronary CTA assessment (142.5 (106.5-215.0) s vs. 195.0 (146.0-265.5) s; p < 0.002) and the total reporting time (274.0 (208.0-377.0) s vs. 350 (264.0-445.5) s; p < 0.02) were lower in the human AI than in the standard group. The number of cases with no, minor, or CAD-RADS relevant changes did not differ significantly between groups (52, 7, 1 vs. 50, 8, 2; p = 0.80). AI-based analysis significantly improves clinical workflow, even in a specialized high-volume setting, by reducing CTA analysis and overall reporting time without compromising diagnostic accuracy.

Unidimensional Longitudinal Strain: A Simple Approach for the Assessment of Longitudinal Myocardial Deformation by Echocardiography.

BACKGROUND: Impaired left ventricular (LV) longitudinal function (LF) is a known predictor of cardiac events in patients with heart failure, but two-dimensional strain imaging, the reference method to measure myocardial deformation, is not always feasible or available. Therefore, reliable and reproducible alternatives are needed. The aim of the present study was to evaluate unidimensional longitudinal strain (ULS) as a simple echocardiographic parameter for the assessment of LV LF. METHODS: Two hundred two patients with dilated cardiomyopathy who had their first presentation in the authors' cardiology department, as well as the same number of age- and gender-matched control subjects, were prospectively included in this study. ULS was compared with global longitudinal strain (GLS), the current gold standard for LV LF assessment by echocardiography. Uni- and multivariate Cox regression analyses were conducted to evaluate the prognostic value of ULS. RESULTS: LV LF was higher in the control group compared with patients: GLS -19.5 ± 1.7% versus -12.6 ± 4.8% and ULS -16.3 ± 1.5% versus -10.2 ± 3.9% (P < .001 for each). Correlation between ULS and GLS was excellent (r = 0.94), while Bland-Altman plots revealed lower values for ULS (bias -2.76%, limits of agreement ±3.31%). During a mean follow-up time of 39 months, the combined end point of cardiovascular death or hospitalization for acute cardiac decompensation was reached by 28 patients (13.9%). GLS (hazard ratio, 1.21; 95% CI, 1.10-1.34; P < .001) and ULS (hazard ratio, 1.24; 95% CI, 1.12-1.39; P < .001) had comparable prognostic impact on patient outcomes. CONCLUSIONS: ULS might be an alternative echocardiographic method for the assessment of LV LF, with similar diagnostic and prognostic value compared with GLS.

Assessment of global longitudinal strain using standardized myocardial deformation imaging: a modality independent software approach.

BACKGROUND: Myocardial deformation measurement is superior to left ventricular ejection fraction in identifying early changes in myocardial contractility and prediction of cardiovascular outcome. The lack of standardization hinders its clinical implementation. The aim of the study is to investigate a novel standardized deformation imaging approach based on the feature tracking algorithm for the assessment of global longitudinal (GLS) and global circumferential strain (GCS) in echocardiography and cardiac magnetic resonance imaging (CMR). METHODS: 70 subjects undergoing CMR were consecutively investigated with echocardiography within a median time of 30 min. GLS and GCS were analyzed with a post-processing software incorporating the same standardized algorithm for both modalities. Global strain was defined as the relative shortening of the whole endocardial contour length and calculated according to the strain formula. RESULTS: Mean GLS values were -16.2 ± 5.3 and -17.3 ± 5.3 % for echocardiography and CMR, respectively. GLS did not differ significantly between the two imaging modalities, which showed strong correlation (r = 0.86), a small bias (-1.1 %) and narrow 95 % limits of agreement (LOA ± 5.4 %). Mean GCS values were -17.9 ± 6.3 and -24.4 ± 7.8 % for echocardiography and CMR, respectively. GCS was significantly underestimated by echocardiography (p < 0.001). A weaker correlation (r = 0.73), a higher bias (-6.5 %) and wider LOA (± 10.5 %) were observed for GCS. GLS showed a strong correlation (r = 0.92) when image quality was good, while correlation dropped to r = 0.82 with poor acoustic windows in echocardiography. GCS assessment revealed only a strong correlation (r = 0.87) when echocardiographic image quality was good. No significant differences for GLS between two different echocardiographic vendors could be detected. CONCLUSIONS: Quantitative assessment of GLS using a standardized software algorithm allows the direct comparison of values acquired irrespective of the imaging modality. GLS may, therefore, serve as a reliable parameter for the assessment of global left ventricular function in clinical routine besides standard evaluation of the ejection fraction.

Reliability of noninvasive assessment of systolic pulmonary artery pressure by Doppler echocardiography compared to right heart catheterization: analysis in a large patient population.

BACKGROUND: Pulmonary artery pressure (PAP) is an important marker in cardiovascular disorders, being closely associated with morbidity and mortality. Noninvasive assessment by Doppler echocardiography is recommended by current guidelines. So far, the reliability of this method has been assessed only in small studies with contradictory results. Therefore, the aim of this study was to analyze the reliability of noninvasive PAP assessment by Doppler echocardiography compared to invasive measurements in a large patient population. METHODS AND RESULTS: We retrospectively analyzed data from a large tertiary cardiology department over 6 years in order to compare invasively measured PAP to estimated PAP from echocardiography examinations. N=15 516 patients fulfilled inclusion criteria and n=1695 patients with timely matched examinations (within 5 days) were analyzed. In n=1221 (72%) patients, pulmonary hypertension (PH) was diagnosed invasively (postcapillary PH: n=1122 [66%]; precapillary PH: n=99 [6%]). Systolic pulmonary artery pressure (sPAP) was 45.3±15.5 mm Hg by Doppler echocardiography and 47.4±16.4 mm Hg by right heart catheterization. Pearson's correlation coefficient was r=0.87 (P<0.0001). Mean right atrial pressure (RAP) was 12.0±5.7 mm Hg by right heart catheterization and was estimated to be 12.1±6.6 mm Hg by echocardiography (r=0.82, P<0.0001). Bland-Altman analysis showed a bias of -2.0 mm Hg for sPAP (95% limits of agreement -18.1 to +14.1 mm Hg) and +1.0 mm Hg for RAP (95% limits of agreement +0.1 to +1.9 mm Hg). Noninvasive diagnosis of pulmonary hypertension with Doppler echocardiography had a good sensitivity (87%) and specificity (79%), positive and negative predictive values (91% and 70%), as well as accuracy (85%) for a sPAP cut-off value of 36 mm Hg (AUC 0.91, P<0.001, CI 0.90 to 0.93). CONCLUSIONS: In this study, Doppler echocardiography proved to be a reliable method for the assessment of sPAP, being well suited to establish the noninvasive diagnosis of pulmonary hypertension in patients with cardiac diseases.

Assessment of left ventricular volumes with echocardiography and cardiac magnetic resonance imaging: real-life evaluation of standard versus new semiautomatic methods.

BACKGROUND: Routine quantitative assessment of left ventricular (LV) volumes with echocardiography is hindered by time-consuming methods requiring a manual trace of the LV cavity from two apical two-dimensional planes. Thus, the aim of this study was to evaluate faster new semiautomatic echocardiographic methods that could represent a feasible alternative for the assessment of LV volumes and ejection fraction (EF) in clinical practice. METHODS: Two semiautomatic methods, the automated EF (Auto-EF) for two-dimensional echocardiography and the 4D Auto LVQ tool for three-dimensional echocardiography (3DE), were compared with the biplane modified Simpson's method and cardiac magnetic resonance (CMR) imaging in 47 patients. To evaluate the accuracy of volumetry, additional in vitro measurements using water-filled latex balloons were performed with both modalities. RESULTS: Results of balloon volumetry by echocardiography and CMR measurements were in good agreement with real balloon volumes. The mean LV EF was 45 ± 11% by Auto-EF, 45 ± 11% by 3DE, 48 ± 11% by Simpson's method, and 54 ± 12% by CMR. Linear regression and Bland-Altman analyses showed good associations for semiautomatic methods with Simpson's method (Auto-EF, r = 0.85, bias = 3%, limits of agreement [LOA] = 12%; 3DE, r = 0.79, bias = 3%, LOA = 14%), as well as with CMR (Auto-EF, r = 0.74, bias = 9%, LOA = 17%; 3DE, r = 0.73, bias = 9%, LOA = 17%). Intra- and interobserver variability were 6% and 12% with Auto-EF and 8% and 11% with 3DE, respectively. CONCLUSIONS: Good correlations between semiautomatic echocardiographic parameters for assessment of LV volumes and EF could be observed when compared with Simpson's method or CMR. However, intertechnique agreement analysis of absolute LV volumes revealed considerable differences, with significant underestimation of volumes and EF with respect to CMR.

Reliability of diffraction enhanced imaging for assessment of cartilage lesions, ex vivo.

OBJECTIVE: The assessment of articular cartilage integrity is of value for the detection of early degenerative joint disease in both the clinical and the research settings. It was the purpose of this study to determine the accuracy and reliability of identifying articular cartilage defects through Diffraction Enhanced Imaging (DEI), a high contrast radiographic imaging technique. DEI provides two new sources of image contrast to radiography: refraction and scatter rejection, besides the absorption of conventional radiography. DESIGN: Cadaveric tali were DEI imaged in the anterior-posterior position at the National Synchrotron Light Source. Two independent observers provided gross score evaluations (on a five point scale) of the trochlear surfaces. The DEI image of each trochlear surface was then graded (on a five point scale) by two additional independent observers who were blinded with regard to the gross evaluation of the articular surfaces. Inter-observer agreement for DEI grades was assessed with the weighted kappa statistic. Correlation of diffraction enhanced image score to the gross score was assessed with Spearman correlation coefficient. RESULTS: The defects of articular cartilage of talar trochleae could be visualized through DEI. The Spearman correlation of gross grades with DEI grades on the 165 talar regions for observers 1 and 2 were 0.91 and 0.91, respectively. The overall weighted kappa value for inter-observer agreement was 0.93, thus considered high agreement. CONCLUSIONS: DEI is accurate and reliable for detection of articular cartilage defects ex vivo. Even early stages of degeneration of cartilage can be visualized with this high contrast technique. Future studies will focus on the application of DEI to the identification of such lesions in vivo.