Assessment of late gadolinium enhancement in hypertrophic cardiomyopathy improves risk stratification based on current guidelines.

BACKGROUND AND AIMS: Identifying patients with hypertrophic cardiomyopathy (HCM) who are candidates for implantable cardioverter defibrillator (ICD) implantation in primary prevention for sudden cardiac death (SCD) is crucial. The aim of this study was to externally validate the 2022 European Society of Cardiology (ESC) model and other guideline-based ICD class of recommendation (ICD-COR) models and explore the utility of late gadolinium enhancement (LGE) in further risk stratification. METHODS: Seven hundred and seventy-four consecutive patients who underwent cardiac magnetic resonance imaging were retrospectively enrolled. RESULTS: Forty-six (5.9%) patients reached the SCD-related endpoint during 7.4 ± 2.5 years of follow-up. Patients suffering from SCD had higher ESC Risk-SCD score (4.3 ± 2.4% vs. 2.8 ± 2.1%, P < .001) and LGE extent (13.7 ± 9.4% vs. 4.9 ± 6.6%, P < .001). Compared with the 2014 ESC model, the 2022 ESC model showed increased area under the curve (.76 vs. .63), sensitivity (76.1% vs. 43.5%), positive predictive value (16.8% vs. 13.6%), and negative predictive value (98.1% vs. 95.9%). The C-statistics for SCD prediction of 2011 American College of Cardiology (ACC)/American Heart Association (AHA), 2014 ESC, 2020 AHA/ACC, and 2022 ESC models were .68, .64, .76 and .78, respectively. Furthermore, in patients without extensive LGE, LGE ≥5% was responsible for seven-fold SCD risk after multivariable adjustment. Whether in ICD-COR II or ICD-COR III, patients with LGE ≥5% and <15% showed significantly worse prognosis than those with LGE <5% (all P < .001). CONCLUSIONS: The 2022 ESC model performed better than the 2014 ESC model with especially improved sensitivity. LGE enabled further risk stratification based on current guidelines.

Gender Disparities in Academic Pain Medicine: A Retrospective, Cross-Sectional Bibliometric Analysis.

Purpose: This study was conducted to characterize the gender disparities within academic pain management departments in the United States, specifically focusing on its relation to research and academic leadership. This will allow for targeted improvements in efforts made to reduce gender gaps within academic pain medicine. Methods: This is a retrospective, cross-sectional analysis study evaluating pain management faculty of various positions at academic institutions across the United States. We utilized publicly available data on faculty positions and sex to analyze research impact, H-index, number of publications and citations through bibliometric and linear regression analysis. Results: Our analysis found that female faculty had significantly less research output to male faculty. The three research measurement indices used in this study including H-index, number of publications, and number of citations were significantly lower in females than in males among associate and full professor faculty ranking. Multivariable analysis did not display any significant disparities of research output at the division director and department chair level. Discussion: As in many areas of medicine, there continues to be a significant gender disparity in academic pain management departments, particularly with regard to leadership positions and research impact within the field. Our study found that female pain physicians had a significantly less research output based on the three variables of H indices, number of publications, and number of citations compared to their male counterparts. This has been shown to have the impact on discrepancies in female faculty ranking. Interestingly, these variables were not significantly different between male and female faculty members of the same level of leadership except for program director. There are various contributory reasons for these disparities, including implicit biases, lack of mentorship, and familial obligations. Addressing some of these factors can help narrow the schism and promote greater gender equality within academic pain management.

Assessing public health and economic loss associated with black carbon exposure using monitoring and MERRA-2 data.

Black carbon (BC) exposure in China continues to be relatively high, prompting researchers to assess BC exposure levels using data from monitoring sites, satellite remote sensing, and models. However, data regarding the application of a combined strategy comprising the analysis of monitoring data and various types of data to simulate BC exposure levels are lacking. Hence, the current study seeks to estimate short- and long-term BC exposure levels by combining national monitoring data with data from the second Modern-Era Retrospective analysis for Research and Applications (MERRA-2). Furthermore, this study attempts to improve the spatio-temporal resolution of BC exposure levels using Bayesian maximum entropy (BME). The BME model performed well in terms of estimating short- (R2 = 0.74 and RMSE = 1.76 µg/m3) and long-term (R2 = 0.76 and RMSE = 1.3 µg/m3) exposure. Premature mortalities and economic losses were also assessed by applying localised concentration-response coefficients simulated in China. A total of 74,500 (95% confidence interval (CI): 23,900-124,500) and 538,400 (95% CI: 495,000-581,300) all-cause premature mortality cases were found to be associated with short- and long-term BC exposure, respectively. Meanwhile, short-term BC exposure was associated with economic losses ranging from 7.5 to 13.2 billion US dollars (USD) (1 USD = 6.36 RMB on January 19, 2022) based on amended human capital (AHC) and willingness to pay (WTP), accounting for 0.06%-0.1% of China's total gross domestic product (GDP) in 2017 (1.2 × 104 billion USD), respectively. The economic losses for long-term exposure varied from 53 to 93.2 billion USD based on AHC and WTP, accounting for 0.4%-0.8% of China's total GDP in 2017, respectively.

Assessment of Left Ventricular Systolic Function by Cardiovascular Magnetic Resonance Compressed Sensing Real-Time Cine Imaging Combined With Area-Length Method in Normal Sinus Rhythm and Atrial Fibrillation.

Background: The most-commonly used multi-slice Simpson's method employed with routine two-dimensional segmented cine images makes it difficult to evaluate left ventricular (LV) volume and function due to endocardial border blurring and beat-to-beat variation during atrial fibrillation (AF) status. Objectives: To assess the feasibility of compressed sensing real-time (CSRT) cine imaging combined with an area-length method for quantification of LV systolic function in normal sinus rhythm (NSR) and AF. Methods: The CSRT cine sequence and routine segmented balanced Steady-State-Free-Precession cine sequence were performed in 71 patients with NSR (n = 36) or AF (n = 35). Image quality and edge sharpness for both sequences were assessed. The LV functional measurements in patients with NSR included end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), ejection fraction (EF), cardiac output (CO), cardiac index (CI), and LV mass (LVM); all were assessed using segmented cine with Simpson's rule in short axis (SegSA_Simpson, as a reference standard) and area-length (AL) method in the two chamber (Seg2CH_AL) or four chamber (Seg4CH_AL) and CSRT cine with AL method in the two chamber (CSRT2CH_AL) or four chamber (CSRT4CH_AL). Finally, the mean, maximum, and minimum values of each LV functional parameter [EDV/ESV/SV/EF/CO/CI/LVM/heart rate (HR)] from 4~5 consecutive heartbeats were measured using CSRT2CH_AL in patients with AF. Results: In patients with NSR, measurements of EDV (p > 0.05), ESV (p > 0.05), SV (p > 0.05), EF (p > 0.05), and LVM (p > 0.05) assessed with CSRT2CH_AL did not differ significantly from those obtained with SegSA_Simpson. In patients with AF, CSRT image quality score (p < 0.001) and edge sharpness (p < 0.001) both were significantly higher than those obtained from segmented cine. The CSRT2CH_AL provided significantly different results among mean, maximum, and minimum values of each LV parameter from 4~5 consecutive heartbeats (all p < 0.001) with strong inter- and intra-observer agreement in AF. Conclusions: The CSRT cine sequence combined with two chamber area-length analysis accurately assessed LV systolic function in NSR. This approach is expected to permit the assessment of multiple parameters in consecutive heartbeats with good inter- and intra-observer reproducibility for beat-to-beat analysis of LV function in AF.

One step further into the blackbox: a pilot study of how to build more confidence around an AI-based decision system of breast nodule assessment in 2D ultrasound.

OBJECTIVES: To investigate how a DL model makes decisions in lesion classification with a newly defined region of evidence (ROE) by incorporating "explainable AI" (xAI) techniques. METHODS: A data set of 785 2D breast ultrasound images acquired from 367 females. The DenseNet-121 was used to classify whether the lesion is benign or malignant. For performance assessment, classification results are evaluated by calculating accuracy, sensitivity, specificity, and receiver operating characteristic for experiments of both coarse and fine regions of interest (ROIs). The area under the curve (AUC) was evaluated, and the true-positive, false-positive, true-negative, and false-negative results with breakdown in high, medium, and low resemblance on test sets were also reported. RESULTS: The two models with coarse and fine ROIs of ultrasound images as input achieve an AUC of 0.899 and 0.869, respectively. The accuracy, sensitivity, and specificity of the model with coarse ROIs are 88.4%, 87.9%, and 89.2%, and with fine ROIs are 86.1%, 87.9%, and 83.8%, respectively. The DL model captures ROE with high resemblance of physicians' consideration as they assess the image. CONCLUSIONS: We have demonstrated the effectiveness of using DenseNet to classify breast lesions with limited quantity of 2D grayscale ultrasound image data. We have also proposed a new ROE-based metric system that can help physicians and patients better understand how AI makes decisions in reading images, which can potentially be integrated as a part of evidence in early screening or triaging of patients undergoing breast ultrasound examinations. KEY POINTS: • The two models with coarse and fine ROIs of ultrasound images as input achieve an AUC of 0.899 and 0.869, respectively. The accuracy, sensitivity, and specificity of the model with coarse ROIs are 88.4%, 87.9%, and 89.2%, and with fine ROIs are 86.1%, 87.9%, and 83.8%, respectively. • The first model with coarse ROIs is slightly better than the second model with fine ROIs according to these evaluation metrics. • The results from coarse ROI and fine ROI are consistent and the peripheral tissue is also an impact factor in breast lesion classification.

Data Engineering for Machine Learning in Women's Imaging and Beyond.

OBJECTIVE. Data engineering is the foundation of effective machine learning model development and research. The accuracy and clinical utility of machine learning models fundamentally depend on the quality of the data used for model development. This article aims to provide radiologists and radiology researchers with an understanding of the core elements of data preparation for machine learning research. We cover key concepts from an engineering perspective, including databases, data integrity, and characteristics of data suitable for machine learning projects, and from a clinical perspective, including the HIPAA, patient consent, avoidance of bias, and ethical concerns related to the potential to magnify health disparities. The focus of this article is women's imaging; nonetheless, the principles described apply to all domains of medical imaging. CONCLUSION. Machine learning research is inherently interdisciplinary: effective collaboration is critical for success. In medical imaging, radiologists possess knowledge essential for data engineers to develop useful datasets for machine learning model development.

CMR assessment and clinical outcomes of hypertrophic cardiomyopathy with or without ventricular remodeling in the end-stage phase.

End-stage phase of hypertrophic cardiomyopathy (ES-HCM) is a recognized part of HCM disease spectrum. Information on cardiac magnetic resonance (CMR) studies for ES-HCM especially for those without ventricular remodeling has been limited. We aimed to evaluate the morpho-functional and tissue features of ES-HCM with or without ventricular remodeling and to explore CMR prognostic value in these patients. We analysed CMR scans of sixty-three ES-HCM patients and divided them into those with ventricular dilatation (D-ES, n = 41) and those with normal ventricular size (N-ES, n = 22). Cox proportional hazards models were used to assess the association between CMR parameters and outcomes. Patients in D-ES showed hypokinetic-dilated HCM phenotype, while patients in N-ES showed hypokinetic-restrictive HCM phenotype. LGE extent was significantly larger in D-ES (34.7% ± 15.4% vs. 22.8% ± 7.7%; P < 0.01). Atrial fibrillation and edema of lower extremity were more common in N-ES (72.7 vs. 29.3% and 54.5 vs. 24.4%, respectively; P < 0.05). Log-rank test found no significant difference between 2 groups in combined end point of cardiovascular events (χ2 = 0.66, P = 0.418). In multivariate analysis, LGE (HR 1.57-1.83 per 10% LGE increase, P < 0.01) and indexed left atrial volume (LAVI) (HR 1.14-1.21 per 20 mL/m2 increase, P < 0.05) remained independently associated with combined end point when adjusted by other risk factors. The CMR features of HCM in end-stage span between two extremes. LGE is more extensive in those with ventricular remodeling and LAVI is larger in those with normal ventricular size. Both LGE and LAVI are significant predictors of poor outcomes.