BackgroundPhase contrast (PC) cardiovascular magnetic resonance (CMR) is widely employed for flow quantification, but analysis typically requires time consuming manual segmentation which can require human correction. Advances in machine learning have markedly improved automated processing, but have yet to be applied to PC-CMR. This study tested a novel machine learning model for fully automated analysis of PC-CMR aortic flow.MethodsA machine learning model was designed to track aortic valve borders based on neural network approaches. The model was trained in a derivation cohort encompassing 150 patients who underwent clinical PC-CMR then compared to manual and commercially-available automated segmentation in a prospective validation cohort. Further validation testing was performed in an external cohort acquired from a different site/CMR vendor.ResultsAmong 190 coronary artery disease patients prospectively undergoing CMR on commercial scanners (84% 1.5T, 16% 3T), machine learning segmentation was uniformly successful, requiring no human intervention: Segmentation time was < 0.01 min/case (1.2 min for entire dataset); manual segmentation required 3.96 ± 0.36 min/case (12.5 h for entire dataset). Correlations between machine learning and manual segmentation-derived flow approached unity (r = 0.99, p < 0.001). Machine learning yielded smaller absolute differences with manual segmentation than did commercial automation (1.85 ± 1.80 vs. 3.33 ± 3.18 mL, p < 0.01): Nearly all (98%) of cases differed by ≤5 mL between machine learning and manual methods. Among patients without advanced mitral regurgitation, machine learning correlated well (r = 0.63, p < 0.001) and yielded small differences with cine-CMR stroke volume (∆ 1.3 ± 17.7 mL, p = 0.36). Among advanced mitral regurgitation patients, machine learning yielded lower stroke volume than did volumetric cine-CMR (∆ 12.6 ± 20.9 mL, p = 0.005), further supporting validity of this method. Among the external validation cohort (n = 80) acquired using a different CMR vendor, the algorithm yielded equivalently small differences (∆ 1.39 ± 1.77 mL, p = 0.4) and high correlations (r = 0.99, p < 0.001) with manual segmentation, including similar results in 20 patients with bicuspid or stenotic aortic valve pathology (∆ 1.71 ± 2.25 mL, p = 0.25).ConclusionFully automated machine learning PC-CMR segmentation performs robustly for aortic flow quantification - yielding rapid segmentation, small differences with manual segmentation, and identification of differential forward/left ventricular volumetric stroke volume in context of concomitant mitral regurgitation. Findings support use of machine learning for analysis of large scale CMR datasets.Electronic supplementary materialThe online version of this article (10.1186/s12968-018-0509-0) contains supplementary material, which is available to authorized users.
Objective To assess the feasibility and acceptability of HPV self‐sampling in Arusha region, northern Tanzania, because the ability for women to self‐collect HPV samples can help reduce the number of health facility visits and improve cervical cancer screening coverage rates. Methods We conducted a facility‐ and community‐based cross‐sectional study among 350 women aged 25–55 years in Arumeru district, Arusha region, northern Tanzania. Women were trained to self‐collect an HPV sample, and follow‐up visits were used to provide results after laboratory testing. Data were analyzed using Stata version 15.1 and summarized using mean and standard deviation for numeric variables and frequencies and percentages for categorical variables. Results Among 350 women, 65 (18.6%) ever screened for cervical cancer, all provided self‐collected samples, and 349 (99.4%) would advise their female friends to undergo the same procedure. The prevalence of positive HPV results was 31 (8.9%), of which 26 (83.9%) were further examined. Two women found with lesions were treated following the national guidelines. Conclusion This study has demonstrated that the HPV self‐sampling intervention for cervical cancer screening is a feasible and acceptable intervention, especially in resource‐limited countries like Tanzania. Scaling‐up policies should consider addressing the potential barriers to the uptake of this intervention.
Background Myocardial strain provides a novel means of quantifying subtle alterations in contractile function; incremental utility post‐MI is unknown. Objectives To test longitudinal—quantified by postprocessing routine echo—for assessment of MI size measured by cardiac magnetic resonance (CMR) and conventional methods, and assess regional and global strain (GLS) as markers of LV thrombus. Methods The population comprised of patients with anterior ST‐segment MI who underwent echo and CMR prospectively. Preexisting echoes were retrieved, re‐analyzed for strain, and compared to conventional MI markers as well as CMR‐evidenced MI, function, and thrombus. Results Seventy‐four patients underwent echo and CMR 4 ± 1 weeks post‐MI; 72% had abnormal GLS. CMR‐quantified MI size was 2.5‐fold larger and EF lower among patients with abnormal GLS, paralleling 2.6–3.1 fold differences in Q‐wave size and CPK (all P ≤ .002). GLS correlated with CMR‐quantified MI (r = .66), CPK (r = .52) and Q‐wave area (r = .44; all P ≤ .001): Regional strain was lower in the base, mid, and apical LV among patients with CMR‐defined transmural MI in each territory (P < .05) and correlated with cine‐CMR regional EF (r = .53–.71; P < .001) and echo wall motion (r = .45–.71; P < .001). GLS and apical strain were ~2‐fold lower among patients with LV thrombus (P ≤ .002): Apical strain yielded higher diagnostic performance for thrombus (AUC: 0.83 [0.72–0.93], P = .001) than wall motion (0.73 [0.58–0.88], P = .02), as did global strain (0.78 [0.65–0.90], P = .005) compared to LVEF (0.58 [0.45–0.72], P = .41). Conclusions Echo‐quantified longitudinal strain provides a marker of MI size and improves stratification for post‐MI LV thrombus beyond conventional indices.
Purpose-Ischemic mitral regurgitation (iMR) augments risk for right ventricular dysfunction (RV DYS). Right and left ventricular (LV) function are linked via common coronary perfusion, but data is lacking regarding impact of LV ischemia and infarct transmurality-as well as altered preload and afterload-on RV performance. Methods-In this prospective multimodality imaging study, stress CMR and 3-dimensional echo (3D-echo) were performed concomitantly in patients with iMR. CMR provided a reference for RV DYS (RVEF<50%), as well as LV function/remodeling, ischemia and infarction. Echo was used to test multiple RV performance indices, including linear (TAPSE, S'), strain (GLS), and volumetric (3D-echo) approaches. Results-90 iMR patients were studied; 32% had RV DYS. RV DYS patients had greater iMR, lower LVEF, larger global ischemic burden and inferior infarct size (all p<0.05). Regarding injury pattern, RV DYS was associated with LV inferior ischemia and infarction (both p<0.05); 80% of affected patients had substantial viable myocardium (<50% infarct thickness) in ischemic inferior segments. Regarding RV function, CMR RVEF similarly correlated with 3D-echo and GLS (r=0.81-0.87): GLS yielded high overall performance for CMR-evidenced RV DYS (AUC: 0.94), nearly equivalent to that of 3D-echo (AUC: 0.95). In multivariable regression, GLS was
Background COVID-19 is associated with cardiac dysfunction. This study tested the relative prognostic role of left (LV), right and bi- (BiV) ventricular dysfunction on mortality in a large multicenter cohort of patients during and after acute COVID-19 hospitalization. Methods/Results All hospitalized COVID-19 patients who underwent clinically indicated transthoracic echocardiography within 30 days of admission at four NYC hospitals between March 2020 and January 2021 were studied. Images were re-analyzed by a central core lab blinded to clinical data. Nine hundred patients were studied (28% Hispanic, 16% African-American), and LV, RV and BiV dysfunction were observed in 50%, 38% and 17%, respectively. Within the overall cohort, 194 patients had TTEs prior to COVID-19 diagnosis, among whom LV, RV, BiV dysfunction prevalence increased following acute infection (p<0.001). Cardiac dysfunction was linked to biomarker-evidenced myocardial injury, with higher prevalence of troponin elevation in patients with LV (14%), RV (16%) and BiV (21%) dysfunction compared to those with normal BiV function (8%, all p<0.05). During in- and out-patient follow-up, 290 patients died (32%), among whom 230 died in the hospital and 60 post-discharge. Unadjusted mortality risk was greatest among patients with BiV (41%), followed by RV (39%) and LV dysfunction (37%), compared to patients without dysfunction (27%, all p<0.01). In multivariable analysis, any RV dysfunction, but not LV dysfunction, was independently associated with increased mortality risk (p<0.01). Conclusions LV, RV and BiV function declines during acute COVID-19 infection with each contributing to increased in- and out-patient mortality risk. RV dysfunction independently increases mortality risk.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.