Philips (Spain)

Developing reliable healthcare AI models requires training with representative and diverse data. In imbalanced datasets, model performance tends to plateau on the more prevalent classes while remaining low on less common cases. To overcome this limitation, we propose DiffUltra, the first generative AI technique capable of synthesizing realistic Lung Ultrasound (LUS) images with extensive lesion variability. Specifically, we condition the generative AI by the introduced Lesion-anatomy Bank, which captures the lesion's structural and positional properties from real patient data to guide the image synthesis. We demonstrate that DiffUltra improves consolidation detection by 5.6% in AP compared to the models trained solely on real patient data. More importantly, DiffUltra increases data diversity and prevalence of rare cases, leading to a 25% AP improvement in detecting rare instances such as large lung consolidations, which make up only 10% of the dataset.

Abstract BackgroundCardiovascular disease and dementia often co‐exist at advanced stages. Yet, midlife longitudinal studies examining the interplay between atherosclerosis and its risk factors on brain health are scarce. We aimed to determine the longitudinal associations between cerebral glucose metabolism, subclinical atherosclerosis and cardiovascular risk factors in middle‐aged asymptomatic individuals.MethodThe Progression of Early Subclinical Atherosclerosis (PESA) is a longitudinal observational cohort study of 4184 asymptomatic individuals aged 40‐54. Participants with subclinical atherosclerosis underwent longitudinal cerebral [18F]fluorodeoxyglucose (FDG)‐PET (outcomes). Cardiovascular risk was quantified with SCORE2 and subclinical atherosclerosis with 3D vascular ultrasound (exposures). Multivariate regression and linear mixed effects models were used to assess outcome‐exposure associations through region‐of‐interest and voxelwise approaches. Additionally, blood‐based biomarkers of neuropathology were quantified and mediation analyses performed.ResultThis longitudinal study included 370 participants (median age: 49·8 [IQR=46·1‐52·2] years; 84% men; follow‐up: 4·7 [4·2‐5·2] years). Persistent high‐risk of cardiovascular disease was associated with an accelerated decline of cortical FDG uptake compared to low‐risk (ß=‐0·008 [95% CI ‐0·013 to ‐0·002]; pFDR=0·040), with plasma neurofilament light chain, a marker of neurodegeneration, mediating this association by 20% (ß=0·198 [0·008 to 0·740]; pFDR=0·050) (Figure 1). Moreover, progression of subclinical carotid atherosclerosis was associated with an additional decline in FDG uptake in Alzheimer’s disease brain regions, not explained by cardiovascular risk (ß=‐0·269 [95% CI ‐0·509 to ‐0·027]; p=0·029) (Figure 2).ConclusionMiddle‐aged asymptomatic individuals with sustained cardiovascular risk and subclinical carotid atherosclerosis already present brain metabolic decline, suggesting that maintaining cardiovascular health during midlife may be key to reducing neurodegenerative disease burden later in life.

Abstract Background Zaire Ebolavirus disease (EVD) outbreaks can be controlled using rVSV-ZEBOV vaccination and other public health measures. People in high-risk areas may have pre-existing antibodies from asymptomatic Ebolavirus exposure that might affect response to rVSV-ZEBOV. Therefore, we assessed the impact pre-existing immunity had on post-vaccination IgG titre, virus neutralization and reactogenicity following vaccination. Methods In this prospective cohort study, 2,115 consenting close contacts (“proches”) of EVD survivors were recruited. Proches were vaccinated with rVSV-ZEBOV and followed up for 28 days for safety and immunogenicity. Anti-GP IgG titre at baseline and day 28 was assessed by ELISA. Samples from a representative subset were evaluated using live virus neutralization. Results Ten percent were seropositive at baseline. At day, 28 IgG in baseline seronegative (GMT 0.106 IU/mL, 95% CI: 0.100 to 0.113) and seropositive (GMT 0.237 IU/ml, 0.210 to 0.267) participants significantly increased from baseline (both p<0.0001). There was strong correlation between antibody concentration and virus neutralization in day 28 samples (Spearman's rho 0.75). Vaccinees with baseline IgG antibodies against Zaire Ebolavirus had similar safety profiles to those without detectable antibodies (63.6% vs 66.1% adults experienced any adverse event; 49.1% vs 60.9% in children), with almost all adverse events graded as mild. No serious adverse events were attributed to vaccination. No EVD survivors tested positive for Ebolavirus by RT-PCR. Conclusions These data add further evidence of rVSV-ZEBOV safety and immunogenicity, including in people with pre-existing antibodies from suspected natural ZEBOV infection, whichdo not blunt rVSV-ZEBOV immune response. Pre-vaccination serological screening is not required.

Machine Learning (ML) applications in complex domains, such as the medical domain, can be highly beneficial, but also hazardous if some concepts are overlooked. In this context, however, health professionals denote expertise in their domain, but they often lack skills in terms of ML. In this sense, to leverage ML applications in the medical domain, it is important to combine both domain expertise and ML-related skills. In previous works, we tackled this challenge in the health context through a visual platform (KoopaML) that enables lay users to build ML pipelines. The present work describes the challenges derived from the first version of the platform and the prototypes for the new features designed to address them. The prototypes have been validated by two experts, obtaining highly valuable feedback.

Pulmonary hypertension (PH) associated with left heart disease is an increasingly prevalent problem, orphan of targeted therapies, and related to a poor prognosis, particularly when pre- and post-capillary PH combine. The current study aimed to determine whether treatment with the selective β3 adrenoreceptor agonist mirabegron improves outcomes in patients with combined pre- and post-capillary PH (CpcPH). The β3 Adrenergic Agonist Treatment in Chronic Pulmonary Hypertension Secondary to Heart Failure (SPHERE-HF) trial is a multicentre, randomized, parallel, placebo-controlled clinical trial that enrolled stable patients with CpcPH associated with symptomatic heart failure. A total of 80 patients were assigned to receive mirabegron (50 mg daily, titrated till 200 mg daily, n=39) or placebo (n=41) for 16 weeks. Of them, 66 patients successfully completed the study protocol and were valid for the main analysis. The primary endpoint was the change in pulmonary vascular resistance (PVR) on right heart catheterization. Secondary outcomes included the change in right ventricular (RV) ejection fraction by cardiac magnetic resonance or computed tomography, other haemodynamic variables, functional class, and quality of life. The trial was negative for the primary outcome (placebo-corrected mean difference of 0.62 Wood units, 95% confidence interval [CI] -0.38, 1.61, p=0.218). Patients receiving mirabegron presented a significant improvement in RV ejection fraction as compared to placebo (placebo-corrected mean difference of 3.0%, 95% CI 0.4, 5.7%, p=0.026), without significant differences in other pre-specified secondary outcomes. SPHERE-HF is the first clinical trial to assess the potential benefit of β3 adrenergic agonists in PH. The trial was negative since mirabegron did not reduce PVR, the primary endpoint, in patients with CpcPH. On pre-specified secondary outcomes, a significant improvement in RV ejection fraction assessed by advanced cardiac imaging was found, without differences in functional class or quality of life.

Abstract Background Myocarditis after SARS-CoV2 infection or vaccination is rare, but seems to be relatively more frequent in young population. Cardiac magnetic resonance (CMR) T2 weighted sequences have the potential to detect subclinical myocarditis. However, there is paucity of data on the potential myocardial involvement after SARS-CoV2 infection or vaccination in asymptomatic adolescents. Purpose To evaluate the presence of subclinical myocardial damage in adolescents who were infected with SARS-CoV2 or vaccinated against SARS-CoV2 using non-contrast CMR imaging. Methods Asymptomatic adolescents enrolled in the “Early ImaginG Markers of unhealthy lifestyles in Adolescents” (EnIGMA) project were scanned using a 3-Tesla CMR scanner between March 2021 and October 2021. CMR scans included CINE imaging and myocardial T2-mapping sequences. SARS-CoV2 IgG antibody testing was performed in capillary blood samples, and date of confirmed SARS-CoV2 infection and/or vaccination if any was collected. Participants were assigned to three different groups according to SARS-CoV2 status: Group 1 (non-infected and non-vaccinated), Group 2 (infected and non-vaccinated), and Group 3 (vaccinated, independently of past infection status). CMR images were analyzed by experienced observers blinded to adolescent's SARS-CoV2 status. ANOVA and multiple regression analysis, together with correlation coefficients, were used to study between-group differences and associations among variables of interest. Results A total of 115 adolescents with a mean age of 16.0 years (standard deviation (SD)=0.4), 54% girls, completed the CMR study and SARS-CoV2 data successfully, and were assigned to Group 1 (n=72), Group 2 (n=22), and Group 3 (n=21). Left and right ventricular ejection fraction (LVEF/RVEF) did not significantly differ among groups: mean LVEF was 62.8% (SD=4.1), 63.0% (SD=3.7) and 60.9% (SD=3.9) [p=0.12] and mean RVEF was 56.5% (SD=4.2), 56.5% (SD=5.5) and 54.5% (SD=5.1) [p=0.23] in Groups 1, 2 and 3, respectively. Similarly, there were no between-group significant differences in myocardial T2 relaxation values: mean T2 values were 44.1 ms (SD=2.2), 44.1 ms (SD=1.8) and 44.4 ms (SD=1.9) in Groups 1, 2, and 3, respectively (p=0.63) (Figure 1). No differences were found either after adjusting for age and gender. Median time (interquartile range) from date of infection or vaccination to CMR acquisition was 133 (121) days and 28 (38) days in Group 2 and Group 3, respectively. No correlation between time from infection/vaccination to CMR acquisition and T2 values was detected (Figure 2). Conclusions This observational study did not find evidence of subclinical myocardial involvement after SARS-CoV2 infection or vaccination in asymptomatic adolescents, as assessed with T2-mapping magnetic resonance imaging. Funding Acknowledgement Type of funding sources: Foundation. Main funding source(s): Instituto de Salud Carlos III (ISCIII)-Fondo de Investigaciόn Sanitaria.European Regional Development Fund/European Social Fund (“A way to make Europe”/“Investing in your future”.