Can I trust my fake data - A comprehensive quality assessment framework for synthetic tabular data in healthcare.

BACKGROUND: Ensuring safe adoption of AI tools in healthcare hinges on access to sufficient data for training, testing and validation. Synthetic data has been suggested in response to privacy concerns and regulatory requirements and can be created by training a generator on real data to produce a dataset with similar statistical properties. Competing metrics with differing taxonomies for quality evaluation have been proposed, resulting in a complex landscape. Optimising quality entails balancing considerations that make the data fit for use, yet relevant dimensions are left out of existing frameworks. METHOD: We performed a comprehensive literature review on the use of quality evaluation metrics on synthetic data within the scope of synthetic tabular healthcare data using deep generative methods. Based on this and the collective team experiences, we developed a conceptual framework for quality assurance. The applicability was benchmarked against a practical case from the Dutch National Cancer Registry. CONCLUSION: We present a conceptual framework for quality assuranceof synthetic data for AI applications in healthcare that aligns diverging taxonomies, expands on common quality dimensions to include the dimensions of Fairness and Carbon footprint, and proposes stages necessary to support real-life applications. Building trust in synthetic data by increasing transparency and reducing the safety risk will accelerate the development and uptake of trustworthy AI tools for the benefit of patients. DISCUSSION: Despite the growing emphasis on algorithmic fairness and carbon footprint, these metrics were scarce in the literature review. The overwhelming focus was on statistical similarity using distance metrics while sequential logic detection was scarce. A consensus-backed framework that includes all relevant quality dimensions can provide assurance for safe and responsible real-life applications of synthetic data. As the choice of appropriate metrics are highly context dependent, further research is needed on validation studies to guide metric choices and support the development of technical standards.

How Real-World Data Can Facilitate the Development of Precision Medicine Treatment in Psychiatry.

Precision medicine has the ambition to improve treatment response and clinical outcomes through patient stratification and holds great potential for the treatment of mental disorders. However, several important factors are needed to transform current practice into a precision psychiatry framework. Most important are 1) the generation of accessible large real-world training and test data including genomic data integrated from multiple sources, 2) the development and validation of advanced analytical tools for stratification and prediction, and 3) the development of clinically useful management platforms for patient monitoring that can be integrated into health care systems in real-life settings. This narrative review summarizes strategies for obtaining the key elements-well-powered samples from large biobanks integrated with electronic health records and health registry data using novel artificial intelligence algorithms-to predict outcomes in severe mental disorders and translate these models into clinical management and treatment approaches. Key elements are massive mental health data and novel artificial intelligence algorithms. For the clinical translation of these strategies, we discuss a precision medicine platform for improved management of mental disorders. We use cases to illustrate how precision medicine interventions could be brought into psychiatry to improve the clinical outcomes of mental disorders.

Disruption of Phosphodiesterase 3A Binding to SERCA2 Increases SERCA2 Activity and Reduces Mortality in Mice With Chronic Heart Failure.

BACKGROUND: Increasing SERCA2 (sarco[endo]-plasmic reticulum Ca2+ ATPase 2) activity is suggested to be beneficial in chronic heart failure, but no selective SERCA2-activating drugs are available. PDE3A (phosphodiesterase 3A) is proposed to be present in the SERCA2 interactome and limit SERCA2 activity. Disruption of PDE3A from SERCA2 might thus be a strategy to develop SERCA2 activators. METHODS: Confocal microscopy, 2-color direct stochastic optical reconstruction microscopy, proximity ligation assays, immunoprecipitations, peptide arrays, and surface plasmon resonance were used to investigate colocalization between SERCA2 and PDE3A in cardiomyocytes, map the SERCA2/PDE3A interaction sites, and optimize disruptor peptides that release PDE3A from SERCA2. Functional experiments assessing the effect of PDE3A-binding to SERCA2 were performed in cardiomyocytes and HEK293 vesicles. The effect of SERCA2/PDE3A disruption by the disruptor peptide OptF (optimized peptide F) on cardiac mortality and function was evaluated during 20 weeks in 2 consecutive randomized, blinded, and controlled preclinical trials in a total of 148 mice injected with recombinant adeno-associated virus 9 (rAAV9)-OptF, rAAV9-control (Ctrl), or PBS, before undergoing aortic banding (AB) or sham surgery and subsequent phenotyping with serial echocardiography, cardiac magnetic resonance imaging, histology, and functional and molecular assays. RESULTS: PDE3A colocalized with SERCA2 in human nonfailing, human failing, and rodent myocardium. Amino acids 277-402 of PDE3A bound directly to amino acids 169-216 within the actuator domain of SERCA2. Disruption of PDE3A from SERCA2 increased SERCA2 activity in normal and failing cardiomyocytes. SERCA2/PDE3A disruptor peptides increased SERCA2 activity also in the presence of protein kinase A inhibitors and in phospholamban-deficient mice, and had no effect in mice with cardiomyocyte-specific inactivation of SERCA2. Cotransfection of PDE3A reduced SERCA2 activity in HEK293 vesicles. Treatment with rAAV9-OptF reduced cardiac mortality compared with rAAV9-Ctrl (hazard ratio, 0.26 [95% CI, 0.11 to 0.63]) and PBS (hazard ratio, 0.28 [95% CI, 0.09 to 0.90]) 20 weeks after AB. Mice injected with rAAV9-OptF had improved contractility and no difference in cardiac remodeling compared with rAAV9-Ctrl after aortic banding. CONCLUSIONS: Our results suggest that PDE3A regulates SERCA2 activity through direct binding, independently of the catalytic activity of PDE3A. Targeting the SERCA2/PDE3A interaction prevented cardiac mortality after AB, most likely by improving cardiac contractility.

Intelligent digital tools for screening of brain connectivity and dementia risk estimation in people affected by mild cognitive impairment: the AI-Mind clinical study protocol.

More than 10 million Europeans show signs of mild cognitive impairment (MCI), a transitional stage between normal brain aging and dementia stage memory disorder. The path MCI takes can be divergent; while some maintain stability or even revert to cognitive norms, alarmingly, up to half of the cases progress to dementia within 5 years. Current diagnostic practice lacks the necessary screening tools to identify those at risk of progression. The European patient experience often involves a long journey from the initial signs of MCI to the eventual diagnosis of dementia. The trajectory is far from ideal. Here, we introduce the AI-Mind project, a pioneering initiative with an innovative approach to early risk assessment through the implementation of advanced artificial intelligence (AI) on multimodal data. The cutting-edge AI-based tools developed in the project aim not only to accelerate the diagnostic process but also to deliver highly accurate predictions regarding an individual's risk of developing dementia when prevention and intervention may still be possible. AI-Mind is a European Research and Innovation Action (RIA H2020-SC1-BHC-06-2020, No. 964220) financed between 2021 and 2026. First, the AI-Mind Connector identifies dysfunctional brain networks based on high-density magneto- and electroencephalography (M/EEG) recordings. Second, the AI-Mind Predictor predicts dementia risk using data from the Connector, enriched with computerized cognitive tests, genetic and protein biomarkers, as well as sociodemographic and clinical variables. AI-Mind is integrated within a network of major European initiatives, including The Virtual Brain, The Virtual Epileptic Patient, and EBRAINS AISBL service for sensitive data, HealthDataCloud, where big patient data are generated for advancing digital and virtual twin technology development. AI-Mind's innovation lies not only in its early prediction of dementia risk, but it also enables a virtual laboratory scenario for hypothesis-driven personalized intervention research. This article introduces the background of the AI-Mind project and its clinical study protocol, setting the stage for future scientific contributions.

Federated Networks for Distributed Analysis of Health Data.

Access to health data, important for population health planning, basic and clinical research and health industry utilization, remains problematic. Legislation intended to improve access to personal data across national borders has proven to be a double-edged sword, where complexity and implications from misinterpretations have paradoxically resulted in data becoming more siloed. As a result, the potential for development of health specific AI and clinical decision support tools built on real-world data have yet to be fully realized. In this perspective, we propose federated networks as a solution to enable access to diverse data sets and tackle known and emerging health problems. The perspective draws on experience from the World Economic Forum Breaking Barriers to Health Data project, the Personal Health Train and Vantage6 infrastructures, and industry insights. We first define the concept of federated networks in a healthcare context, present the value they can bring to multiple stakeholders, and discuss their establishment, operation and implementation. Challenges of federated networks in healthcare are highlighted, as well as the resulting need for and value of an independent orchestrator for their safe, sustainable and scalable implementation.

Low Concordance Between T-Cell Densities in Matched Primary Tumors and Liver Metastases in Microsatellite Stable Colorectal Cancer.

BACKGROUND: The subtype, density and location of tumor infiltrating T-cells are being explored as prognostic and predictive biomarkers in primary colorectal cancer (pCRC) and colorectal liver metastases (CLM). Very limited data exist comparing findings in pCRC and matched CLM. PATIENTS AND METHODS: Fifty-eight patients with available pCRC and matched CLM (57/58 microsatellite stable) were included in this OSLO-COMET substudy. In immunohistochemically stained sections, total (Ttot), helper (TH), cytotoxic (CTL), and regulatory (Treg) T-cells were manually counted in hotspots from the invasive margin (IM), intratumor (IT), and tumor adjacent regions to determine T-cell densities. RESULTS: A striking accumulation of T-cells was found in IM of both pCRC and CLM with much lower densities in the IT region, exemplified by Ttot of 2838 versus 340 cells/mm2, respectively, in CLM. The correlation at the individual level between T-cell densities in pCRC and corresponding CLM was poor for all regions and T-cell subtypes; for instance, the correlation coefficient (R2) for IM Ttot was 0.07. The IT TH : CTL and Treg : TH ratios were 2.94 and 0.44, respectively, in pCRC, and 1.84 and 0.24, respectively, in CLM. CONCLUSION: The observed accumulation of T-cells in the IM regions of pCRC and CLM with low penetration to the IT regions, combined with high TH : CTL and Treg : TH ratios, point to the presence of an immune suppressive microenvironment. T-cell densities of CLM differed markedly from the matched pCRC, indicating that to evaluate T-cell biomarkers in metastasis, the commonly available pCRC cannot serve as a surrogate for the metastatic tumor.

Neoadjuvant chemotherapy is associated with a transient increase of intratumoral T-cell density in microsatellite stable colorectal liver metastases.

Patients with colorectal liver metastases (CLM) commonly receive neoadjuvant chemotherapy (NACT) prior to surgical resection. NACT may induce immunogenic cell death with subsequent recruitment of T-cells to the tumor microenvironment, which could be exploited by immune checkpoint inhibition (ICI). In theory, this could expand the use of ICI to obtain responses also in microsatellite stable colorectal cancer, but evidence to suggest optimal treatment schedules are lacking. In this study, densities of total-, cytotoxic-, helper- and regulatory T-cells were quantified by immunohistochemistry in resected CLM from 92 patients included in the OSLO-COMET trial (NCT01516710). All but one patient had microsatellite stable tumors (91/92). Associations between T-cell densities and clinicopathological parameters were analyzed. Fluoropyrimidine-based NACT (in most cases with addition of oxaliplatin or irinotecan) was administered to 45 patients completed median 8 weeks prior to surgical resection. No overall association was found between NACT administration and intratumoral T-cell densities. However, within the NACT group, a short time interval (<9.5 weeks) between NACT completion and CLM resection was strongly associated with high intratumoral T-cell densities compared to the long-interval and no NACT groups (medians 491, 236, and 292 cells/mm2, respectively; P < .0001). The results from this study suggest that the observed increase in intratumoral T-cells after NACT administration may be transient. The significance of this finding should be further explored to ensure that optimal treatment schedules are chosen for studies combining cytotoxic chemotherapy and ICI.

Effect of lipid and fatty acid composition of phospholipid vesicles on long-term stability and their response to Staphylococcus aureus and Pseudomonas aeruginosa supernatants.

Phospholipid vesicles have been the focus of attention as potential vehicles for drug delivery, as they are biomimetic, easy to produce, and contain an aqueous compartment which can be used to carry hydrophilic material, such as drugs or dyes. Lipid vesicles used for this purpose present a particular challenge, as they are not especially stable and can rapidly break down and release their contents away from the target area, especially at physiological temperatures/environments. This study aims to investigate optimum methods for vesicle stabilization where the vesicles are employed as part of a system or technology that signals the presence of pathogenic bacteria via the effect of secreted cytolytic virulence factors on a sensor interface. A number of approaches have been investigated and are presented here as a systematic study of the long-term (14 day) stability at 37 °C, and at various pHs. The response of vesicles, both in suspension and within hydrogels, to Staphylococcus aureus (RN 4282) and Pseudomonas aeruginosa (PAO1) whole bacteria, and supernatants from overnight cultures of both (containing secreted proteins but free of cells), was measured via a sensitive encapsulated carboxyfluorescein release assay. The results showed that lipid chain length, cholesterol concentration, and stabilization via photopolymer stable components were critical in achieving stability. Finally, dispersion of the optimum vesicle formulation in hydrogel matrixes was investigated, culminating in the in vivo demonstration of a simple prototype wound dressing.