Integrating Multi-Omics Data With EHR for Precision Medicine Using Advanced Artificial Intelligence.

With the recent advancement of novel biomedical technologies such as high-throughput sequencing and wearable devices, multi-modal biomedical data ranging from multi-omics molecular data to real-time continuous bio-signals are generated at an unprecedented speed and scale every day. For the first time, these multi-modal biomedical data are able to make precision medicine close to a reality. However, due to data volume and the complexity, making good use of these multi-modal biomedical data requires major effort. Researchers and clinicians are actively developing artificial intelligence (AI) approaches for data-driven knowledge discovery and causal inference using a variety of biomedical data modalities. These AI-based approaches have demonstrated promising results in various biomedical and healthcare applications. In this review paper, we summarize the state-of-the-art AI models for integrating multi-omics data and electronic health records (EHRs) for precision medicine. We discuss the challenges and opportunities in integrating multi-omics data with EHRs and future directions. We hope this review can inspire future research and developing in integrating multi-omics data with EHRs for precision medicine.

Explainable synthetic image generation to improve risk assessment of rare pediatric heart transplant rejection.

Expert microscopic analysis of cells obtained from frequent heart biopsies is vital for early detection of pediatric heart transplant rejection to prevent heart failure. Detection of this rare condition is prone to low levels of expert agreement due to the difficulty of identifying subtle rejection signs within biopsy samples. The rarity of pediatric heart transplant rejection also means that very few gold-standard images are available for developing machine learning models. To solve this urgent clinical challenge, we developed a deep learning model to automatically quantify rejection risk within digital images of biopsied tissue using an explainable synthetic data augmentation approach. We developed this explainable AI framework to illustrate how our progressive and inspirational generative adversarial network models distinguish between normal tissue images and those containing cellular rejection signs. To quantify biopsy-level rejection risk, we first detect local rejection features using a binary image classifier trained with expert-annotated and synthetic examples. We converted these local predictions into a biopsy-wide rejection score via an interpretable histogram-based approach. Our model significantly improves upon prior works with the same dataset with an area under the receiver operating curve (AUROC) of 98.84% for the local rejection detection task and 95.56% for the biopsy-rejection prediction task. A biopsy-level sensitivity of 83.33% makes our approach suitable for early screening of biopsies to prioritize expert analysis. Our framework provides a solution to rare medical imaging challenges currently limited by small datasets.

Aristolochic acids exposure was not the main cause of liver tumorigenesis in adulthood.

Aristolochic acids (AAs) have long been considered as a potent carcinogen due to its nephrotoxicity. Aristolochic acid I (AAI) reacts with DNA to form covalent aristolactam (AL)-DNA adducts, leading to subsequent A to T transversion mutation, commonly referred as AA mutational signature. Previous research inferred that AAs were widely implicated in liver cancer throughout Asia. In this study, we explored whether AAs exposure was the main cause of liver cancer in the context of HBV infection in mainland China. Totally 1256 liver cancer samples were randomly retrieved from 3 medical centers and a refined bioanalytical method was used to detect AAI-DNA adducts. 5.10% of these samples could be identified as AAI positive exposure. Whole genome sequencing suggested 8.41% of 107 liver cancer patients exhibited the dominant AA mutational signature, indicating a relatively low overall AAI exposure rate. In animal models, long-term administration of AAI barely increased liver tumorigenesis in adult mice, opposite from its tumor-inducing role when subjected to infant mice. Furthermore, AAI induced dose-dependent accumulation of AA-DNA adduct in target organs in adult mice, with the most detected in kidney instead of liver. Taken together, our data indicate that AA exposure was not the major threat of liver cancer in adulthood.

Electrophysiology and Arrhythmogenesis in the Human Right Ventricular Outflow Tract.

BACKGROUND: Right ventricular outflow tract (RVOT) is a common source of ventricular tachycardia, which often requires ablation. However, the mechanisms underlying the RVOT's unique arrhythmia susceptibility remain poorly understood due to lack of detailed electrophysiological and molecular studies of the human RVOT. METHODS: We conducted optical mapping studies in 16 nondiseased donor human RVOT preparations subjected to pharmacologically induced adrenergic and cholinergic stimulation to evaluate susceptibility to arrhythmias and characterize arrhythmia dynamics. RESULTS: We found that under control conditions, RVOT has shorter action potential duration at 80% repolarization relative to the right ventricular apical region. Treatment with isoproterenol (100 nM) shortened action potential duration at 80% repolarization and increased incidence of premature ventricular contractions (P=0.003), whereas acetylcholine (100 µM) stimulation alone had no effect on action potential duration at 80% repolarization or premature ventricular contractions. However, acetylcholine treatment after isoproterenol stimulation reduced the incidence of premature ventricular contractions (P=0.034) and partially reversed action potential duration at 80% repolarization shortening (P=0.029). Immunolabeling of RVOT (n=4) confirmed the presence of cholinergic marker VAChT (vesicular acetylcholine transporter) in the region. Rapid pacing revealed RVOT susceptibility to both concordant and discordant alternans. Investigation into transmural arrhythmia dynamics showed that arrhythmia wave fronts and phase singularities (rotors) were relatively more organized in the endocardium than in the epicardium (P=0.006). Moreover, there was a weak but positive spatiotemporal autocorrelation between epicardial and endocardial arrhythmic wave fronts and rotors. Transcriptome analysis (n=10 hearts) suggests a trend that MAPK (mitogen-activated protein kinase) signaling, calcium signaling, and cGMP-PKG (protein kinase G) signaling are among the pathways that may be enriched in the male RVOT, whereas pathways of neurodegeneration may be enriched in the female RVOT. CONCLUSIONS: Human RVOT electrophysiology is characterized by shorter action potential duration relative to the right ventricular apical region. Cholinergic right ventricular stimulation attenuates the arrhythmogenic effects of adrenergic stimulation, including increase in frequency of premature ventricular contractions and shortening of wavelength. Right ventricular arrhythmia is characterized by positive spatial-temporal autocorrelation between epicardial-endocardial arrhythmic wave fronts and rotors that are relatively more organized in the endocardium.

Performance Evaluation of a Wearable Tattoo Electrode Suitable for High-Resolution Surface Electromyogram Recording.

OBJECTIVE: High-density surface electromyography (HD-sEMG) has been utilized extensively in neuromuscular research. Despite its potential advantages, limitations in electrode design have largely prevented widespread acceptance of the technology. Commercial electrodes have limited spatial fidelity, because of a lack of sharpness of the signal, and variable signal stability. We demonstrate here a novel tattoo electrode that addresses these issues. Our dry HD electrode grid exhibits remarkable deformability which ensures superior conformity with the skin surface, while faithfully recording signals during different levels of muscle contraction. METHOD: We fabricated a 4 cm×3 cm tattoo HD electrode grid on a stretchable electronics membrane for sEMG applications. The grid was placed on the skin overlying the biceps brachii of healthy subjects, and was used to record signals for several hours while tracking different isometric contractions. RESULTS: The sEMG signals were recorded successfully from all 64 electrodes across the grid. These electrodes were able to faithfully record sEMG signals during repeated contractions while maintaining a stable baseline at rest. During voluntary contractions, broad EMG frequency content was preserved, with accurate reproduction of the EMG spectrum across the full signal bandwidth. CONCLUSION: The tattoo grid electrode can potentially be used for recording high-density sEMG from skin overlying major limb muscles. Layout programmability, good signal quality, excellent baseline stability, and easy wearability make this electrode a potentially valuable component of future HD electrode grid applications. SIGNIFICANCE: The tattoo electrode can facilitate high fidelity recording in clinical applications such as tracking the evolution and time-course of challenging neuromuscular degenerative disorders.

Skin-integrated wireless haptic interfaces for virtual and augmented reality.

Traditional technologies for virtual reality (VR) and augmented reality (AR) create human experiences through visual and auditory stimuli that replicate sensations associated with the physical world. The most widespread VR and AR systems use head-mounted displays, accelerometers and loudspeakers as the basis for three-dimensional, computer-generated environments that can exist in isolation or as overlays on actual scenery. In comparison to the eyes and the ears, the skin is a relatively underexplored sensory interface for VR and AR technology that could, nevertheless, greatly enhance experiences at a qualitative level, with direct relevance in areas such as communications, entertainment and medicine1,2. Here we present a wireless, battery-free platform of electronic systems and haptic (that is, touch-based) interfaces capable of softly laminating onto the curved surfaces of the skin to communicate information via spatio-temporally programmable patterns of localized mechanical vibrations. We describe the materials, device structures, power delivery strategies and communication schemes that serve as the foundations for such platforms. The resulting technology creates many opportunities for use where the skin provides an electronically programmable communication and sensory input channel to the body, as demonstrated through applications in social media and personal engagement, prosthetic control and feedback, and gaming and entertainment.

Long-Lived, Transferred Crystalline Silicon Carbide Nanomembranes for Implantable Flexible Electronics.

Implantable electronics are of great interest owing to their capability for real-time and continuous recording of cellular-electrical activity. Nevertheless, as such systems involve direct interfaces with surrounding biofluidic environments, maintaining their long-term sustainable operation, without leakage currents or corrosion, is a daunting challenge. Herein, we present a thin, flexible semiconducting material system that offers attractive attributes in this context. The material consists of crystalline cubic silicon carbide nanomembranes grown on silicon wafers, released and then physically transferred to a final device substrate (e.g., polyimide). The experimental results demonstrate that SiC nanomembranes with thicknesses of 230 nm do not experience the hydrolysis process (i.e., the etching rate is 0 nm/day at 96 °C in phosphate-buffered saline (PBS)). There is no observable water permeability for at least 60 days in PBS at 96 °C and non-Na+ ion diffusion detected at a thickness of 50 nm after being soaked in 1× PBS for 12 days. These properties enable Faradaic interfaces between active electronics and biological tissues, as well as multimodal sensing of temperature, strain, and other properties without the need for additional encapsulating layers. These findings create important opportunities for use of flexible, wide band gap materials as essential components of long-lived neurological and cardiac electrophysiological device interfaces.

Ultrathin, Transferred Layers of Metal Silicide as Faradaic Electrical Interfaces and Biofluid Barriers for Flexible Bioelectronic Implants.

Actively multiplexed, flexible electronic devices represent the most sophisticated forms of technology for high-speed, high-resolution spatiotemporal mapping of electrophysiological activity on the surfaces of the brain, heart, and other organ systems. Materials that simultaneously serve as long-lived, defect-free biofluid barriers and sensitive measurement interfaces are essential for chronically stable, high-performance operation. Recent work demonstrates that conductively coupled electrical interfaces of this type can be achieved based on the use of highly doped monocrystalline silicon electrical " via" structures embedded in insulating nanomembranes of thermally grown silica. A limitation of this approach is that dissolution of the silicon in biofluids limits the system lifetimes to 1-2 years, projected based on accelerated testing. Here, we introduce a construct that extends this time scale by more than a factor of 20 through the replacement of doped silicon with a metal silicide alloy (TiSi2). Systematic investigations and reactive diffusion modeling reveal the details associated with the materials science and biofluid stability of this TiSi2/SiO2 interface. An integration scheme that exploits ultrathin, electronic microcomponents manipulated by the techniques of transfer printing yields high-performance active systems with excellent characteristics. The results form the foundations for flexible, biocompatible electronic implants with chronic stability and Faradaic biointerfaces, suitable for a broad range of applications in biomedical research and human healthcare.

Conductively coupled flexible silicon electronic systems for chronic neural electrophysiology.

Materials and structures that enable long-term, intimate coupling of flexible electronic devices to biological systems are critically important to the development of advanced biomedical implants for biological research and for clinical medicine. By comparison with simple interfaces based on arrays of passive electrodes, the active electronics in such systems provide powerful and sometimes essential levels of functionality; they also demand long-lived, perfect biofluid barriers to prevent corrosive degradation of the active materials and electrical damage to the adjacent tissues. Recent reports describe strategies that enable relevant capabilities in flexible electronic systems, but only for capacitively coupled interfaces. Here, we introduce schemes that exploit patterns of highly doped silicon nanomembranes chemically bonded to thin, thermally grown layers of SiO2 as leakage-free, chronically stable, conductively coupled interfaces. The results can naturally support high-performance, flexible silicon electronic systems capable of amplified sensing and active matrix multiplexing in biopotential recording and in stimulation via Faradaic charge injection. Systematic in vitro studies highlight key considerations in the materials science and the electrical designs for high-fidelity, chronic operation. The results provide a versatile route to biointegrated forms of flexible electronics that can incorporate the most advanced silicon device technologies with broad applications in electrical interfaces to the brain and to other organ systems.