The Singapore National Precision Medicine Strategy.

Precision medicine promises to transform healthcare for groups and individuals through early disease detection, refining diagnoses and tailoring treatments. Analysis of large-scale genomic-phenotypic databases is a critical enabler of precision medicine. Although Asia is home to 60% of the world's population, many Asian ancestries are under-represented in existing databases, leading to missed opportunities for new discoveries, particularly for diseases most relevant for these populations. The Singapore National Precision Medicine initiative is a whole-of-government 10-year initiative aiming to generate precision medicine data of up to one million individuals, integrating genomic, lifestyle, health, social and environmental data. Beyond technologies, routine adoption of precision medicine in clinical practice requires social, ethical, legal and regulatory barriers to be addressed. Identifying driver use cases in which precision medicine results in standardized changes to clinical workflows or improvements in population health, coupled with health economic analysis to demonstrate value-based healthcare, is a vital prerequisite for responsible health system adoption.

High-Resolution Digital Phenotypes From Consumer Wearables and Their Applications in Machine Learning of Cardiometabolic Risk Markers: Cohort Study.

BACKGROUND: Consumer-grade wearable devices enable detailed recordings of heart rate and step counts in free-living conditions. Recent studies have shown that summary statistics from these wearable recordings have potential uses for longitudinal monitoring of health and disease states. However, the relationship between higher resolution physiological dynamics from wearables and known markers of health and disease remains largely uncharacterized. OBJECTIVE: We aimed to derive high-resolution digital phenotypes from observational wearable recordings and to examine their associations with modifiable and inherent markers of cardiometabolic disease risk. METHODS: We introduced a principled framework to extract interpretable high-resolution phenotypes from wearable data recorded in free-living conditions. The proposed framework standardizes the handling of data irregularities; encodes contextual information regarding the underlying physiological state at any given time; and generates a set of 66 minimally redundant features across active, sedentary, and sleep states. We applied our approach to a multimodal data set, from the SingHEART study (NCT02791152), which comprises heart rate and step count time series from wearables, clinical screening profiles, and whole genome sequences from 692 healthy volunteers. We used machine learning to model nonlinear relationships between the high-resolution phenotypes on the one hand and clinical or genomic risk markers for blood pressure, lipid, weight and sugar abnormalities on the other. For each risk type, we performed model comparisons based on Brier scores to assess the predictive value of high-resolution features over and beyond typical baselines. We also qualitatively characterized the wearable phenotypes for participants who had actualized clinical events. RESULTS: We found that the high-resolution features have higher predictive value than typical baselines for clinical markers of cardiometabolic disease risk: the best models based on high-resolution features had 17.9% and 7.36% improvement in Brier score over baselines based on age and gender and resting heart rate, respectively (P<.001 in each case). Furthermore, heart rate dynamics from different activity states contain distinct information (maximum absolute correlation coefficient of 0.15). Heart rate dynamics in sedentary states are most predictive of lipid abnormalities and obesity, whereas patterns in active states are most predictive of blood pressure abnormalities (P<.001). Moreover, in comparison with standard measures, higher resolution patterns in wearable heart rate recordings are better able to represent subtle physiological dynamics related to genomic risk for cardiometabolic disease (improvement of 11.9%-22.0% in Brier scores; P<.001). Finally, illustrative case studies reveal connections between these high-resolution phenotypes and actualized clinical events, even for borderline profiles lacking apparent cardiometabolic risk markers. CONCLUSIONS: High-resolution digital phenotypes recorded by consumer wearables in free-living states have the potential to enhance the prediction of cardiometabolic disease risk and could enable more proactive and personalized health management.

Towards More Efficient Data Valuation in Healthcare Federated Learning using Ensembling.

Federated Learning (FL) wherein multiple institutions collaboratively train a machine learning model without sharing data is becoming popular. Participating institutions might not contribute equally - some contribute more data, some better quality data or some more diverse data. To fairly rank the contribution of different institutions, Shapley value (SV) has emerged as the method of choice. Exact SV computation is impossibly expensive, especially when there are hundreds of contributors. Existing SV computation techniques use approximations. However, in healthcare where the number of contributing institutions are likely not of a colossal scale, computing exact SVs is still exorbitantly expensive, but not impossible. For such settings, we propose an efficient SV computation technique called SaFE (Shapley Value for Federated Learning using Ensembling). We empirically show that SaFE computes values that are close to exact SVs, and that it performs better than current SV approximations. This is particularly relevant in medical imaging setting where widespread heterogeneity across institutions is rampant and fast accurate data valuation is required to determine the contribution of each participant in multi-institutional collaborative learning.

Semi-supervised classification of radiology images with NoTeacher: A teacher that is not mean.

Deep learning models achieve strong performance for radiology image classification, but their practical application is bottlenecked by the need for large labeled training datasets. Semi-supervised learning (SSL) approaches leverage small labeled datasets alongside larger unlabeled datasets and offer potential for reducing labeling cost. In this work, we introduce NoTeacher, a novel consistency-based SSL framework which incorporates probabilistic graphical models. Unlike Mean Teacher which maintains a teacher network updated via a temporal ensemble, NoTeacher employs two independent networks, thereby eliminating the need for a teacher network. We demonstrate how NoTeacher can be customized to handle a range of challenges in radiology image classification. Specifically, we describe adaptations for scenarios with 2D and 3D inputs, with uni and multi-label classification, and with class distribution mismatch between labeled and unlabeled portions of the training data. In realistic empirical evaluations on three public benchmark datasets spanning the workhorse modalities of radiology (X-Ray, CT, MRI), we show that NoTeacher achieves over 90-95% of the fully supervised AUROC with less than 5-15% labeling budget. Further, NoTeacher outperforms established SSL methods with minimal hyperparameter tuning, and has implications as a principled and practical option for semi-supervised learning in radiology applications.

Self-Path: Self-Supervision for Classification of Pathology Images With Limited Annotations.

While high-resolution pathology images lend themselves well to 'data hungry' deep learning algorithms, obtaining exhaustive annotations on these images for learning is a major challenge. In this article, we propose a self-supervised convolutional neural network (CNN) framework to leverage unlabeled data for learning generalizable and domain invariant representations in pathology images. Our proposed framework, termed as Self-Path, employs multi-task learning where the main task is tissue classification and pretext tasks are a variety of self-supervised tasks with labels inherent to the input images. We introduce novel pathology-specific self-supervision tasks that leverage contextual, multi-resolution and semantic features in pathology images for semi-supervised learning and domain adaptation. We investigate the effectiveness of Self-Path on 3 different pathology datasets. Our results show that Self-Path with the pathology-specific pretext tasks achieves state-of-the-art performance for semi-supervised learning when small amounts of labeled data are available. Further, we show that Self-Path improves domain adaptation for histopathology image classification when there is no labeled data available for the target domain. This approach can potentially be employed for other applications in computational pathology, where annotation budget is often limited or large amount of unlabeled image data is available.

A Vital Signs Telemonitoring Programme Improves the Dynamic Prediction of Readmission Risk in Patients with Heart Failure.

Heart failure (HF) is a leading cause of hospital readmissions. There is great interest in approaches to efficiently predict emerging HF-readmissions in the community setting. We investigate the possibility of leveraging streaming telemonitored vital signs data alongside readily accessible patient profile information for predicting evolving 30-day HF-related readmission risk. We acquired data within a non-randomized controlled study that enrolled 150 HF patients over a 1-year post-discharge telemonitoring and telesupport programme. Using the sequential data and associated ground truth readmission outcomes, we developed a recurrent neural network model for dynamic risk prediction. The model detects emerging readmissions with sensitivity > 71%, specificity > 75%, AUROC ~80%. We characterize model performance in relation to telesupport based nurse assessments, and demonstrate strong sensitivity improvements. Our approach enables early stratification of high-risk patients and could enable adaptive targeting of care resources for managing patients with the most urgent needs at any given time.

Artificial Intelligence and Radiology in Singapore: Championing a New Age of Augmented Imaging for Unsurpassed Patient Care.

Artificial intelligence (AI) has been positioned as being the most important recent advancement in radiology, if not the most potentially disruptive. Singapore radiologists have been quick to embrace this technology as part of the natural progression of the discipline toward a vision of how clinical medicine, empowered by technology, can achieve our national healthcare objectives of delivering value-based and patient-centric care. In this article, we consider 3 core questions relating to AI in radiology, and review the barriers to the widespread adoption of AI in radiology. We propose solutions and describe a "Centaur" model as a promising avenue for enabling the interfacing between AI and radiologists. Finally, we introduce The Radiological AI, Data Science and Imaging Informatics (RADII) subsection of the Singapore Radiological Society. RADII is an enabling body, which together with key technological and institutional stakeholders, will champion research, development and evaluation of AI for radiology applications.

Sparsity enables estimation of both subcortical and cortical activity from MEG and EEG.

Subcortical structures play a critical role in brain function. However, options for assessing electrophysiological activity in these structures are limited. Electromagnetic fields generated by neuronal activity in subcortical structures can be recorded noninvasively, using magnetoencephalography (MEG) and electroencephalography (EEG). However, these subcortical signals are much weaker than those generated by cortical activity. In addition, we show here that it is difficult to resolve subcortical sources because distributed cortical activity can explain the MEG and EEG patterns generated by deep sources. We then demonstrate that if the cortical activity is spatially sparse, both cortical and subcortical sources can be resolved with M/EEG. Building on this insight, we develop a hierarchical sparse inverse solution for M/EEG. We assess the performance of this algorithm on realistic simulations and auditory evoked response data, and show that thalamic and brainstem sources can be correctly estimated in the presence of cortical activity. Our work provides alternative perspectives and tools for characterizing electrophysiological activity in subcortical structures in the human brain.

Reference-free removal of EEG-fMRI ballistocardiogram artifacts with harmonic regression.

Combining electroencephalogram (EEG) recording and functional magnetic resonance imaging (fMRI) offers the potential for imaging brain activity with high spatial and temporal resolution. This potential remains limited by the significant ballistocardiogram (BCG) artifacts induced in the EEG by cardiac pulsation-related head movement within the magnetic field. We model the BCG artifact using a harmonic basis, pose the artifact removal problem as a local harmonic regression analysis, and develop an efficient maximum likelihood algorithm to estimate and remove BCG artifacts. Our analysis paradigm accounts for time-frequency overlap between the BCG artifacts and neurophysiologic EEG signals, and tracks the spatiotemporal variations in both the artifact and the signal. We evaluate performance on: simulated oscillatory and evoked responses constructed with realistic artifacts; actual anesthesia-induced oscillatory recordings; and actual visual evoked potential recordings. In each case, the local harmonic regression analysis effectively removes the BCG artifacts, and recovers the neurophysiologic EEG signals. We further show that our algorithm outperforms commonly used reference-based and component analysis techniques, particularly in low SNR conditions, the presence of significant time-frequency overlap between the artifact and the signal, and/or large spatiotemporal variations in the BCG. Because our algorithm does not require reference signals and has low computational complexity, it offers a practical tool for removing BCG artifacts from EEG data recorded in combination with fMRI.

Reference-free harmonic regression technique to remove EEG-fMRI ballistocardiogram artifacts.

Obtaining high quality electroencephalogram (EEG) data simultaneously with functional MRI (fMRI) recordings is increasingly relevant for the study of cognitive and clinical brain states - as EEG-fMRI offers uniquely high spatiotemporal resolution imaging of brain activity. However, the utility of this technique is limited by ballistocardiogram (BCG) artifacts induced in the EEG by cardiac pulsation and head movement inside the magnetic field. In this paper, we introduce a novel model-based harmonic regression technique to remove BCG artifacts from EEG recorded in the MR scanner. Our technique uses physically motivated parametric models of the BCG artifact and the true EEG signal, and incorporates maximum likelihood approaches to identify model parameters, estimate and subtract the BCG from corrupted EEG measurements. We show that this method effectively removes BCG artifacts from EEG recorded in the MR scanner, restores simulated oscillatory signatures and enables over 20-fold improvement in SNR in bands of interest. Further, unlike common BCG removal techniques that rely on cardiac or motion reference signals, our approach is reference-free and thus is useful when reference signals are corrupted or difficult to acquire.

Speed adaptation in a powered transtibial prosthesis controlled with a neuromuscular model.

Control schemes for powered ankle-foot prostheses would benefit greatly from a means to make them inherently adaptive to different walking speeds. Towards this goal, one may attempt to emulate the intact human ankle, as it is capable of seamless adaptation. Human locomotion is governed by the interplay among legged dynamics, morphology and neural control including spinal reflexes. It has been suggested that reflexes contribute to the changes in ankle joint dynamics that correspond to walking at different speeds. Here, we use a data-driven muscle-tendon model that produces estimates of the activation, force, length and velocity of the major muscles spanning the ankle to derive local feedback loops that may be critical in the control of those muscles during walking. This purely reflexive approach ignores sources of non-reflexive neural drive and does not necessarily reflect the biological control scheme, yet can still closely reproduce the muscle dynamics estimated from biological data. The resulting neuromuscular model was applied to control a powered ankle-foot prosthesis and tested by an amputee walking at three speeds. The controller produced speed-adaptive behaviour; net ankle work increased with walking speed, highlighting the benefits of applying neuromuscular principles in the control of adaptive prosthetic limbs.

Human leg model predicts ankle muscle-tendon morphology, state, roles and energetics in walking.

A common feature in biological neuromuscular systems is the redundancy in joint actuation. Understanding how these redundancies are resolved in typical joint movements has been a long-standing problem in biomechanics, neuroscience and prosthetics. Many empirical studies have uncovered neural, mechanical and energetic aspects of how humans resolve these degrees of freedom to actuate leg joints for common tasks like walking. However, a unifying theoretical framework that explains the many independent empirical observations and predicts individual muscle and tendon contributions to joint actuation is yet to be established. Here we develop a computational framework to address how the ankle joint actuation problem is resolved by the neuromuscular system in walking. Our framework is founded upon the proposal that a consideration of both neural control and leg muscle-tendon morphology is critical to obtain predictive, mechanistic insight into individual muscle and tendon contributions to joint actuation. We examine kinetic, kinematic and electromyographic data from healthy walking subjects to find that human leg muscle-tendon morphology and neural activations enable a metabolically optimal realization of biological ankle mechanics in walking. This optimal realization (a) corresponds to independent empirical observations of operation and performance of the soleus and gastrocnemius muscles, (b) gives rise to an efficient load-sharing amongst ankle muscle-tendon units and (c) causes soleus and gastrocnemius muscle fibers to take on distinct mechanical roles of force generation and power production at the end of stance phase in walking. The framework outlined here suggests that the dynamical interplay between leg structure and neural control may be key to the high walking economy of humans, and has implications as a means to obtain insight into empirically inaccessible features of individual muscle and tendons in biomechanical tasks.