Publications by authors named "A K Ng"

2,304 Publications

  • Page 1 of 1

Contrastive learning of heart and lung sounds for label-efficient diagnosis.

Patterns (N Y) 2022 Jan 7;3(1):100400. Epub 2021 Dec 7.

Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA.

Data labeling is often the limiting step in machine learning because it requires time from trained experts. To address the limitation on labeled data, contrastive learning, among other unsupervised learning methods, leverages unlabeled data to learn representations of data. Here, we propose a contrastive learning framework that utilizes metadata for selecting positive and negative pairs when training on unlabeled data. We demonstrate its application in the healthcare domain on heart and lung sound recordings. The increasing availability of heart and lung sound recordings due to adoption of digital stethoscopes lends itself as an opportunity to demonstrate the application of our contrastive learning method. Compared to contrastive learning with augmentations, the contrastive learning model leveraging metadata for pair selection utilizes clinical information associated with lung and heart sound recordings. This approach uses shared context of the recordings on the patient level using clinical information including age, sex, weight, location of sounds, etc. We show improvement in downstream tasks for diagnosing heart and lung sounds when leveraging patient-specific representations in selecting positive and negative pairs. This study paves the path for medical applications of contrastive learning that leverage clinical information. We have made our code available here: https://github.com/stanfordmlgroup/selfsupervised-lungandheartsounds.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.patter.2021.100400DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8767307PMC
January 2022

Improving triage accuracy through a modified nurse-administered emergency department assessment of chest pain score on patients with chest pain at triage (EDACT): A prospective observational study.

Int Emerg Nurs 2022 Jan 22;61:101130. Epub 2022 Jan 22.

Emergency Medicine Department, National University Hospital, National University Health System, Singapore; Department of Surgery, Yong Loo Lin School of Medicine, National University of Singapore, Singapore.

Background: Chest pain triage in our emergency department (ED) prioritize patients for consultation based on unstructured nursing gestalt. The Emergency Department Assessment of Chest Pain Score (EDACS) identifies patients at low-risk for major adverse cardiac events and may provide standardization for chest pain triage in ED.

Methods: We conducted a prospective observational study, including adult patients with chief complaint of chest pain who were self-conveyed. We aimed to evaluate the overall diagnostic performance of a modified EDACS in triaging these patients.

Results: Data was collected over 6 weeks, with 284 patients included in the final analysis. Nursing gestalt had higher sensitivity (97.6%, 95% confidence interval [CI] 87.4% to 99.9% versus EDACS 45.2%, 95% CI 29.8% to 61.3%), while modified EDACS provided higher specificity (76.4%, 95% CI 70.6% to 81.6%, versus nursing gestalt 29.8%, 95% CI 24.1% to 35.9%). EDACS with electrocardiogram had significantly better area under the receiver operating characteristic curve statistic (0.712; 95% CI 0.631 to 0.793) than EDACS alone (0.608; 95% CI 0.528 to 0.689) and nursing gestalt (0.637; 95% CI 0.600 to 0.674) (p = 0.0324).

Conclusions: Further studies should explore if modified EDACS, together with nursing gestalt, can improve triage accuracy for ED patients with chest pain.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ienj.2021.101130DOI Listing
January 2022

Longitudinal Study of SNCA Rep1 Polymorphism on Executive Function in Early Parkinson's Disease.

J Parkinsons Dis 2022 Jan 20. Epub 2022 Jan 20.

Department of Neurology, National Neuroscience Institute, Tan Tock Seng Hospital, Jalan Tan Tock Seng, Singapore.

The alpha-synuclein gene promoter (SNCA-Rep1) is associated with Parkinson's disease (PD), but its relationship with performance across individual cognitive domains in early PD is unknown. This study aims to investigate Rep1 polymorphism and longitudinal change in cognition in early PD. In this longitudinal study, Rep1 allele lengths ("long" and "short") were determined in 204 early PD patients. All participants underwent annual neuropsychological assessments and followed up for 3 years. Linear-mixed model was performed to investigate the association of Rep1 status and longitudinal change in individual cognitive domains. At 3 years, significant decline in executive function was observed in long Rep1 allele carriers vs short allele carriers, controlling for potential confounders. This is the first longitudinal study demonstrating that long Rep1 allele carriers are at higher risk for executive dysfunction in early PD.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.3233/JPD-213029DOI Listing
January 2022

Multicenter Study on the Genetics of Glomerular Diseases among Southeast and South Asians: The DRAGoN Study.

Clin Genet 2022 Jan 22. Epub 2022 Jan 22.

Molecular Diagnosis Centre, National University Hospital, Singapore.

Multinational studies have reported monogenic etiologies in 25-30% of children with steroid-resistant nephrotic syndrome. Such large studies are lacking in Asia. We established DRAGoN (Deciphering Diversities: Renal Asian Genetics Network) and aimed to describe the genetic and clinical spectrums in Asians. We prospectively studied a cohort of 183 probands with suspected genetic glomerulopathies from South and Southeast Asia, of whom 17% had positive family history. Using multi-gene panel sequencing, we detected pathogenic variants in 26 (14%) probands, of whom one-third had COL4A4 or COL4A5 variants (n = 9, 5%). Of those with COL4A5 defects, only 25% had features suggestive of Alport syndrome. Besides traditional predictors for genetic disease (positive family history and extrarenal malformations), we identified novel predictors, namely older age (6.2 versus 2.4 years; p = 0.001), hematuria (OR 5.6; 95%CI 2.1-14.8; p < 0.001) and proteinuria in the absence of nephrotic syndrome (OR 4.6; 95%CI 1.8-11.8; p = 0.001) at first manifestation. Among patients who first presented with proteinuria without nephrotic syndrome, the genetic diagnostic rates were > 60% when a second risk factor (positive family history or extrarenal manifestation) co-existed. The genetic spectrum of glomerulopathies appears different in Asia. Collagen IV genes may be included in sequencing panels even when suggestive clinical features are absent. This article is protected by copyright. All rights reserved.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1111/cge.14116DOI Listing
January 2022

Underdiagnosis of hepatitis C in the community: do not forget vulnerable populations.

Authors:
Alexander Ng

Eur J Gastroenterol Hepatol 2021 Dec;33(1S Suppl 1):e1095

Faculty of Laws, University College London, London, UK.

View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1097/MEG.0000000000002211DOI Listing
December 2021
-->