Raimond L. Winslow, PhD Raj and Neera Singh Professor ...


1 Οκτ 2013 (πριν από 4 χρόνια και 7 μήνες)

205 εμφανίσεις

Raimond L. Winslow, PhD

Raj and Neera Singh Professor, Biomedical Engineering

Director, Institute for Computational Medicine

Center for Cardiovascular Bioinformatics and Modeling

The Johns Hopkins University

Theodore P. Abraham, M.D., F.A.C.C.,


Associate Professor of Medicine

The Johns Hopkins University

Stephen J. Granite, MS, MBA

Director of Database and Software Development

Center for Cardiovascular Bioinformatics and Modeling

Institute for Computational Medicine

The Johns Hopk
ins University

Research Interests:

Hypertrophic cardiomyopathy (HCM) is the most common inherited heart
condition. It presents with arrhythmias, heart failure and sudden cardiac death. It is in fact the
most common cause of sudden death in athletes and
those below 30 years of age. A common
finding in subjects with HCM is electrocardiographic (ECG) abnormalities. However, not every
HCM patient has ECG abnormalities and the implications of ECG abnormalities in HCM remain
vague. Dr. Abraham recently consti
tuted a HCM consortium consisting of HCM Centers of
Excellence across the world with a view to bring together a large, well
characterized cohort of
HCM patients. One of the primary objectives of this consortium is to investigate the potential
role of ECG a
bnormalities in predicting risk.

In their work with

the CardioVascular Research Grid (CVRG
; http://www.cvrgrid.org
), Dr.
Winslow and Mr. Granite have
developed tools for the

management, presentation and annotation
of time series data, specifically (ECGs
). ECGs are ubiquitous in medicine, from measurements of
general health to specific diagnoses of cardiovascular disease. However, aggregation and
sharing of ECG annotations is quite difficult due to the absence of agreement in the
ation of these measurements and diagnoses.

Need for NCBO Services:

One approach to addressing the challenges of classification mapping and standardization is the
use of an


ontology to centralize these terms and present them to the community in

rm manner

REST Services address

this problem by providing a
rapid response
method to retrieve a
n ontology

term and return the definition of the selected term

another request

Specific Uses of NCBO Services:

Within the ECGrid Tool
kit, Dr. Winslow, Mr. Granite and other members of the CVRG team
have developed a fully functional web interface that uses web services to store, retrieve, visualize
and annotate, and analyze ECG time
series in an access
controlled manner (Figure 1). Thro
the visualization interface, end users can display time
series data for one or more arrays within
an ECG dataset (i.e., an ECG lead’s data). By clicking on a point in the time
series, the Toolkit
invokes an annotation screen. Within the annotation sc
reen, the CVRG team embedded the
OntologyTree Widget developed by the NCBO, to call the NCBO REST Services and display a
hierarchical rep
resentation of the ECG Ontology (Figure 2). Upon selection of a term, the
Toolkit calls the NCBO REST Services again,
retrieving the definition for the ontology term
selected. More information on the Toolkit itself can be found at the CVRG website.

The availability of the ECGrid Toolkit in the CVRG allows the HCM consortium to process
large volumes of ECGs, gain access

to several automated ECG analysis algorithms and annotate
ECGs and the analysis results with terms from the ECG Ontology. The same work done
manually would not only take several individuals and many years but is likely not feasible in the
current environm
ent. The ECGrid Toolkit will allow us to generate and test new hypotheses
concerning the value of ECG in HCM management.

Figure 1. ECGrid Toolkit Interface Components.

Figure 2. ECG Annotation Example.

Raimond L. Winslow, PhD

Theodore P. Abraham,
M.D., F.A.C.C., F.A.S.E.

Stephen J. Granite, MS, MBA