WEB MARKER: a Java-based decision-support system
for differential diagnosis of hematologic neoplasms using
immunophenotyping data obtained with flow cytometry.
Andy Nguyen,M.D./ UT-Medical School at Houston, Pathology/
Last Revision on: 11/4/98
- Abstract:
We describe the implementation of a Java-based application for differential diagnosis of
hematopoietic neoplasms using immunophenotyping by flow cytometry. The current version
of this Java applet includes the knowledge-base for 33 hematopoietic neoplasms and 43
diagnostic immunophenotyping markers. Java, a new object-oriented computing language,
helps facilitate development of this applet, a platform-independent module that can be
implemented on the World Wide Web.
The core materials in Web-Marker were derived from those developed for the differential
diagnostic module of CD-Marker, a Microsoft Access database for personal computer
designed previously by the authors. The search
engine in this program, written in Microsoft Visual Basic for Access, was based on backward-chaining inference. Further details of the components, inference algorithm, and
validation results for CD-Marker had previously been described. Conversion of the materials
from the PC-based version to the Java-based version required a complete rewriting of all the
PC software codes to accommodate the web medium.
As the Web rapidly becomes more accessible to users
around the world, Web-based software may eventually form the core of decision-support
systems in clinical setting. Java-based applications, such as the one described in this paper,
are expected to contribute significantly in this area.
- Development platforms:
Web-Marker was implemented in Java using Symantec Visual Cafe 2.0 (Cupertino,
California). The codes for the graphic user interface (GUI) and the inference algorithm were
compiled into a Java byte-code. This byte-code was embedded into a Web page written in
Hypertext Markup Language (HTML). Web-Marker was installed in a Microsoft Windows
NT 4.0 server running Microsoft Internet Information Server 3.0 in the Department of
Pathology and Laboratory Medicine, University of Texas-Houston, Medical School.
- Publication:
Andy N.D. Nguyen, John D. Milam, Kathy A. Johnson, Euginio I. Banez.
A Java-Based Application for Differential Diagnosis of Hematopoietic Neoplasms Using
Immunophenotyping by Flow Cytometry. Computers in Biology and Medicine, Volume 30, Issue 4,
15 July 2000, pp 225-235