A KNOWLEDGE-BASED SYSTEM FOR THE INTERPRETATION OF FLOW-CYTOMETRY DATA IN LEUKEMIAS AND LYMPHOMAS

Citation
Lw. Diamond et al., A KNOWLEDGE-BASED SYSTEM FOR THE INTERPRETATION OF FLOW-CYTOMETRY DATA IN LEUKEMIAS AND LYMPHOMAS, Cytometry, 17(3), 1994, pp. 266-273
Citations number
24
Categorie Soggetti
Cytology & Histology","Biochemical Research Methods
Journal title
ISSN journal
01964763
Volume
17
Issue
3
Year of publication
1994
Pages
266 - 273
Database
ISI
SICI code
0196-4763(1994)17:3<266:AKSFTI>2.0.ZU;2-E
Abstract
Flow cytometry immunophenotyping and analysis of DNA ploidy and prolif erative activity have become extremely helpful techniques for diagnosi ng and subclassifying hematopoietic cell populations in a modern, full -service hematopathology laboratory. The number of physicians with spe cial training in the interpretation of these studies is limited. A kno wledge-based computer system has been designed tb aid in the interpret ation of immunophenotyping and DNA flow cytometry results in hematopoi etic disease. The system, known as ''Professor Fidelio,'' is a heurist ic classification system that reasons on the basis of defined diagnost ic patterns. In this study, Fidelio was tested as a stand-alone system on 366 specimens from two large tertiary medical centers. Fidelio's i nterpretation was considered to be appropriate in all cases. In 300 of 366 (82%) specimens, the system's interpretation agreed with the diag nosis of record. Many of the disagreements could be traced to errors i n the recording of the original diagnosis and minor differences in dia gnostic criteria between Fidelio's knowledge base and the criteria in use at the medical centers. When used in a stand-alone mode, Fidelio's interpretation was less specific than the diagnosis of record in cert ain lymphoproliferative disorders that require morphologic information for subclassification. Professor Fidelio is one module in a workstati on for the diagnostic hematology laboratory. This workstation is desig ned for interpretive reporting, education, and database functions for clinical research. Clinical and morphologic information are shared bet ween Fidelio and the other modules for peripheral blood analysis, bone marrow morphology, and lymph node interpretation by means of a relati onal database. The system will be useful in hospitals that lack indivi duals specially trained in flow cytometry. The consistency of interpre tation from case to case will also be valuable for education and resea rch. (C) 1994 Wiley-Liss, Inc.