Structural identifiability analysis of a cardiovascular system model. Issue 5 (May 2016)

• Record Type: Journal Article
• Authors: Pironet, Antoine; Dauby, Pierre C.; Chase, J. Geoffrey; Docherty, Paul D.; Revie, James A.; Desaive, Thomas
• Is Part Of: Medical engineering & physics. Volume 38:Issue 5(2016:May)
• Page Start: 433
• Page End: 441
• Publication Date: 2016-05
• Subjects: Identifiability -- Parameter identification -- Lumped-parameter model -- Physiological model; CVS cardiovascular system; Biomedical engineering -- Periodicals; Biomedical Engineering -- Periodicals; Physics -- Periodicals; Génie biomédical -- Périodiques; Biomedical engineering; Electronic journals; Periodicals; 610.28
• DOI: 10.1016/j.medengphy.2016.02.005
• Languages: English
• ISSNs: 1350-4533

Abstract

Highlights: We show that lumped-parameter models of the cardiovascular system have non-identifiability issues. We present a clinically available output set and demonstrate that the cardiovascular system model of interest is identifiable from this output set. We show that with a supplementary assumption, this output set can be further reduced. This analysis implies that the parameter identification procedure can theoretically be performed. Thus, the model may be considered suitable for use in diagnosis. Graphical abstract: Abstract: The six-chamber cardiovascular system model of Burkhoff and Tyberg has been used in several theoretical and experimental studies. However, this cardiovascular system model (and others derived from it) are not identifiable from any output set. In this work, two such cases of structural non-identifiability are first presented. These cases occur when the model output set only contains a single type of information (pressure or volume). A specific output set is thus chosen, mixing pressure and volume information and containing only a limited number of clinically available measurements. Then, by manipulating the model equations involving these outputs, it is demonstrated that the six-chamber cardiovascular system model is structurally globally identifiable. A further simplification is made, assuming known cardiac valve resistances. Because of the poor practical identifiability of these four parameters, this assumption is usual. Under this hypothesis, the six-chamber cardiovascular system model is structurally identifiable from an even smaller dataset. As a consequence, parameter values computed from limited but well-chosen datasets are theoretically unique. This means that the parameter identification procedure can safely be performed on the model from such a well-chosen dataset. Thus, the model may be considered suitable for use in diagnosis.