Jeffrey L. Sponsler, Michael Senta, and Anastasia Kendrick-Adey
stroke, expert system, CADASIL, optimization
BACKGROUND: A stroke diagnosis expert system prototype StrokeDx has been developed. StrokeDx has been extended to include knowledge about a genetic stroke syndrome, Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy (CADASIL). A clinical discovery of 2 new CADASIL mutations is reported. Mutation 1 is a G to T transversion at nucleotide position 2077 with an amino acid substitution glycine to cysteine. Mutation 2 is a C to T transversion at nucleotide position 2227 with an amino acid substitution arginine to cysteine. CADASIL is frequently misdiagnosed and software tools to aid diagnosis are required. GOAL: Augment StrokeDx with knowledge (clinical, genetics, examination, MRI analysis) about CADASIL. METHODS: A new CADASIL specific rule set was encoded using the StrokeDx prolog-like system. Test case data were analyzed by StrokeDx. RESULTS: StrokeDx computed diagnostic confidence factor of 1.0 for the CADASIL case and computed much lower factors for other non-CADASIL strokes. CONCLUSIONS: Stroke Dx has demonstrated technical flexibility, encoding of CADASIL-specific rules was efficient, and the resulting expert system is more powerful AI system.
Important Links:
Go Back
