טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentSharon Harell
SubjectA Computerized Approach for Quantitative
Analysis of the "Myocardial Blush Grade" in
Patients Undergoing Angiography
DepartmentDepartment of Biomedical Engineering
Supervisors Full Professor Beyar Rafael
Professor Porat Moshe
Dr. Luis Gruberg


Abstract

Myocardial Blush Grade is a clinical tool for assessment of tissue-level microvascular perfusion from a coronary angiogram. Still, the blush index is not in use in real-time conditions. Developing a quantitative tool for blush grading is therefore expected. In this research, a user oriented, ready for use at the cath-lab blush system was built and statistically tested. The investigations conducted, reveal the following: 1) Qualitative blush analysis: agreement between experts: 56%, disagreement ±1: 33%   2) Pre/Post PCI analysis: after treatment, maximum blush increases (+37%, p<0.05), time to peak decreases (-21%, p<0.05), Wash-in mean slope increases (+41%, p<0.05) and Wash-out mean slope decreases (-39%, p<0.05) .Experts' qualitative blush grades hardly change (+3%, p=0.5). 3) Parameters by Experts' blush grades analysis: quantitative parameters, when divided into 3 blush groups, indeed have significantly different characteristics from each other. In particular, the combined quantitative blush grade increases for each higher blush grade lowest at poor blush and highest at healthy impaired blush. (Blush1≠Blush2≠Blush3 p<0.005, Blush2&3-Blush1: +61% p<0.005,    Blush2-Blush1: +30% p<0.05, Blush3-Blush2: +69%, p<0.005). The results of the different investigations prove the sensitive and subjective nature of the qualitative analysis and re-enforce the need for the quantitative approach. It is shown that knowing the nature of the change in a given parameter before and after PCI can be used to evaluate the effectiveness of treatment by comparison of their given to their expected behavior. Lastly, the combined blush index is an important step towards building a computerized blush classifier.