Date on Master's Thesis/Doctoral Dissertation

5-2018

Document Type

Master's Thesis

Degree Name

M. Eng.

Department

Chemical Engineering

Degree Program

JB Speed School of Engineering

Committee Chair

Berson, R. Eric

Committee Member

Willing, Gerold A.

Committee Member

Pantalos, George M.

Author's Keywords

blood flow modeling; stenosis; arterial renderings; CFD; mean age theory; FFR

Abstract

Over one million invasive coronary angiography procedures are performed annually in patients who experience chest pain or are known to have coronary artery disease. The procedure is carried out to ascertain the degree of arterial blockage (stenosis) that hinders blood flow to the heart. A cardiologist performing the procedure determines the physiological degree of a stenosis by either visual estimation, which is routine practice, or by invasively measuring fractional flow reserve (FFR), which is the current gold standard that has been demonstrated to improve patient outcomes and temper the cost of healthcare. Nevertheless, FFR is performed in only 10–20% of patients because it is invasive, expensive, and requires more radiation exposure.

New computational methods utilizing three-dimensional renderings processed from coronary angiograms can provide an accurate, highly sensitive, non-invasive method to assess stenotic significance without using FFR. While beneficial, this technique requires intensive computer processing power and calculation runtimes on the order of several hours. An approach to reduce computational time involves alike computing of two-dimensional arterial slices cut from the three-dimensional source renderings.

The main objective was to determine if two-dimensional processing can also provide an accurate and highly sensitive method to assess stenotic significance at a fraction of the computational expense. Blood flow was analyzed in five patient cases below and five patient cases above the commonly accepted FFR threshold value for intervention of 0.80. Following the generation of two orthogonal slices from DICOM-derived three-dimensional renderings, pulsing blood flow was simulated with CFD, and multiphase mean age theory was applied to calculate the mean age of red blood cells as a diagnostic metric.

Two-dimensional processing typically exhibited a correlation with FFR only in the geometries of vertically-oriented slices. This was ascribed to the possibility of uncaptured stenotic blood flow characteristics in the limited testing of only two angles of a full arterial segment.Mean ages for the three-dimensional cases were many orders of magnitude higher than those of the corresponding two-dimensional cases. This was attributed to red blood cell collisions and distal recirculatory eddies near a stenosisbeing less expressed in the simplicity of the two-dimensional slices when compared to the complexity of the three-dimensional source renderings.

A mean age threshold for determining stent intervention was estimated for the two-dimensional cases since limited sample size disallowed rigorous statistical analysis. The data suggested an arbitrary value equal to ~2.5. Nine out of ten cases correlated with FFR, with just one false negative diagnosis. In published virtual FFR techniques, false diagnosis typically occurs in 10–13% of the cases.

Computational runtime for two-dimensional cases was less than 2% of the runtime for corresponding three-dimensional cases. Preliminary results indicate two-dimensional processing may efficiently detect and assess stenoses non-invasively, provided that it holds up to rigorous statistical analysis following testing of at least 80–100 more cases, plus several additional slice angles.

Share

COinS