Analysis identifying minimal governing parameters for clinically accurate in silico fractional flow reserve

Abstract

Background Personalized hemodynamic models can accurately compute fractional flow reserve (FFR) from coronary angiograms and clinical measurements (FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM1"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">baseline</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> ), but obtaining patient-specific data could be challenging and sometimes not feasible. Understanding which measurements need to be patient-tuned vs. patient-generalized would inform models with minimal inputs that could expedite data collection and simulation pipelines. Aims To determine the minimum set of patient-specific inputs to compute FFR using invasive measurement of FFR (FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM2"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">invasive</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> ) as gold standard. Materials and Methods Personalized coronary geometries ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM3"><mml:mi>N</mml:mi><mml:mo>=</mml:mo><mml:mn>50</mml:mn></mml:math> ) were derived from patient coronary angiograms. A computational fluid dynamics framework, FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM4"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">baseline</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> , was parameterized with patient-specific inputs: coronary geometry, stenosis geometry, mean arterial pressure, cardiac output, heart rate, hematocrit, and distal pressure location. FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM5"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">baseline</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> was validated against FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM6"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">invasive</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> and used as the baseline to elucidate the impact of uncertainty on personalized inputs through global uncertainty analysis. FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM7"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">streamlined</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> was created by only incorporating the most sensitive inputs and FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM8"><mml:msub><mml:mi/><mml:mrow><mml:mtext>semi-streamlined</mml:mtext></mml:mrow></mml:msub></mml:math> additionally included patient-specific distal location. Results FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM9"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">baseline</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> was validated against FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM10"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">invasive</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> via correlation ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM11"><mml:mi>r</mml:mi><mml:mo>=</mml:mo><mml:mn>0.714</mml:mn></mml:math> , <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM12"><mml:mi>p</mml:mi><mml:mo>&lt;</mml:mo><mml:mn>0.001</mml:mn></mml:math> ), agreement (mean difference: <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM13"><mml:mn>0.01</mml:mn><mml:mo>±</mml:mo><mml:mn>0.09</mml:mn></mml:math> ), and diagnostic performance (sensitivity: 89.5%, specificity: 93.6%, PPV: 89.5%, NPV: 93.6%, AUC: 0.95). FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM14"><mml:msub><mml:mi/><mml:mrow><mml:mtext>semi-streamlined</mml:mtext></mml:mrow></mml:msub></mml:math> provided identical diagnostic performance with FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM15"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">baseline</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> . Compared to FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM16"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">baseline</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> vs. FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM17"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">invasive</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> , FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM18"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">streamlined</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> vs. FFR <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM19"><mml:msub><mml:mi/><mml:mrow><mml:mrow><mml:mi mathvariant="normal">invasive</mml:mi></mml:mrow></mml:mrow></mml:msub></mml:math> had decreased correlation ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM20"><mml:mi>r</mml:mi><mml:mo>=</mml:mo><mml:mn>0.64</mml:mn></mml:math> , <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM21"><mml:mi>p</mml:mi><mml:mo>&lt;</mml:mo><mml:mn>0.001</mml:mn></mml:math> ), improved agreement (mean difference: <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="IM22"><mml:mn>0.01</mml:mn><mml:mo>±</mml:mo><mml:mn>0.08</mml:mn></mml:math> ), and comparable diagnostic performance (sensitivity: 79.0%, specificity: 90.3%, PPV: 83.3%, NPV: 87.5%, AUC: 0.90). Conclusion Streamlined models could match the diagnostic performance of the baseline with a full gamut of patient-specific measurements. Capturing coronary hemodynamics depended most on accurate geometry reconstruction and cardiac output measurement.