E Computed coronary microvascular resistance: optimisation, validation and application in a landmark clinical trial

Introduction Computed coronary physiology (virtual FFR) is now established in clinical practice. Our group recently developed a model for predicting absolute flow and hyperaemic coronary microvascular resistance (hMVR) using pressure wire and angiographic data. However, the original model could not regionalise flow and therefore lacked accuracy for interrogating clinical data. I aimed to address these deficiencies by: 1) mathematically optimising simulated hMVR; 2) validate against invasive measurements; and 3) investigate the importance of hMVR in a landmark clinical trial (ORBITA). Methods I implemented side branch flow using a scaling law relating flow (Q) to side branch diameter (D) (Q ∝ Dc). Simulated hMVR was then validated against invasive results obtained with a thermodilution catheter (Rayflow). I performed a global sensitivity analysis (GSA) to identify the model parameters most influential on hMVR results. This informed a systematic review and meta-analysis of the optimal flow-diameter scaling exponent, c, in coronary arteries. I implemented the results into an updated hMVR model. Finally, I used the software to compute hMVR in patients recruited to ORBITA and evaluated the relationship with trial outcomes. Results For validation experiments, hMVR was quantified in 21 mild-moderately stenosed coronary arteries from 19 patients. Correlation was strong (r=0.70, p2.39). I computed hMVR for 131 ORBITA patients (66 PCI and 65 placebo). For patients with low (20th centile) hMVR, PCI increased exercise time by 48 seconds versus placebo (95% CI 6 to 92 seconds, probability of significant difference (Pr)=98.5%). Low hMVR also predicted a benefit of PCI for complete freedom from angina (Pr=98.8%), reduced angina frequency (Pr=97.8%) and improved stress echocardiography score (Pr=99.9%). In patients with high (80th centile) hMVR, PCI offered no benefit versus placebo for any outcome. Conclusions I have optimised virtual hMVR computations by refining a long held scaling law for coronary physiology. The model can now accurately predict absolute flow throughout the branching epicardial tree. I have used it to demonstrate the importance of hMVR in predicting ORBITA trial outcomes. For patients with severe single vessel disease, taking optimal medical therapy, microvascular disease may attenuate the functional and symptomatic benefits of PCI. Introduction Four years ago, an angiography-derived computational fluid dynamics (CFD) model was described which quantifies absolute coronary flow and hMVR.1 The original model did not consider side-branches, so was unable to regionalise flow throughout the coronary tree. This significance of this was uncertain, but flow should correspond to the same location as distal pressure measurement. This communication first focuses on work implementing inclusion of side branch flow, demonstrating its importance throughout the coronary tree and optimising hMVR predictions. In the second half, I used the CFD model to examine the relationship between hMVR with the placebo-controlled effect of percutaneous coronary intervention (PCI) in stable coronary artery disease. This was a retrospective analysis of ORBITA – the first placebo-controlled trial of PCI as symptomatic treatment for stable angina.2 In the original ORBITA trial, PCI did not significantly improve exercise capacity or anginal symptoms compared with a placebo intervention, an effect which was independent of invasive lesion physiology.3 However, pre-randomisation testing did not assess for coronary microvascular dysfunction (CMD). I used the updated model to compute hMVR and assessed its impact on the effect of PCI on ORBITA endpoints. Methods The CFD technique reconstructs an anatomically representative, single lumen coronary artery and applies pressures measured during fractional flow reserve (FFR) assessment to compute hMVR.1 A vascular scaling law was used to compute the expected side-branch loss (Q) from diameter (D) of the tapering main vessel: 1) The original theoretical law, derived a century ago, considered bifurcations in isolation to derive a flow-diameter exponent (c) of 3.04 but this has been disputed by newer derivations5 and my own work on invasive data.6 Flow was regionalised to local taper, utilising a stenosis filter (figure 1). hMVR results were validated against invasive measurements, taken with continuous infusion thermodilution,7patients with FFR2 Interaction between ORBITA outcomes and hMVR was assessed using regression models within a Bayesian framework. The follow-up value was conditioned on the pre-randomisation value, with non-linearity allowed using a restricted cubic spline with 3 knots (10th, 50th, 90th centiles), and randomisation arm and hMVR. A contrast of a 'typical' patient at the 20th against the 80th centile of hMVR is presented, representing the expected placebo-controlled benefit of PCI for patients with low and high hMVR respectively. This is accompanied by 95% Credible Interval (CrI) and probability of significant benefit favouring PCI versus placebo (Pr). The interactional effect between the randomisation arm and hMVR (Printeraction) was also computed. Results For the validation study, hMVR was computed in 21 arteries, taken from 19 patients. Mean FFR was 0.85±0.06 and median hMVR was 0.35 0.29 – 0.46 mmHg.min/mL. Correlation between computed and measured hMVR was strong (r=0.70, p8 GSA considered 165 million simulation results. For mild-moderate stenoses (virtual FFR 0.75 – 1.00), side branch flow-diameter exponent and reconstruction stenosis severity were the co-dominant factors for flow results. For lower virtual FFR values, stenosis severity became more significant (figure 1). Given the GSA results showed influence of the flow-diameter exponent, but the true value was disputed,4 5 I conducted a systematic review and meta-analysis. From a total of 4524 articles, retrieved from five databases, 18 were suitable for meta-analysis. Studies included data from 1070 unique coronary trees, taken from 372 humans and 112 animals. The pooled flow diameter exponent across both epicardial and transmural arteries was 2.39 (95% CI 2.24 – 2.54, I2 = 99%). Study quality was generally judged as fair or good. There was no evidence of systematic reporting bias (Egger test, P=0.15, figure 2).9 From the 200 patients in ORBITA, 131 (66 from PCI, 65 from placebo) were included. Almost all patients (n=128, 97.7%) had at least one positive ischemia test at randomisation (table 1). Computed median hMVR was 1.38 0.89 – 2.09 mmHg.min/mL. Mean pre-randomisation exercise time was 495±188 seconds. For a patient with low hMVR (0.77 mmHg.min/mL), PCI increased exercise time by 48 seconds versus placebo (95% CrI 6 to 92, Pr=98.5%). For a patient with high hMVR (2.43 mmHg.min/mL), PCI increased exercise time by 16 seconds (95% CrI -29 to 61, Pr=75.2%). There was modest evidence for this interaction (Printeraction=83.1%). For a patient with low hMVR, PCI also decreased (improved) DSE score by 0.83 units (95% CrI -1.42 to -0.30, Pr=99.9%), increased log odds of complete freedom from angina by 1.24 (95% CrI 0.16 to 2.38, Pr=98.8%) and improved angina frequency score by 7.67 points (95% CrI 0.25 to 15.6, Pr=97.8%) versus placebo. These effects all became non-significant for patients with high hMVR (figure 3). There was little evidence of interaction with hMVR for any other ORBITA outcome variable (table 2). Discussion We have developed and validated a computational technique for computing hMVR which considers side-branch flow. I also derived an optimal law of coronary scaling, which suggests the Huo-Kassab law5 most accurately describes bifurcation morphology and has additional implications for bifurcation PCI. Our retrospective analysis of the ORBITA trial suggest a relationship between CMD and response to PCI across multiple outcome measures. The benefits of PCI in those with low hMVR included improved exercise time, a 3.5-times greater likelihood of complete freedom from angina, reduced angina frequency and improved DSE scores. These benefits all became non-significant in patients with high hMVR. Therefore, CMD may be a determinant of otherwise successful PCI performed for relief of anginal symptoms. Statement of Contribution DJT Validation study: Coronary reconstruction, side branch optimisation methodology, simulations, statistical analysis, manuscript writing. Sensitivity analysis: Study conception, coronary reconstruction, statistical analysis, manuscript writing. Meta-analysis: Study conception, literature searching, manuscript selection, statistical analysis, manuscript writing. ORBIAT sub-study: Coronary reconstruction, simulations, statistical analysis, manuscript writing. EY: Simulations and coronary reconstruction verification; TN: simulations; HS: Sensitivity analysis simulations, mathematical methods, and statistical analysis; RG: simulations; XX: simulations; KC, AN, RH: development of the original and side branch CFD methodology; AB: simulation and coronary reconstruction technical support; PT, MvV: clinical data collection; MSS: ORBITA trail investigator and statistical analysis; RAL: ORBITA trial investigator; IH: supervision, simulations, mathematical methods, statistical analysis, CFD support; JG: supervision, study conception, coronary reconstruction verification, development of the original CFD methodology; PM: supervision, study conception, development of original CFD methodology, statistical analysis, coronary reconstruction verification, manuscript writing. References Morris PD, Gosling R, Zwierzak I, Evans H, Aubiniere-Robb L, Czechowicz K, et al. A novel method for measuring absolute coronary blood flow and microvascular resistance in patients with ischaemic heart disease. Cardiovascular Research 2021;117(6):1567–77. Al-Lamee R, Thompson D, Dehbi H-M, Sen S, Tang K, Davies J, et al. Percutaneous coronary intervention in stable angina (ORBITA): a double-blind, randomised controlled trial. The Lancet 2018;391(10115):31–40. Al-Lamee R, Howard JP, Shun-Shin MJ, Thompson D, Dehbi H-M, Sen S, et al. Fractional flow reserve and instantaneous wave-free ratio as predictors of the placebo-controlled response to percutaneous coronary intervention in stable single-vessel coronary artery disease: physiology-stratified analysis of ORBITA. Circulation 2018;138(17):1780–92. Murray CD. The physiological principle of minimum work: I. the vascular system and the cost of blood volume. Proc Natl Acad Sci U S A. 1926;12(3):207–14. Huo Y, Kassab GS. Intraspecific scaling laws of vascular trees. Journal of The Royal Society Interface 2012;9(66):190–200. Taylor DJ, Feher J, Halliday I, Hose DR, Gosling R, Aubiniere-Robb L, et al. Refining our understanding of the flow through coronary artery branches; revisiting Murray's law in human epicardial coronary arteries. Frontiers in Physiology 2022;13:871912. van 't Veer M, Adjedj J, Wijnbergen I, Tóth GG, Rutten MC, Barbato E, et al. Novel monorail infusion catheter for volumetric coronary blood flow measurement in humans: in vitro validation. EuroIntervention 2016;12(6):701–7. Taylor DJ, Feher J, Czechowicz K, Halliday I, Hose D, Gosling R, et al. Validation of a novel numerical model to predict regionalized blood flow in the coronary arteries. European Heart Journal-Digital Health 2023;4(2):81–9. Taylor DJ, Saxton H, Halliday I, Newman T, Hose D, Kassab GS, et al. A systematic review and meta-analysis of Murray's Law in the coronary arterial circulation. American Journal of Physiology-Heart and Circulatory Physiology 2024.