A Step Closer to Acute Care: MEI Lab Presents Research on Cardiovascular Bio Digital Twin

Scientists from the NTT Research Medical & Health Informatics (MEI) Lab delivered two posters and two presentations at a leading biomedical engineering conference this summer. The work presented at the 45th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), held July 24–27, 2023, in Sydney Australia, provided an exciting glimpse of the MEI Lab’s work-in-progress. 

Described as “an essential destination for global biomedical engineers,” this year’s EMBC featured sessions spanning a dozen themes, nine separate workshops, 17 exhibitors and sponsors and daily networking functions. The MEI Lab’s four contributions to EMBC 2023 reflected the team’s collaboration with the National Cerebral and Cardiovascular Center (NCVC) and its ongoing work on a cardiovascular bio-digital twin, i.e., a clinical system that can recommend and even implement therapy in an acute care setting. They included research on model parameter assignment, therapeutic drug effects, renal circulation and cardiac contractility:

  • “From Digital Model to Digital Twin: Probabilistic Estimation of Cardiovascular Bio Digital Twin Parameters,” Iris Shelly, Jon Peterson, Giulia Cornali, Jennifer Rohrs, Joe Alexander. In this poster, Shelly, et al., used a statistical technique (Bayesian inference) to infer parameter values from clinical measurements. The results are highly encouraging and illustrate the importance of considering the clinical relevance of available measurements when choosing how to use them for parameter estimation.
  • “Estimation of Renal Arterial Resistances and Glomerular Capillary Pressure via High-Resolution Renal Arterial Impedance,” Yukiko Fukuda, Yasuyuki Kataoka, Toru Kawada, Jon Peterson, Keita Saku, Joe Alexander, Kenji Sunagawa. In this poster, Fukuda, et al., examined the effect of angiotensin-II and telmisartan (a pharmacological blocker) on renal impedance. By fitting their results to a simple 3-element windkessel model of the kidney, they found that these drugs respectively increased and decreased resistance in efferent arterioles. These data points are consistent with previous studies and demonstrate the utility of the windkessel modeling approach.
  • “Analytical Representation of Four-dimensional Hemodynamics for Drug Therapy Simulation in Acute Heart Failure Treatment,” Yasuyuki Kataoka, Yukiko Fukuda, Jon Peterson, Shohei Yokota, Kazunori Uemura, Keita Saku, Joe Alexander, Kenji Sunagawa. In this paper, Kataoka, et al., propose an analytical representation of four clinically important hemodynamic parameters: left atrial pressure, cardiac output, mean arterial pressure and cardiac oxygen consumption. They derive these as functions of cardiovascular parameters (systemic vascular resistance, cardiac contractility, heart rate and stressed blood volume). By creating a database of how different drugs affect these parameters, they can recommend a combination of drugs to stabilize the patient’s hemodynamics while minimizing the energy requirements of the heart.
  • “Inverse ESPVR Estimation with Singularity Avoidance via Constrained EDPVR Parameter Optimization,” Yasuyuki Kataoka, Yukiko Fukuda, Iris Shelly, Jon Peterson, Shohei Yokota, Kazunori Uemura, Keita Saku, Joe Alexander, Kenji Sunagawa. In an acute care situation, it is both technically challenging and clinically desirable to measure the pumping ability of the heart. In this paper, Kataoka, et al., propose a new technique to estimate cardiac contractility in real time. The result is a stable signal of end systolic elastance that can be implemented in the ICU/CCU setting.