Development of an adaptive pulmonary simulator for in vitro analysis of patient populations and patient-specific data

Mock circulatory loops used for testing cardiac assist devices: A review of computational and experimental models

Heart failure is a major health risk, and with limited availability of donor organs, there is an increasing need for developing cardiac assist devices (CADs). Mock circulatory loops (MCL) are an important in-vitro test platform for CAD's performance assessment and optimisation. The MCL is a lumped parameter model constructed out of hydraulic and mechanical components aiming to simulate the native cardiovascular system (CVS) as closely as possible. Further development merged MCLs and numerical circulatory models to improve flexibility and accuracy of the system; commonly known as hybrid MCLs. A total of 128 MCLs were identified in a literature research until 25 September 2020. It was found that the complexity of the MCLs rose over the years, recent MCLs are not only capable of mimicking the healthy and pathological conditions, but also implemented cerebral, renal and coronary circulations and autoregulatory responses. Moreover, the development of anatomical models made flow visualisation studies possible. Mechanical MCLs showed excellent controllability and repeatability, however, often the CVS was overly simplified or lacked autoregulatory responses. In numerical MCLs the CVS is represented with a higher order of lumped parameters compared to mechanical test rigs, however, complex physiological aspects are often simplified. In hybrid MCLs complex physiological aspects are implemented in the hydraulic part of the system, whilst the numerical model represents parts of the CVS that are too difficult to represent by mechanical components per se. This review aims to describe the advances, limitations and future directions of the three types of MCLs.

Length of stay prediction for ICU patients using individualized single classification algorithm

BACKGROUND AND OBJECTIVE

In intensive care units (ICUs), length of stay (LOS) prediction is critical to help doctors and nurses select appropriate treatment options and predict patients' condition. Considering that most hospitals use universal models to predict patients' condition, which cannot meet the individual needs of special ICU patients. Our goal is to create a personalized model for patients to determine the number of hospital stays.

METHODS

In this study, a new combination of just-in-time learning (JITL) and one-class extreme learning machine (one-class ELM) is proposed to predict the number of days a patient stays in hospital. This combination is shortened as one-class JITL-ELM, where JITL is used to search for personalized cases for a new patient and one-class ELM is used to determine whether the patient can be discharged within 10 days.

RESULTS

The experimental results show that the one-class JITL-ELM model has an area under the curve (AUC) index of 0.8510, lift value of 2.1390, precision of 1, and G-mean is 0.7842. Its accuracy, specificity, and sensitivity were found as 0.82, 1, and 0.6150, respectively. Moreover, a novel simple mortality risk level estimation system that can determine the mortality rate of a patient by combining LOS and age is proposed. It has an accuracy rate of 66% and the miss rate of only 6.25%.

CONCLUSIONS

Overall, the one-class JITL-ELM can accurately predict hospitalization days and mortality using early physiological parameters. Moreover, a simple mortality risk level estimation system based on a combination of LOS and age is proposed; the system is simple, highly interpretable, and has strong application value.

Finite state machine implementation for left ventricle modeling and control

Background

Simulation of a left ventricle has become a critical facet of evaluating therapies and operations that interact with cardiac performance. The ability to simulate a wide range of possible conditions, changes in cardiac performance, and production of nuisances at transition points enables evaluation of precision medicine concepts that are designed to function through this spectrum. Ventricle models have historically been based on biomechanical analysis, with model architectures constituted of continuous states and not conducive to deterministic processing. Producing a finite-state machine governance of a left ventricle model would enable a broad range of applications: physiological controller development, experimental left ventricle control, and high throughput simulations of left ventricle function.

Methods

A method for simulating left ventricular pressure-volume control utilizing a preload, afterload, and contractility sensitive computational model is shown. This approach uses a logic-based conditional finite state machine based on the four pressure-volume phases that describe left ventricular function. This was executed with a physical system hydraulic model using MathWorks’ Simulink^® and Stateflow tools.

Results

The approach developed is capable of simulating changes in preload, afterload, and contractility in time based on a patient’s preload analysis. Six pressure–volume loop simulations are presented to include a base-line, preload change only, afterload change only, contractility change only, a clinical control, and heart failure with normal ejection fraction. All simulations produced an error of less than 1 mmHg and 1 mL of the absolute difference between the desired and simulated pressure and volume set points. The acceptable performance of the fixed-timestep architecture in the finite state machine allows for deployment to deterministic systems, such as experimental systems for validation.

Conclusions

The proposed approach allows for personalized data, revealed through an individualized clinical pressure–volume analysis, to be simulated in silico. The computational model architecture enables this control structure to be executed on deterministic systems that govern experimental left ventricles. This provides a mock circulatory system with the ability to investigate the pathophysiology for a specific individual by replicating the exact pressure–volume relationship defined by their left ventricular functionality; as well as perform predictive analysis regarding changes in preload, afterload, and contractility in time.