System architecture
The EIT system Pulmo EIT-100 (Fourth Military Medical University, FMMU, Xi’an, China) consists of five main parts, which are impedance measurement module, power supply module, PC all-in-one machine, medical cart and accessories. The impedance measurement module is the main part of the system, including the control system unit, communication interface management unit, current source, and voltage measurement unit. The EIT system contains 16 measurement channels. The overall functions and configuration of the system are controlled through a PC. The main board contains two processors. One processor is responsible for data pre-processing and communication. The other one is for current signal generation and voltage measurement. The overall design is illustrated in Fig. 1.
Overall design of the system. PC personal computer, ADC analog–digital converter, PGA programmable gain amplifier, MCU microcontroller unit, In-Amp instrumentation amplifier.
PC and software introduction
The PC with Ethernet port (RJ45) is used as the imaging and control device. The PC exchanges commands and data with the lower computer through the high-speed Ethernet port. Our human–machine interface mainly displays the lung impedance time difference image, and the region of interest can be selected for sub-regional display of the impedance waveform. A screenshot of our software display is shown in Fig. 2.
Screenshot of the software display.
Measurement board main control unit
The function of the measurement main board (PCB) is mainly completed by MCU and FPGA control. The former is responsible for the overall scheduling of the lower computer system. It establishes communication with the PC to receive commands and configuration information from the upper computer. It also controls the FPGA for the operation of the measurement function and transmits the measurement information from the FPGA to the PC for image reconstruction.
Current generator
The demodulation method of PulmoEIT-100 is digital. The current generation unit adopts the driving form of dual current sources, which generate 50–200 kHz single-frequency sine wave current source with a phase difference of 180°. There is a 20 MHz synchronous clock source inside the FPGA to provide synchronous clock signal to other modules. In total it has 16 channels and the frame rate can be up to 50 fps.
Th VCCS module circuit principle is illustrated in Fig. 3. The V/I converter can achieve accurate constant current output with wide bandwidth and high output impedance. The signal output from DAC is buffered and filtered into VCCS module for voltage source to current source conversion. The system adopts three op-amps VCCS circuit (the selected op-amps are AMP03 and TH4032). The working bandwidth of AMP03 is 3 MHz. Its amplification setting resistors are built-in, which can better ensure the accuracy of constant current source.
Illustration of the VCCS module circuit principle.
Voltage acquisition unit
The voltage acquisition unit consists of differential amplifier, bandpass filter and gain amplifier. It converts the differential voltage between a pair of electrodes into a single-ended signal. Further, it carries out band-pass filtering to minimize DC component and high-frequency noise to ensure the accuracy and stability of the acquired signal.
System performance evaluation
The performance of the system current source and voltage measurement unit was tested. For the current source, the frequency stability, current stability and output impedance of the output current signal were examined. For the voltage measurement unit, the signal-to-noise ratio (SNR) and stability were examined. To be specific, the device was connected to resistive network and the resistance between every pair channel was 100 Ω. The injected current was set to 5 mA and 100 kHz. Data were recorded continuously for 1 h. The SNR was calculated according to Eqs. (1) and (2):
$${mathrm{SNR}}_{i}= -20mathrm{lg}left(frac{sqrt{frac{1}{N-1}sum_{n=1}^{N}{left({V}_{n}-overline{V}right)}^{2}}}{overline{V}}right)$$
(1)
$$overline{mathrm{SNR}} = frac{1}{{ch}_{max}} sum_{i=1}^{{ch}_{max}}{SNR}_{i}$$
(2)
where i is the channel number; N is the total number of testing sample (N = 1000 in the present study); Vn is the voltage of the n testing sample; (overline{V}) is the average of V; chmax is the number of total channels. The current frequency was tested with Fluke 8845A 6.5 Digit Precision Multimeter (FLUKE, Everett, US).
Further, the system performance was evaluated on human subjects. The study protocol was approved by the ethics committees of the Fourth Military Medical University (KY20213003-1) and all subjects signed the informed consent form before the experiment. A total of 50 healthy lung volunteers were prospectively examined (male:female 33:17; age, 47 ± 15 years; height, 167 ± 8 cm; weight, 66 ± 9 kg). Subjects were asked to perform repetitive slow vital capacity (SVC) maneuvers5 with a spirometer (HI-101; CHEST M.I., INC., Tokyo, Japan). EIT measurements were performed in the following sequence during each SVC with: (1) Pulmo EIT-100 from FMMU, (2) PulmonVista500 from Dräger Medical, (3) Pulmo EIT-100 and (4) PulmonVista500. The electrode belts from the devices were attached and detached from the subjects’ thorax, ~ 4th intercostal space. The level of the electrode plane was marked to ensure the repeatability of electrode placements for every measurement. Relax stable tidal breathing and functional residual capacity level were observed before the SVC maneuver was conducted. Sampling rate of both devices was set to 50 Hz. GREIT algorithm6 was used for offline analysis of EIT data from both devices to eliminate the influence of reconstruction methods.
Tidal variation images (the difference between end-inspiration and end-expiration) were calculated. Impedance values were normalized to the corresponding SVC in milliliters. The clinically widely EIT parameters were used to evaluate the differences of the measured data7. In brief, the following measures are calculated:
(a-b) linearity, i.e. the ratio between the tidal variation and impedance changes obtained during SVC, as well as the correlation between volume and impedance changes in individual subjects. Linear interpolation was applied for spirometry data to match the number of data points recorded by EIT.
(c-d) global ventilation distribution, indicated by the tidal variation in the right and left lungs, as well as in the ventral and dorsal regions.
(e–f) spatial ventilation distribution, indicated by the global inhomogeneity (GI) index8,9 and the center of ventilation (CoV)9,10,11.
(g) temporal ventilation distribution, indicated by standard deviation of regional ventilation delay (RVD)12,13.
Statistical analysis
Two-sample equivalence test was conducted used to compare the differences of EIT measures between two devices. Bland–Altman plots were used to illustrate the differences. Differences between two measurements from the same devices were divided by the average to show the repeatability of the maneuver. A p value < 0.05 was considered statistically significant. EIT data and statistical analysis were performed using MATLAB R2015a (the MathWorks Inc., Natick, MA). Since no information regarding the mean and standard deviation of the EIT measures in healthy volunteers was available, sample size was predefined. The post-hoc power was calculated based on the global ventilation distribution indicated by the tidal variation in the right and left lungs.
Ethics approval
All methods of this study were carried out in accordance with the relevant guidelines and regulations of the Fourth Military Medical University. The study was approved by the Ethics Committee of the First Affiliated Hospital of the Forth Military Medical University (No. KY20213003-1).
