Virtual biopsies in Barrett's esophagus using an impedance probe

Preliminary results of electrical impedance measurements in squamous and columnar epithelia in rat and human tissues are presented. The aim of this work is to show the possibility of differentiating these two types of epithelia in terms of their electrical characteristics. For the measurements, we employed a 1.95-m-long, 3.2-mm-diameter, four-electrode probe designed to be used transendoscopically in the diagnosis of Barrett's esophagus (BE). BE is a condition in which the normal squamous epithelium of the esophagus is replaced by columnar epithelium of the intestinal type. This metaplasia is considered as a premalignant condition that puts patients at a 30-125-fold risk of developing adenocarcinoma of the esophagus. The diagnosis and surveillance of BE involve taking multiple biopsies, an expensive and time-consuming procedure. This study constitutes the first stage in the replacement of tissue biopsy by "virtual biopsies".

Information from time-varying vibrotactile stimuli

Experiments have been carried out to investigate the information transfer available via a single vibrator on the fingertip. In a first experiment, for stimuli with durations 80 to 320 ms, discrimination of a one-octave step change in frequency at the halfway point was investigated. Results were similar for three stimulus types--sinewave, monophasic pulse and tetraphasic pulse--suggesting temporal cues are more important than spectral cues in this task. In a second experiment, subjects were required to perceive changes in a sequence of stimulus elements. A presentation rate of 6.25 elements s-1 was found to give better results than a rate of 12.5 elements s-1. In the former case, the potential information transfer per element was estimated to be approximately 1.0 bits, corresponding to an information transfer rate of around 6 bits s-1. Implications for the design of a tactile aid to lipreading are discussed.

Model for the dielectric properties of human lung tissue against frequency and air content

Electrical impedance tomographic spectroscopy measurements of the lungs are taken from nine normal subjects, in the frequency range 9.6 kHz-1.2 MHz. The results show that resistivity rho'FRC relative to functional residual capacity increases almost linearly with inspiration volume V, with the slope of the curve increasing with frequency f. Resistivity rho'9.6 kHz relative to 9.6 kHz decreases with f. rho'9.6 kHz increases with V, at any given frequency. Curves for rho'9.6 kHz show a roughly linear trend with log10(f). Based on a discussion of the measurement results, a mathematical lung tissue model is designed that involves extra-capillary blood vessels and alveoli, the walls of which consist of blood-filled capillaries, epithelial cells and intercellular liquid. Using this model, the increase in rho'FRC with V is explained by the thinning of alveolar walls with increasing air content. The almost linear shape of curves for rho'9.6 kHz is attributed to four partly overlapping main dispersions caused by extra-capillary blood vessels, epithelial cells, blood and the capillary network.

EITS changes following oleic acid induced lung water

We present the results of using electrical impedance tomographic spectroscopy (EITS) to follow the changes in lung water induced by oleic acid. Measurements were made on three goats before and after the injection of oleic acid. In addition to the EITs measurements, lung water was also measured using a double-indicator technique. Large falls in lung electrical impedance were seen as a result of the increase in lung water but the size of the fall was a function of the frequency at which the measurements were made. These changes have been modelled using the Cole equation. Four-electrode measurements were also made on two extracted porcine lungs and Cole equation modelling carried out following the introduction of saline into the lungs. Results were similar in the two sets of animal experiments.

Cardiac-related changes in lung resistivity as a function of frequency and location obtained from EITS images

ECG-gated electrical impedance tomographic spectroscopy (EITS) measurements of the lungs were taken on seven normal subjects in the frequency range 9.6 kHz to 614.4 kHz. The results show that in late systole the resistivity p' relative to the R-wave (i.e. p' = 1 at the R-wave) decreases consistently within the lung. In addition there arises an increase in p' in early systole towards the periphery of the lung. Frequency behaviour of p' changes with location. At all times after the R-wave, in the centre of the lung p' is higher at higher frequency f whereas in the periphery it is lower at higher f. The principal decrease in p' can be explained by increasing pulmonary blood volume due to cardiac contraction. The early systolic increase is presumably due to venous return to the left atrium locally leading blood output from the right ventricle which is delayed by the windkessel effect. Based on a model taking extracapillary and capillary blood volume increase into account, the change in frequency behaviour of p' is explained by regional variations in extracapillary blood vessel size determining the relative contributions of extracapillary blood volume and capillary blood volume change to p' at a certain frequency.

Reproducibility of electrical impedance tomographic spectroscopy (EITS) parametric images of neonatal lungs

The reproducibility of electrical impedance tomographic spectroscopy (EITS) images of neonatal lungs have been investigated in 11 clinically stable babies. We have used the Sheffield Mark IIIa EITS system. An average inspiration frame was generated from the data frames associated with maximum inspiration. Frequency images were reconstructed from these frames. The frequency images were analysed to locate the pixel with the maximum change in the right lung field. The change was defined as the 614 kHz measurement relative to 9.6 kHz. A 3 x 3 pixel region of interest was centred at this point. The changes in impedance with frequency for this region of interest show good overall reproducibility between electrode applications for eight frequencies (95% limits of agreement +/- 28%). This reproducibility is improved (95% limits of agreement +/- 13%) by omitting the highest frequency (1.2 MHz) which is most subject to system noise. The parameters for the Cole model derived from data with the highest frequency omitted are less reproducible between electrode applications (95% limits of agreement, R/S +/- 0.83, fc +/- 81.6, RC +/- 0.52, SC +/- 0.39). We suspect that the parametric model used may have an effect on this. The signals recorded at the highest frequency (1.2 MHz) are a major source of variability. The reproducibility results are improved by omitting this frequency from the analysis.

Modelling of cardiac-related changes in lung resistivity measured with EITS

Resistivity data from 9.6 kHZ to 1.2 MHz were recorded from eight normal subjects using an electrical impedance tomographic spectroscopy (EITS) system and then averaged to a mean cardiac cycle using the ECG gating technique. The Cole-Cole model, that is, extracellular resistance R connected in parallel with intracellular resistance S and membrane capacitance C in series, with a distribution parameter a, was applied to model the frequency characteristics and to produce parametric images. During systole, SC and RC were found to decrease and FR increase. The changes in R/S were not consistent among the subjects. We estimated the peak changes in R, S and C to be -2.5%, -3.3% and -7.6% respectively. The results can be explained by considering the blood vessels as spheres of different sizes with blood inside them. The decrease in R during systole might be caused by the increased blood content in relatively large vessels, whereas that in S by the increased blood volume in relatively small vessels. The capacitance of blood is normally smaller than that of lung tissue, whereas FR blood is higher than that of lung tissue. Hence, as blood content increases, C should decrease and FR increase.

Can electrical impedance tomography be used to detect gastro-oesophageal reflux?

A number of experiments have been carried out to investigate the feasibility of using electrical impedance tomography as an alternative method for identifying gastro-oesophageal reflux. Five subjects have been studied using simultaneous EIT and intraoesophageal pH measurements. A refluxogenic meal was given during the course of the recording to induce reflux in the subjects. Results show that there is some change in conductivity in the stomach region during some of the reflux episodes, but EIT is not able to detect all of the pH changes associated with the reflux. Other large conductivity changes, relating to gastric motility can be detected at times when no reflux is occurring.

Three-dimensional electrical impedance tomography

The electrical resistivity of mammalian tissues varies widely and is correlated with physiological function. Electrical impedance tomography (EIT) can be used to probe such variations in vivo, and offers a non-invasive means of imaging the internal conductivity distribution of the human body. But the computational complexity of EIT has severe practical limitations, and previous work has been restricted to considering image reconstruction as an essentially two-dimensional problem. This simplification can limit significantly the imaging capabilities of EIT, as the electric currents used to determine the conductivity variations will not in general be confined to a two-dimensional plane. A few studies have attempted three-dimensional EIT image reconstruction, but have not yet succeeded in generating images of a quality suitable for clinical applications. Here we report the development of a three-dimensional EIT system with greatly improved imaging capabilities, which combines our 64-electrode data-collection apparatus with customized matrix inversion techniques. Our results demonstrate the practical potential of EIT for clinical applications, such as lung or brain imaging and diagnostic screening.

Parametric modelling for electrical impedance spectroscopy system

Three parametric modelling approaches based on the Cole-Cole model are introduced. Comparison between modelling only the real part and modelling both the real and imaginary parts is carried out by simulations, in which random and systematic noise are considered, respectively. The results of modelling the in vitro data collected from sheep are given to reach the conclusions.

Determination of the relationship between the pH and conductivity of gastric juice

Studies of gastric secretion were carried out on 14 subjects, some of whom had taken acid secretion inhibitors. In vitro studies were performed in an attempt to ascertain the effect of H+ and Na+ ions on conductivity. There is a strong correlation between intragastric pH and conductivity for pH < 2, but none of the gastric samples were isotonic. The measured conductivity of the samples was therefore considerably lower than predicted for isotonic gastric juice.

Identifying oesophageal contents using electrical impedance tomography

Investigations have been carried out using the Sheffield mark II real-time EIT system, to look at changes in conductivity associated with swallowing. A ring of 16 electrodes was placed around the neck of 10 subjects, who then performed swallows with four liquids of different conductivities, ranging from water (sigma = 0.03 mS cm-1) to salty soup (sigma = 35.8 mS cm-1). Results showed that the conductive and non-conductive liquids could be distinguished. Bolus transit times were calculated from region of interest curves, and the average transit time for the 10 subjects was found to be 320 +/- 100 ms.

Multifrequency and parametric EIT images of neonatal lungs

The aims of the study were to investigate the problems involved in making multifrequency EIT measurements on neonates and to compare the images obtained with the results from a group of normal adults. The Sheffield electrical impedance tomographic spectroscopy (EITS) system acquires multifrequency data using a set of eight drive and eight receive electrodes. EITS measurements were made on an inhomogeneous group of 10 neonates admitted to the special care baby unit for observation and feeding. R/S, characteristic frequency, RC and SC parameters were generated using the Cole equation. Comparisons of the parameters were made with data collected from normal adults in another study. We have shown that it is possible to obtain EITS parametric images of neonatal lungs and that there are some differences in Cole parameters between the adult and neonatal groups.

A fast parametric modelling algorithm with the Powell method

This paper presents a model that comprises only two parameters (R/S, fr) and the application of three function minimization algorithms (simplex, Powell and modified Powell) to this model to obtain parametric images. Comparisons among the three algorithms in terms of efficiency and reliability were carried out. It was found that, with proper initialization by taking the shape of the modelled data into consideration, the minimization function can be approximated by a quadratic function near the minimum point, therefore the iteration times can be minimized in the modified Powell method. The results show that with the modified Powell method a substantial reduction of computation time can be achieved in the parametric imaging. This makes it possible to obtain a 16 x 16 parametric image in 1 s.

Measured and expected Cole parameters from electrical impedance tomographic spectroscopy images of the human thorax

Electrical impedance tomographic spectroscopy (EITS) images have been recorded from a group of 12 normal subjects using frequencies from 9.6 kHz to 1.2 MHz. The impedance changes with frequency have been modelled on a pixel by pixel basis to produce parametric images as a means of characterizing tissue. The modelling was based on the Cole equation. The lungs are seen as areas of high characteristic frequency and low time constants SC and RS. The R/S images are much less uniform over the region of the lungs. Values characterizing the lung and cardiac regions are given. The results appear to be consistent with a model for the lungs whereby the model parameters can be related to alveolar structure and composition.

Detecting oesophageal-related changes using electrical impedance tomography

Preliminary work has been carried out using the Sheffield mark II real time EIT system, looking for changes in conductivity which occur in the stomach and oesophagus following a swallow of a small volume of either a conducting or a non-conducting liquid. This has been done using three different configurations: a conventional transverse array placed around the thorax, a rosette array on the abdomen, and a rosette array placed on the back. Results show a significant difference between the two liquids, which can be detected equally well by the three electrode configurations.

A real-time electrical impedance tomography system for clinical use--design and preliminary results

An instrument is described which produces images of the electrical impedance distribution within the body at a rate of 25 frames per second, allowing lung ventilation and lung perfusion to be observed in real time. The instrument makes impedance measurements using an array of 16 electrodes on the surface of the body, and reconstructs the images using a weighted backprojection technique. The design of the data acquisition electronics and the reconstruction and display processor are described. Some preliminary in vitro and in vivo results from the system are presented.

Cardiac and respiratory related electrical impedance changes in the human thorax

Electrical impedance measurements have been made from the human trunk over the frequency range 9.6 kHz to 614 kHz. Measurements have been made from 12 normal subjects and the amplitude of the impedance changes associated with the cardiac and respiratory cycles have been recorded. It was found that the real part of the impedance fell to 64.0% of its low frequency value over the measured range of frequencies and that the changes associated with respiration fell in a similar manner. However, the cardiac related changes fell more rapidly with increasing frequency to 28.2% of the low frequency value. The origin of the measured changes is discussed with a view to understanding why the cardiac related changes fall more rapidly. It is not possible to relate in any simple way the frequency dispersion of a single component to that of the whole trunk. However, the results are consistent with the lungs being the major origin of both the cardiac and respiratory related components. The origin of the cardiac related impedance changes could be the pulsatile volume changes in the pulmonary tree. These could be shunted by nonpulsatile lung tissue that has decreasing impedance at high frequency and thus decreases the relative magnitude of the cardiac related changes. This hypothesis needs to be tested using localized measurements from the thorax and 3-D modeling of the trunk.

Detection of changes in intrathoracic fluid in man using electrical impedance tomography

1. The Sheffield electrical impedance tomography system produces information on changes in the distribution of resistivity within tissue. We report on the assessment of electrical impedance tomography in monitoring changes in lung resistivity during a fluid challenge in normal man. 2. Eight normal subjects were studied. Electrical impedance tomography recordings were made at three different lung volumes before, during and after the intravenous infusion of 1 litre of 0.9% NaCl (saline). 3. The mean fall in lung resistivity during the infusion was -22% at total lung capacity (range -10% to -28%), -24% at tidal breathing (-15% to -37%) and -11% at residual volume (-5% to -19%) (P < 0.05 mean pre-infusion resistivity compared with the nadir value after infusion, Wilcoxon). 4. These changes in lung resistivity were probably due to a combination of a fall in haematocrit and an expansion of pulmonary blood volume.

Multi-frequency imaging and modelling of respiratory related electrical impedance changes

Two studies concerning multi-frequency impedance measurements are presented. The first uses tetrapolar measurements made on the thorax and the second electrical impedance tomography images, also made from the thorax. The way in which the impedance and the changes in impedance with ventilation depend upon frequency are investigated using Cole-Cole modelling and also a physiological model of lung tissue. There is an excellent fit to the Cole-Cole model, and the results show that it should be possible to identify tissue on the basis of the impedance spectrum and the spectrum of the changes in impedance during breathing.

Noise equalization within EIT images

We first describe experiments designed to measure the spatial distribution of noise within an electrical impedance tomographic image of a saline-filled tank. These experiments show that the noise increases by a factor of up to 30 from the periphery towards the centre of the image. We then propose a method of noise equalization that depends upon spending longer collecting the smaller potentials than the larger potentials. The method is tested and shown to give an improvement in noise uniformity of about 16 dB.

A comparison of ventilatory and cardiac related changes in EIT images of normal human lungs and of lungs with pulmonary emboli

EIT images have been recorded from the upper thorax of 10 normal subjects and from two patients with pulmonary emboli. The Sheffield Mk2 system was used to obtain the EIT images during quiet tidal breathing and the images were then analysed to extract the cardiac and respiratory related components. In the 10 normal subjects the mean measured change in resistivity during tidal breathing was 9% (SD 3%) with no significant difference in four lung regions. The mean changes during the cardiac cycle were different in the four regions, ranging from -0.9% to -2.6%. The two patients showed very different cardiac related changes from those found in the normals in the posterior lung regions. The sign of the changes was positive, whereas it was negative in the normals. The changes in the anterior lung regions were within the range found in our normal group.

Clinical applications of electrical impedance tomography in the monitoring of changes in intrathoracic fluid volumes

We have previously shown that EIT can detect changes in intrathoracic fluid volumes in normal subjects undergoing a fluid challenge. We now report the preliminary results of two studies designed to investigate the ability of EIT to monitor changes in intrathoracic fluid in disease states. All recordings were made using the Sheffield Mark 1 system with the 16 electrode array placed around the lower thorax. Ten patients were studied during the sequential aspiration of unilateral pleural effusions. Following recording of a baseline dataset during tidal breathing (600 cycles) fluid was aspirated in 300 ml aliquots up to a maximum of 1000 ml, 600 cycles of data being collected after each aspiration. On the side of the effusion there was a progressive increase in intrathoracic resistivity in all patients as the fluid was aspirated. The mean increase in resistivity per 100 ml aspirated was 7%, range 3%-13%, p < 0.01 pre- versus post-aspiration resistivity. In the contralateral lung there was a smaller increase in all subjects (mean 1.8%, range 1%-3%) consistent with a reduction in mediastinal shift following aspiration. In these preliminary observations we have shown that EIT can be used in the clinical environment to detect small changes in intrathoracic fluid.

Vibrotactile and electrotactile perception of time-varying pulse trains

To establish the best strategy for transmitting speech-derived information via a single tactile channel, measurements were made on the perception of frequency- and/or amplitude-modulated pulse-train stimuli, with a comparison of the electrotactile and vibrotactile modalities. In one experiment, vibrotactile perception of 2-oct step changes in stimulus frequency was found to be significantly better than electrotactile on a time-scale appropriate for the transmission of speech features (e.g., with practiced subjects, information transfer of 69% with 200-ms vibrotactile stimuli, 32% with 200-ms electrotactile stimuli). Perception of step changes in stimulus amplitude was similar in the two modalities when changes in amplitude were tailored to match the different dynamic ranges available. In a second experiment, vibrotactile-perception of voice fundamental frequency with various codings was investigated. Both experiments showed information transfer for vibrotactile stimuli to be greater when frequency and amplitude modulation were used together rather than with one or the other in isolation (sentence-stress identification scores: 66% for FM stimuli, 69% for AM stimuli, 80% for FM/AM stimuli). It is concluded that frequency- and amplitude-modulated vibratory stimulation is a good choice in a practical device for the profoundly hearing impaired.