Summary Background When an electrical potential is applied to human tissue, the pattern of the resulting current flow is determined by the shapes, arrangements, and internal structure of the tissue cells. By measurement of the electrical current patterns over a range of frequencies, and use of an inverse modelling procedure, electrical variables describing the tissue structure can be calculated. We used this method to develop a screening technique for the detection of cervical precancers. Methods We used a pencil probe (diameter 5 mm) to measure electrical impedance spectra from eight points on the cervix in 124 women with abnormal cervical smears. Variables that should be sensitive to the expected tissue changes were calculated. These were compared with the colposcopic results. Findings The measured electrical impedance changes were those predicted on the basis of the expected tissue structures. Measurements made on normal squamous tissues were well separated from those made on precancerous tissues. We constructed receiver-operating-characteristic curves, comparing measurements made on normal tissue and that showing cervical intraepithelial neoplasia grade 2/3; the area under the curve was 0·951. These groups of women could be separated with a sensitivity of 0·92 and a specificity of 0·92. Interpretation Characteristics of the electrical impedance spectra of tissues can be explained by changes in cell arrangements (layering) and in the size of the nuclei. This relation opens the way to deriving tissue structure from electrical impedance spectral measurements. We show that this approach can be used to give good separation of normal and precancerous cervical tissues. Introduction Biological tissues have complex electrical impedance, which is a function of frequency, because tissues contain components that have both resistive and charge storage (capacitive) properties. The magnitude of the impedance and its dependence on frequency are a function of tissue composition. There have been both practical 1,2 and theoretical 3 demonstrations that different tissue structures are associated with different frequency bands within an impedance spectrum. At high frequencies (>1 GHz) molecular structure is the determining factor, whereas at low frequencies (<100 Hz) charge accumulation at large membrane interfaces dominates. At frequencies of a few kHz to 1 MHz, sometimes referred to as the  dispersion region, cell structures are the main determinant of tissue impedance. Within the  dispersion region, low-frequency current can be thought of as passing through the extracellular space. Because the current has to pass around the cells, the resistance to flow depends on the cell spacings and how they are arranged. However, at higher frequencies, current can penetrate the cell membranes and hence passes through both intracellular and extracellular spaces. The current will thus be determined by intracellular volume and, we propose, the size of the nucleus. This study aimed to assess the agreement between practical measurements of electrical impedance spectra and the predictions made by taking into account the known cell arrangements in cervical tissue. The major changes in cervical tissue in the precancerous stages are the breaking down of superficial cell layering and increases in the size of cell nuclei. 4 The large amount of published material on the histology of cervical tissue in both the normal and pathological states forms the basis for predicting the tissue impedance spectra to be associated with these states. Cancer of the cervix is the second commonest cancer affecting women in the world and the commonest cause of death from cancer. However, in more developed countries, the disorder is potentially preventable 5 by screening and treatment of the precancerous phase, cervical intraepithelial neoplasia (CIN). All screening programmes to date have used exfoliative cervical cytology. However, this technique is associated with low sensitivity and specificity, and the result of the test routinely takes many weeks, thereby creating undue anxiety. 6 The cost burden to health-care programmes of supporting the infrastructure of a cytology screening programme is immense. A test based on the measurement of an electrical impedance spectrum would give a result to the operator (most likely a primary-care physician or nurse) in the same time as it takes for a smear-test sample to be collected. A negative result could be relayed immediately to the patient. A positive result, on the other hand, would enable more rapid referral to a colposcopy clinic. Methods Impedance measurements were made with a pencil probe of 5·5 mm in diameter, with four gold electrodes (1 mm diameter) mounted flush with the face of the probe and spaced equally on a circle of radius 1·65 mm (figure 1). A current of 10 A peak-to- peak was passed between an adjacent pair of electrodes, and the real part of the resulting potential was measured between the two remaining electrodes. The ratio of the measured potential to the amplitude of the imposed current determines a transfer impedance. Measurements were made at eight frequencies by doubling the frequency in steps between 4·8 kHz and 614 kHz. Measurements were made serially at 67 frames per s, and entered onto a computer. In nearly all cases two separate sets of data (each of 100 measurements recorded over 1·5 s) were recorded in succession to check reproducibility. Only the first sets of measurements in each woman are used for the results presented here. EARLY REPORT 892 THE LANCET • Vol 355 • March 11, 2000 Relation between tissue structure and imposed electrical current flow in cervical neoplasia Brian H Brown, John A Tidy, Karen Boston, Anthony D Blackett, Rod H Smallwood, Frank Sharp Lancet 2000; 355: 892–95 Department of Medical Physics, University of Sheffield, Royal Hallamshire Hospital, Sheffield S10 2JF, UK (Prof B H Brown PhD, K Boston MSc, Prof R H Smallwood PhD) and Department of Obstetrics and Gynaecology, University of Sheffield, Clinical Sciences Centre, Northern General Hospital Trust, Sheffield (J A Tidy MD, A D Blackett PhD, Prof F Sharp MD) Correspondence to: Prof Brian H Brown (e-mail: b.h.brown@sheffield.ac.uk) Early report