THE LANCET • Vol 352 • August 1, 1998 359 biological variability of these electrical phenomena will be important in the improvement of this test. Studies in populations with a lower cancer prevalence are needed to assess further the diagnostic value of this approach. Lancet 1998; 351: 359–63 Introduction The diagnosis of breast cancer usually begins with the discovery of a breast mass, either by the patient or her physician, or with the detection of an abnormality on screening mammography. The majority of women with these suspicious physical or radiological findings will have benign lesions. 1–3 However, most of these women are referred for multiple diagnostic tests, which can include additional mammography, ultrasound, fine-needle aspiration, and in some cases open surgical biopsy. 4,5 Such unnecessary testing of benign symptoms is a consequence of the low specificity of current screening and diagnostic modailities, and results in anxiety for patients and excessive health-care expenditure. 6,7 There is thus a need for a cost-effective means of safely reducing the number of diagnostic procedures undergone by these women. In this study we examined a new modality based on the analysis of breast electropotentials measured at the skin surface by an array of specially designed sensors placed in the regions of a suspicious breast lesion. 8,9 Previous studies have suggested that rapidly proliferating and transformed cells have electrically depolarised cell membranes compared with normal cells. 10,11 Furthermore, epithelial cells, which are electrically polarised because of a distinct apical and basolateral domain, also lose their transepithelial potential during carcinogenesis. 12–14 This effect is thought to extend from the cancerous area to the adjacent regions, and is measurable at the skin surface above the breast lesion. Field carcinogenesis, in which abnormal proliferation extends beyond a frankly malignant lesion in other regions of the affected organ, has been recognised in the breast and in other epithelial cancers for some time. 15,16 Thus, altered skin-surface electropotentials may indicate the presence of underlying abnormal proliferation that is indicative of cancer. The measurement of skin electrical potentials is non-invasive and does not expose the patient to ionising radiation. Such a test could be used on young patients and as often as needed without risk. In addition, the test result is immediate and objective. These combined advantages, along with promising results from pilot studies, 8 have stimulated interest in further assessment of this method. We carried out a multicentre European study to test this technique in women scheduled to have an open biopsy. Summary Background Proliferative changes in breast epithelium are an intrinsic aspect in the development of breast cancer, and result in regions of epithelial electrical depolarisation within the breast parenchyma, which can extend to the skin surface. Diagnostic information might be obtained from a non-imaging and non-invasive test based on skin-surface electropotentials. Methods In 661 women, scheduled for open biopsy at eight European centres, we studied whether measurements of breast electrical activity with surface sensors could distinguish benign from malignant breast disease. A depolarisation index was developed. Results We found a highly significant trend of progressive electrical changes according to the proliferative characteristics of the biopsied tissue. Discriminatory information was obtained in both premenopausal and postmenopausal women, and the index was not related to age. The best test performances were for women with palpable lesions. The median index was 0·398 for non- proliferative benign lesions, 0·531 for proliferative benign lesions, and 0·644 for cancer (ductal carcinoma-in-situ and invasive). A specificity of 55% was obtained at 90% sensitivity for women with palpable lesions when a discriminant based on age and the depolarisation index was used. Interpretation This new modality may have diagnostic value, especially in reducing the number of unnecessary diagnostic tests among women with inconclusive findings on physical examination. Understanding and control of the Electropotential measurements as a new diagnostic modality for breast cancer Jack Cuzick, Roland Holland, Volker Barth, Richard Davies, Mark Faupel, Ian Fentiman, H J Frischbier, J L LaMarque, Mirella Merson, Virgilio Sacchini, Daniel Vanel, Umberto Veronesi Early reports Imperial Cancer Research Fund, London, UK (J Cuzick PhD); National Expert and Training Center for Breast Cancer Screening, Nijmegen, Netherlands (R Holland MD); Esslingen Hospital, Esslingen, Germany (Prof V Barth MD); Hackensack University Medical Center and UMD-New Jersey Medical School, Hackensack and Newark, NJ (Prof R Davies MD); SpectRx Inc, Norcross, GA, USA(M Faupel PhD); Guy’s Hospital, London, UK (Prof I Fentiman FRCS); Hamburg University, Hamburg, Germany (Prof H J Frischbier MD); Institut Montpellier D’Imagerie Medico-Biologique et de Mastologie, Montpellier, France (Prof J L LaMarque MD); Istituto Nazionale Tumori, Milan, Italy (M Merson MD); Istituto Europeo di Oncologia, Milan, Italy (V Sacchini MD, Prof U Veronesi MD); and Institut Gustave-Roussy, Villejuif, France (D Vanel MD) Correspondence to: Dr Jack Cuzick, Department of Mathematics, Statistics, and Epidemiology, Imperial Cancer Research Fund, PO Box 123, 61 Lincoln’s Inn Fields, London WC2A 3PX, UK (e-mail: j.cuzick@icrf.icnet.uk) EARLY REPORTS