Bonifacino, A., Petrocelli, V., Pisani, T., Simone Giannerini, Giovagnoli, A., Vecchione, A., Mingazzini, P. L., Giovagnoli, M. R., A. BONIFACINO, V. PETROCELLI, T. PISANI, S. GIANNERINI, GIOVAGNOLI A., A. VECCHIONE, P.L MINGAZZINI, and M.R. GIOVAGNOLI
Purpose: This paper reports a prospective study designed to measure the accuracy of mammography, ultrasonography, fine needle aspiration cytology and one of the most recently introduced techniques, vacuum biopsy, in the diagnosis of breast cancer. Experimental Design: A sample of 146 breast lesions on 135 patients were examined at our Breast Centre. The lesions had been examined previously at different outpatient centres, without any definite conclusion being reached. The design of the study made it possible to compare MRx, US, FNAC and VB directly, because it excluded several confounding variables. The histological diagnosis on surgical specimen, when available (45 cases), was considered as the gold standard. In the other 101 cases the biopsy results complemented by the negative clinical results of the follow-up were taken to be the true outcome. Statistical indicators (sensitivity, specificity, predictive values PPV and NPV, FN and FP rates, suspicious plus indeterminate rate and likelihood ratios) have been calculated. Results: The NPV of MRx and US are remarkably high (92.4% and 97.9% respectively), confirming previous reports. Complete sensitivity of FNAC turns out to be 80%, while specificity and LR of positive test were 99.1% and 88.8 %. The combined score of FNAC, US and MRx results in a good increase in complete sensitivity (97.1%), if compared with the results of the single diagnostic tests evaluated separately. Absolute sensitivity of VB is 97.1% and specificity is 100%. Conclusions: Considered together, MRx, Us and FNAC appear to be reliable diagnostic procedures and when they are all negative the possibility of a cancer is extremely low (NPV = 0.988 in our sample of women) although it cannot be completely ruled out. The VB test, obviously, has the highest absolute sensitivity among all methods compared and is significantly more sensitive than the Combined Score. Therefore, this technique could be considered conclusive in diagnostic doubtful cases, avoiding open surgical biopsy. 1. Introduction Breast cancer accounts for one-third of cancer diagnoses (more than 192,200 new cancer cases in 2001) and 15% of cancer deaths in U.S. women1,2 and is believed to be the most common cause of malignancy for women in most other Western countries too. In Europe breast cancer accounts for 34% of all prevalent cancers in females and in Italy over 33000 new breast cancer cases are expected in 20023. Over the last 25 years the diagnostic approaches and therapeutic strategies for breast cancer have changed dramatically and mortality rates have fallen by 21% as a result of earlier detection and improved treatment. The diagnostic tools have gradually become more diversified and sophisticated, with a consequent improvement in their resolution limits and accuracy, the tools referred to being digital mammography, ultrasound (US associated with Colour Doppler, Power Doppler, Contrast Agents), magnetic resonance, fine needle aspiration biopsy (FNAC), Core Biopsy (CB), Vacuum Biopsy (VB) and advanced breast biopsy instrument (ABBI). A crucial role has been played by the introduction of national screening programmes4. At present, according to the European Quality Assurance Guidelines in Mammography screening5 the “Triple Approach” is recommended: combined physical examination, mammography (MRx) and FNAC. By introducing mammography screening programmes, however, the size of detected breast lesions has become smaller and the diagnostic problems greater6 so that it seems necessary to redefine the conditions and indications in the use of the new diagnostic tools recently developed. The aim of this paper is to describe a prospective study carried out in a well defined clinical and experimental setting in order to report the accuracy of certain well-established techniques, such as MRx, US and FNAC in our sample of women, and to measure the accuracy of a recently introduced one, VB, in the diagnosis of breast cancer.