RESUMEN
INTRODUCTION: Breast cancer screening increased the ratio of small tumours. These tumours have a low lymph node metastatic potential. Sentinel node detection allows detecting axillary lymph node invasion without the morbidity of complete axillary lymph node dissection. OBJECTIVES: In this study we report the results of the learning curve of sentinel node detection in the Institut Salah-Azaïz of Tunis. MATERIALS AND METHODS: It is a prospective study between January 2004 and December 2005 in which 115 patients were included with breast cancer less than 3 cm without antecedents of breast surgery. All these women had sentinel node dissection by a colorimetric method and 30% had a combined method (colorimetric and isotopic). RESULTS: The rate of detection was 97.3% (n = 112). An extemporaneous examination was performed in 91 patients. The rate of negative forgery of the extemporaneous examination was 4.3% and the sensitivity of 95.7%. There are no false positive with the extemporaneous exam. The sentinel lymph node was the only node invaded in 15 patients (44%). In 3 patients, the sentinel node was healthy whereas the axillary dissection was positive, so the false negative rate is about 2.6%. CONCLUSION: Sentinel node dissection is a reliable and feasible technique. It however requires a training of the surgeon, the pathologist and the nuclear doctor. It allows to reduce the morbidity of the treatment of the breast cancer by avoiding "useless" axillary dissection out in patients without node invasion. The increase in the number of the small cancers discovered during screening makes it possible to increase the number of patients who can profit from this technique.