ABSTRACT
Context: Nipple areola (NA) sparing mastectomy has an acceptable complication rate, is oncologically safe and facilitates an improved cosmetic result, aiding greatly in reducing psychological trauma associated with breast loss. Questions regarding preoperative case selection for NA sparing mastectomy are pertinent. Aims: The principle objective was to develop a simple model based on correlation of malignant involvement of NA with morphological factors in breast cancer cases to accurately predict the cancerous involvement of nipple areola preoperatively. Settings and Design: The present cross-sectional study was carried out on 136 patients of breast cancer. The period of study spanned 3 years from 2004 to 2007. Materials and Methods: We evaluated 17 different morphological parameters which had proven prognostic significance in breast cancer cases for their relationship with NA involvement. Data regarding cytological parameters were available in 120 cases out of the total number of 136 cases. Simple and conventional methods appropriate for any under-resourced set-up were employed to enhance the economic viability and acceptability of the project. Statistical Analysis used: Statistical analysis in this study was mostly done using SPSS version: 14 software. P-value < 0.05 was considered significant when assessing correlation between two parameters. Results: The frequency of NA involvement detected in this study was 19.1%. In univariate analysis, 13 of the 17 morphological parameters were found to have strong statistical association (P<0.05) with NA involvement. In multivariate analysis, only four parameters-macroscopic NA changes, tumor-NA distance (<1.5cm), histological lymph node grade and extra capsular extension in lymph node were found to have independent role for NA involvement prediction. This multivariate Cox and Snell Regression model with Cox and Snell Regression Square of 0.551 can predict accurately 98.5% cases of nipple involvement using the 4 parameters as variables. Conclusions: By application of this simple multivariate model, accurate prediction of NA involvement would be possible preoperatively. NA sparing mastectomy may be performed on those cases predicted to have no NA involvement thus substantially reducing the burden of psychological morbidity.