Detection of DCIS and invasive ductal carcinoma of the breast in human serum by protein profiling (SELDI)

Abstract

Introduction: Several methods are suited to classify different cancers by testing RNA or protein expression. No method is available to screen for DCIS. This study is performed to differentiate serum samples of healthy women and patients with benign, premalignant and malignant lesions by protein profiling. Additionally the diagnostic value of SELDI-MS (surface-enhanced laser desorption ionisation- mass spectroscopy) for detection of DCIS and invasive ductal carcinoma is investigated. Methods: SELDI is a protein chip based procedure and is used for quantitative measurement of protein mixtures like serum. Here a weak cationic chip is chosen. Obtained protein profiles are combined in an algorithm to correlate expression patterns and histological diagnosis. Blood samples from 112 women are analysed (28 from healthy controls, 27 with benign breast lesions, 25 with DCIS and 32 with invasive ductal carcinoma). Results: SELDI reveals 27 significantly discriminative protein peaks (3000–15000 Da). In an algorithm various combinations of four significant peaks are used. SELDI analysis and data entry into the algorithm reveal a good differentiation between premalignant and malignant lesions and normal controls. Statistical analysis shows a high sensitivity of 0.89 and 0.88 for detection of DCIS and invasive ductal carcinoma in patient serum. Conclusion: Identification of differentially expressed proteins in serum samples related to women with tumors of different malignant progression, by SELDI technique, turned out to be possible. A high prediction strength of nearly 90% for DCIS underlines the importance of this new method. SELDI may become an important screening tool for detection of DCIS and early invasive ductal carcinoma.