Renal cell carcinoma (RCC) is the most lethal urological cancer, and survival greatly depends on early diagnosis. Therefore, reliable, new biomarkers for detection of RCC are required. We assessed serum protein profiles of samples from two institutes with SELDI-TOF MS in duplicate on CM10 chips at pH 6.0 (set 1: 37 RCC + 32 healthy controls (HC), set 2: 20 RCC + 25 HC). Mean peak intensities of detected proteins were compared between RCC and HC with non-parametric testing. Classification trees were built with discriminating peaks using one sample set as training set and the other as independent validation set. We found 15 peaks significantly different (p<0.01) between RCC and HC. Two classification trees could be built with these peaks. Tree A achieved 75% sensitivity and 85% specificity for cross-validation and 76 and 65% for independent validation. Tree B had 71 and 62% sensitivity and specificity for cross-validation and 83 and 82% for independent validation. Although two serum protein profiles comprising 5 protein peaks were found that could separate RCC from HC, the sensitivity and specificity is not sufficient to recommend large scale use. Upon structural identification and quantitative validation, however, these proteins might prove suitable markers in the follow up of RCC patients.

Related Articles