Treatment options for Hepatocellular carcinoma (HCC) are limited. Surgical resection or transplantation are most effective, but patients are often not candidates due to late presentation. Alpha fetoprotein (AFP) has traditionally been employed as a serum screening test, but is only 60% sensitive and 80% specific. Other serum markers are needed to compliment AFP in the early detection of HCC. Methods: In this study serum was obtained preoperatively in patients undergoing orthotopic liver transplantation. Whole serum was spotted onto immobilized metal affinity chips (IMAC30 ProteinChip) bound with copper and allowed to incubate. A mass range of 1400-50000 Daltons was analyzed by SELDI-Time of Flight mass spectroscopy (SELDI-TOF). SELDI-TOF proteomic spectra were then correlated with ex-planted liver pathology. Initially, 58 serum samples with a known surgical pathologic diagnosis were analyzed as a “training set”. Researchers were then blinded to the presence or absence of HCC in a subsequent prospective “test set”. Statistical analysis was performed using Biomarker Pattern Software (Bio-Rad, Inc.). Results: A total of 88 serum samples were analyzed; 29 from patients with cancer and 59 from patients without cancer. These were further divided into a training set (19 cancer, 39 non-cancer) and a prospective test set (10 cancer, 20 non-cancer) for statistical analysis. The training set produced two decision trees which classified the cancer samples from the non-cancer samples. The first decision tree used two peaks as nodes to split the groups (mass/charge values of 2783 and 2896Da). Using these two values, decision tree analysis on the prospective test set resulted in five false positive and four false negative classifications (sensitivity 71%, specificity 80%). The second decision tree used five peaks as nodes to split the groups (mass/charge values of 1966, 2783, 2896, 2945, and 4082Da). Using these five values, decision tree analysis on the prospective test set resulted in two false positive and four false negative classifications (sensitivity 71%, specificity 91%). Conclusions: SELDI-TOF mass spectroscopy of preoperative serum samples is able to distinguish patients with HCC vs. patients without HCC prior to orthotopic liver transplantation with greater sensitivity and specificity than reported for AFP. Our proteomic analysis of these patient serum samples indicated that the SELDI markers identified on the IMAC chips increased the accuracy of identifying the presence of HCC nearly 30% by a combination of reducing the number of false positives identified by ~10% and reducing the number of false negatives identified by nearly 20%.