The oncogene HER-2 is overexpressed in 25-30% of breast cancers and is the target for therapeutic strategies with monoclonal antibodies (Herceptin®). Anyway a soluble HER-2 (Shed-Antigen) has been characterized in the serum of patients with various carcinomas, however the correlation between tumor expression of HER-2 and respective HER-2 serum levels is still unknown. Therefore, we prospectively analyzed sera and tumor samples of 88 patients with advanced CRC and 20 healthy controls (only serum levels) for the expression of HER-2 using standard ELISA, immunohistochemistry (Hercept-Test®, Dako) and in situ hybridization techniques and tried to correlate it with lab parameters, tumor markers CEA and CA19-9, patients spezific parameters and overall survival. Results: HER-2 serum levels were significantly higher in pts with CRC compared with the control group (median 2511 vs. 2184 HNU; p=0.038). 67 of 88 pts (75%) showed elevated serum levels (>3000 HNU), but in only 16 of 74 pts (21.6%) weak expression of HER-2 could be detected in tumor samples (9x 1+, 7x 2+, 0x 3+ Dako score!). Only 1 of 6 tested tumor samples that yielded a Dako score of 2+ was positiv in CISH. HER-2 serum levels significantly correlate with CEA (p=0.0159), bilirubin (0.0198), alkaline phosphatase (p=0.0050) and with ongoing anti-tumor therapy (p=0.0046), but not with number of metastatic sites, CA 19-9, CRP, LDH, creatinine, hematological parameters (hemoglobin, leukocytes, platelets) and age. Patients with liver metastases were more likely to demonstrate elevated serum levels of HER-2 (p=0.0033), but did not show HER-2 overexpression in the respective primary tumors. No positive correlation could be found between serum levels and tissue expression of HER-2 (p=0.035). The IHC (Dako score 0 vs. 1/2) showed a significant correlation with the presence of liver metastases (p=0.047) and serum alkaline phosphatase (p=0.0398), but not with number of metastatic sites, CEA, CA 19-9, LDH, CRP, bilirubin, and age. Surprisingly, patients with normal serum levels of HER-2 ( Conclusions: Very few colorectal cancers overexpress HER-2 when assessed by standard IHC and in situ hybridisation (here: CISH). Therefore, molecular approaches targeting this oncogene in CRC are not necessarily warranted. The pathophysiological role of elevated levels of soluble HER-2 in the serum of a substantial number of patients with CRC has yet to be determined. And so the soluble HER-2 can be an useful prognostic factor or a paramter for the clinical course of desease in CRC.