To screen for glycoproteins showing aberrant sialylation patterns in sera of cancer patients and apply such information for biomarker identification, we performed SELDI-TOF MS analysis coupled with lectin-coupled ProteinChip arrays (Jacalin or SNA) using sera obtained from lung cancer patients and control individuals. Our approach consisted of three processes (i) removal of 14 abundant proteins in serum, (ii) enrichment of glycoproteins with lectin-coupled ProteinChip arrays, and (iii) SELDI-TOF MS analysis with acidic glycoprotein-compatible matrix. We identified 41 protein peaks showing significant differences (p<0.05) in the peak levels between the cancer and control groups using the Jacalin- and SNA-ProteinChips. Among them, we identified loss of Neu5Ac (alpha2,6) Gal/GalNAc structure in apolipoprotein C-III (apoC-III) in cancer patients through subsequent MALDI-QIT-TOF MS/MS. Furthermore, subsequent validation experiments using an additional set of 60 lung adenocarcinoma patients and 30 normal controls demonstrated that there is a higher frequency of serum apoC-III with loss of alpha2,6-linkage Neu5Ac residues in lung cancer patients compared to controls. Our results have demonstrated that lectin-coupled ProteinChip technology allows the high-throughput and specific recognition of cancer-associated aberrant glycosylations, and implied a possibility of its applicability to studies on other diseases.