Hepatocellular carcinoma (HCC) is the seventh most common cancer in the world and accounts for over 800,000 deaths annually, making it the second most lethal cancer. Poor chemotherapeutic agents against HCC have necessitated the importance of early detection and surgical intervention in patient management. However, the search for biomarkers of HCC have proved difficult, as serum and tissue proteomics, transcriptomics and genomics have proved inconclusive. We have utilized a novel spatial-omics based glycan imaging method to identify glycan changes that occur directly in HCC tissue. Glycan changes included branched, fucosylated and high mannose glycan. Importantly, using matching serum samples, we were able to identify the glycans that are altered in tissue and also in the serum of individuals with HCC. Glycoproteomics was used to identify the proteins containing these glycan structures. To translate these findings, we utilized a novel multi-plexed antibody panel based glycan imaging method, which allowed us to examine all the glycans associated with the identified glycoproteins. These glycan changes, found in serum, and directly associated with the tumor, were used to create a diagnostic algorithm to detect HCC. These spatial omics based, machine learning (SOML) algorithms were tested in a discovery set (n=201) and an external validation set (n=192). AUROC values ranged from 0.95 in the discovery set and 0.91 in the independent external validation set. In conclusion, we present the development and application of a new biomarker platform that can be used to examine the glycans associated with disease.