Aberrant glycosylation is a hallmark of various disease states, including cancer, and effective detection and discrimination between healthy and diseased cells is an important challenge for the diagnosis and treatment of many diseases. Here, we describe the use of boronic acid functionalized synthetic lectins (SLs) in an array format for the differentiation of structurally similar cancer associated glycans and cancer cell lines; discrimination is based on subtle variations in glycosylation patterns. We further demonstrate the utility of our SLs in recognizing glycoproteins with up to 50-fold selectivity, even in 95% human serum. Given their robust and selective nature, these SLs were able to effectively distinguish (a) five structurally similar glycans with 94% accuracy; (b) seven normal, cancerous and metastatic colon cancer cell lines, including three isogenic cell lines, with 92% accuracy; and (c) these same seven cell lines using a guided statistical analysis to improve our analysis to 97% accuracy. In total, these data suggest that an SL-based array will be useful for the diagnosis of cancer.