The goal of brain cancer surgery is to strike a balance between removing diseased tissue and leaving healthy tissue. But to the surgeon, cancer tissue and normal tissue can be difficult to tell apart. Imaging methods based on chemical signatures might be able to help distinguish between the two. To that end, a research team led by Harvard University’s X. Sunney Xie has used label-free, two-color stimulated Raman scattering microscopy to detect human glioblastoma in mice and in tumors removed from human patients (Sci. Transl. Med. 2013, DOI: 10.1126/scitranslmed.3005954). The researchers acquired images of excised brain tissue using Raman signals at 2,845 cm−1 and 2,930 cm−1, which are representative of proteins and lipids, respectively. They colored the lipid signal green and the protein signal blue. Tumors show up as mostly blue. Neuropathologists classified brain tissue in the samples using both Raman and conventional histological staining, making only two errors out of 225 Raman samples.