Lipids in tissues present something of a challenge to histologists and pathologists. Lipids are relatively unreactive, chemically speaking. Even unsaturated lipids have few available sites to which stain molecules can bind. We've seen that osmium tetroxide will covalently add to lipids, but OsO4 is pretty nasty stuff to handle and it's extremely expensive. As a routine light microscopic stain for lipids it has little real practical application.
Another problem is that lipids don't remain in the tissues in routine processing: they get leached out by the numerous solvents used in preparing the typical wax sections for staining (i.e., xylene and alcohol). Hence, by the time the stains get applied, the presence of lipids is inferred by their absence in the section: large areas without any staining and without, in fact, any tissue at all. This isn't a problem in plastic-embedded tissues, which are processed without the prolonged solvent immersions needed for wax; but 99% of all clinical labs still use wax embedment because of its low cost, its suitability to automatic tissue processors, and the long-established and well-tested staining routines that they use. Plastic sections are superior in definition and detail, but wax is faster and cheaper, and for diagnostic purposes, almost as good.
Most lipid "stains" aren't truly stains at all. They don't form covalent bonds with lipid components, and they can't for ionic bonds or hydrogen bonds, either. Instead, lipid "stains" depend on preferential solubility of some chemicals. Twp examples are Oil Red O and Sudan Black. Both of these use preferential partitioning into the lipid compartment to achieve their results.
It should be obvious, though, that in order to be dissolved in the lipid, there has to be lipid present. The way around the extraction difficulty is to use tissue sections that have been frozen, not embedded in wax or plastic. The tissue is rapidly frozen rock-hard and held at liquid nitrogen temperatures. It's cut in a special refrigerated microtome called a cryostat (right) just as easily as embedded specimens are. These are then thawed and treated with a lipid stain with the results you see above.
Another advantage to frozen sections is that they're fast. They're sometimes used for quick biopsies while a patient is still under anesthesia. In that case they're stained with Toluidine Blue, and the surgeon has a pathology "read" within minutes so he can decide what to do. The principal drawback to frozen sections is that they can't be stored: since they're usually not fixed in formaldehyde, they deteriorate rapidly.