We present 350 μm continuum images of 24 massive star formation regions obtained with the Caltech Submillimeter Observatory equipped with the SHARC focal-plane array. At this wavelength the spatial resolution is 11''. The 350 μm emission is compared with thermal radio continuum emission and OH, H2O, and CH3OH masers. Emission at 350 μm is believed to be thermal emission from dust heated by embedded or nearby stars. Compact radio continuum sources are usually present in the mapped 350 μm fields, and more than 60% of the 350 μm peaks coincide with radio continuum peaks. This association lends strong support to the notion that the dust is heated primarily by hot stars. Masers are also a common property of massive star formation regions. Usually OH, H2O, and/or CH3OH masers are found near, but generally not coincident with, the 350 μm peaks. Less than 25% of the 350 μm peaks have no reported masers located close by. In most of the observed regions, the 350 μm dust maps show one or more components surrounded by fainter extended emission. In total, we identify 28 separate 350 μm components. Ten of the 28 components do not have radio continuum counterparts. These are luminous sources and should produce detectable H II regions. It is postulated that these sources may be undergoing such rapid accretion that the infalling matter quenches the H II region very close to the protostar, thereby making the H II region undetectable in free-free emission. Objects in this evolutionary state may well represent the long-sought precursors to ultracompact H II regions and should have the properties of accreting massive protostars (e.g., accretion disks, bipolar molecular outflows, hot shocked gas, infall with spin-up toward the protostar). We suggest that the 10 sources in this category in our sample merit further observational study for these properties. Two-temperature graybody models constrained by the observed infrared spectral energy distributions were calculated to estimate the total mass, luminosity, average dust temperature, and hydrogen column and number densities for each source. The graybody models do not account for possible low-level emission from the coolest dust (<25 K) and therefore may underestimate the total mass. There is, however, no evidence for emission in excess of the models at 1.3 mm, and we conclude that the mass contribution of dust colder than ~25 K cannot be large. Thus, the graybody models shown provide useful average global properties required to understand massive star formation environments.