I recently read an [interview][1] with Sergei Novikov where he claimed that he introduced "toric neighborhoods" in his proof of the topological invariance of Pontrjagin classes. He draws a comparison to Grothendieck's invention of "generalized open covers" (now known as Grothendieck topologies), and remarks that the same kind of construction was used by Kirby et al. in later developments. [1]: https://homepage.mi-ras.ru/~snovikov/998.pdf