From my North American engineering perspective:

The main thing is that when you look at a new transit project in isolation, the cheapest thing to design and build is a transit system that interacts with as few existing parts as possible. There are plenty of exceptions to this, such as using an existing rail tunnel through a mountain vs. digging a new one, or if an existing maintenance facility has enough spare capacity to avoid needing to build a new one. But in every place the new line interacts with existing infrastructure, it's a cost to determine how the design can best integrate it both during and after construction. Take an interchange station with another line. Will it go beside, above, underneath? Can it be made without needing the close the station or the entire line, or with as little disruption to existing service as possible? Interchange stations are great for usability, but planners and design bidders evaluate and present cost-to-benefit tradeoff scenarios that will get approved or denied for both fiscal and political reasons. Let's say between two distinct lines you want to have shared track. If the electrification type, rail gauge, signalling system, platform length, vehicle profile, boarding level height are different, you will have to spend a solid amount of engineering effort figuring out how to harmonize it.

There are many examples of light rail vehicles using track or right-of-way shared with heavy rail in full or partial sections of the line. Waterloo, Canada's ION tram is one. Unfortunately it's much more difficult and costly to have it the other way around-- i.e. heavy rail vehicles on light rail tracks. The tracks and roadbed are not meant to handle the weight and vehicles may not make it through the smaller bridges tunnels, and curves. To make a light rail system compatible without knowing what heavy rail trains would use it, is a major cost incurred for no forseeable benefit.