DH+ (and RIO, same technology, different protocol) is a very forgiving network.
Sure, there is the "correct" way to install and terminate it, and the "correct" way is well documented.
However, getting down to fundamentals...
The purpose of the terminating resistor at each end of the "trunk-line" is to suppress signal reflection. When the transmitted data reaches a "dead-end" some of the signal is reflected back along the trunk-line, which interferes with the signal coming at it.
Consider that each node on the network, even if daisy-chain connected, has a "drop", from the terminations to the transceiver circuits. These are very short, so any signal reflection caused by the high impedance of the transceiver circuit is "in-phase" with the signal on the network, and any interference will be negligable. BUT, the last node on the cable, or the trunk-line just after where the last node is "dropped" needs a terminator.
I would suggest that whatever the topology, if you were to make sure that you have a terminating resistor at the end of each long run, then you may well be able to use a "star" configuration. I've seen it done several times with no issues.
Of course, RA tech support will only support networks installed "to the book", but when push comes to shove, and you just have to hook-up in a non-standard way, just think about where a terminating resistor should be installed.
On a side note : in a brewery I did a lot of work in, one DH+ network would go "intermittent" at roughly the same time, once a week. It was a puzzler, because the network had been installed to the book, using a high-level trunk-line, Station Connectors, and drops in galvanised conduit to the control panels.
Eventually, after the brewery engineers had failed to find any problems, RA Tech Support were called in. They did signal strength and quality measurements at various places on the network, and couldn't find anything wrong. So they did a full physical survey.
What they found was a drop into a decommissioned control panel was still connected to the trunk-line, there was about 15 foot of "blue hose" coiled up in the bottom of the panel, and the bare ends of the conductors were just sitting there.
The reason why the network went bad at a specific time in the week was due to scheduled cleaning of the area with high-pressure washers. The decommissioned panel had had all its push-buttons, switches and lamps removed from the doors when the internal kit was removed, so there were lots of holes for water to get in.
The survey found the coil of blue hose (with its bare ends), sitting in a pool of water in the bottom of the old panel !!