But China didn't had the same frontiers at all during all the imperial dynastic rule : it ruled over core areas and not necessarily all of them in the same time (the traditional North/South contrast). Similarily, an imperial Roman state apparatus could have survived as a dynamic and varying entity either as ERE as IOTL, either as an Italo-Roman WRE maintaining an hegemonic patronage over various warlord post-imperial states (either Barbarian as ITOL, either Roman).
You'd "just" need maintain of imperial state apparatus and its predominance to reach something close enough : for instance, in a no-Islam TL, Roman Empire centered in ERE, Italy and Africa as equivalent of Chinese
"protectorates", Francia as Korea, etc.
The idea that Rome have either to expand, or at least keep ideal borders to survive (either Rhine/Danube as IOTL, either Elbe/Vistula/Prout because it looks good enough on a blank map) doesn't seem quite obvious to me, giving that most long-lasting empires did well enough with varying cores and borders.
So, really, up to a point (let's say the VIIth century), you can go wild well enough.