True in general. But it's okay for the special case of a reversible
process.
Actually you don't even need a "reversible" path.
Any path will do -- any fixed, simply-connected path.
A path of constant T will do, and is often useful.
A path of constant T+S will do, but is not very often
useful or physically significant.
You just integrate TdS along such a path, call the result
the Q du jour, and differentiate it to get dQ.
Caveats:
-- You need to specify *which* path, either by decorating
Q with subscripts, or perhaps by reeeally explicit context.
-- You cannot extend this type of Q-on-a-path to cover the
(T,S) plane, or any nontrivial subset thereof, or any other
nontrivial more-than-one-dimensional region ... because
the Qs for different paths will be mutally incompatible.
Do not imagine that there is any potential of the form
Qreversible(T,S) any more than there is a Q(T,S).
For details including pictures, see http://www.av8n.com/physics/thermo-forms.htm
and references therein.