Not all cannabinoids in the smoke are carboxylated, or at least lets agree that the efficiency of your average conversion is not 100%. Some degree of non-carboxylated cannabinoids can ride those particles too .. for example at the beginning of a joint, the smoke is being drawn through THC fields which have not yet been burned, but may be dislodged and part of the draw as the heated substrate is pulled into the lungs ..
Agreed, there may be some applicability of this study on smokers (albeit at a reduced effect size). Worth betting on it being protective? Or just a rationalization?