Thought id chime in on this.
I do alot of walnut blasting here on CP-E. There are very few motors that we work on that aren't direct injection. Starting with the MZR Mazdaspeed motor (which is a direct predecessor to the 2.0 and 2.3 ecoboost motors). Those motors have EGR which makes the problem a little worse.
At 19.5K I removed the intake manifold, took pictures, and blasted all the intake valves. they were suprisingly bad for 20k miles. I tuned my car starting at 700 miles on the odo. I installed a mishimoto catch can at 19.5k after this proceedure. Car has pretty much had stage 3 bolt ons since the beginning.
So far ive blasted about 20 other cars with no negative side effects. However, I am yet to dyno a car, blast and then redyno without making any other changes. I also recently did a second blast on an N54 BMW (also DI) after I blasted and installed a catch can 20k before that. The catch can did not seem to help much, however, it was pretty full of crap.
Before
After
I plan on removing intake manifold again at 40k and checking the effectiveness of the catch can. Even if the catch can isn't doing much to help the valves, I still believe it is useful in keeping oil filled air out of the motor which reduces the effective octane.
All the cars I do with ~100k are so much worse. You are effectively blocking the passageway of air into the motor. I haven't confirmed this but you should see a decrease in LTFT as the car is higher mileage as at the same load levels, less air will be getting into the cylinder, and less fuel is needed to reach the desired AFR. By how much? Who really knows. There are too many other factors that can also effect your volumetric efficiency over the life of the motor. Also, as this carbon buildup gets worse, chunks can go into the cylinder and build up on the piston and create hotspots or go through the exhaust and potentially cause other damage.
Just upgraded my blaster. Excited to get better results!