Making it work I am going to punt on part of this to John Bradford of Makino below.
I would add to his great recommendations that you test the runout in the spindle at operating temp. Use a good 50 Millionths test indicator and a mag base or gage stand with fine adjust shafts. We will part from our normal advice and say that pre-balanced holders will work here (the endmill has so little mass) and that the holder’s projection or diameter will not make a difference one way or another. Runout and balance are key, but the tool is still very flexible and dynamics are in play. While we can't tap test miniature endmills there is some science that can be applied. The Harmonizer by MLI can record the sound of the cut and determine resonate speeds where the flexing is at its worst and need to be avoided. This will work with spindle speeders and air spindles if you can adjust the output speed. https://goo.gl/WLeUy4
On standard spindles and toolholders, RCSA can be used to model the miniature endmill and predict stable speeds. Once you get a process dialed in you will want repeatability of the tool stick-out. The rings used on PCB drills might be a good solution if the shanks are 1/8" or 3mm.
Sharing information about high performance milling technologies, the result of 30 years of research.