How could a particle be isolated to avoid decoherence? The question aims to this issue : if there is some technological arrangement (or action) to take over the particle/system in order to keep it in a coherent state, then the field, (force or whatever) keeping it away from  interacting with "an external system" isn't it itself an interaction? 
I mean, supposing you reach  enough isolation to avoid decoherence.
How do you know the particle is still there?
Thanks
 A: If you're talking about building a quantum computer, then there are some modes of the system which you need to keep isolated so that you can make sure that any coherence of these modes is preserved, but there are other modes of the system that you use to control the system, and these aren't isolated. This idea is also used in quantum error correction. This process uses active control on certain modes of a system to suppress the decoherence of other modes of the system. You can be sure that the system is still there by observing the modes that don't need to be isolated. A similar idea is used in building a  quantum logic clock, which is the most accurate clock ever built. 
A: There's a lot of creativity in constructing an experiment, exploiting many different interactions in many different configurations, so there isn't a one size fits all Answer to your Question.
One source of decoherence, however, are thermal fluctuations of the EM field (the immediate environment), which are driven by the thermal fluctuations of whatever surrounds the system of experimental interest. To reduce the effects of decoherence, we can surround the system by something cold, so that the EM field is more driven by the cold surrounding, and much less driven by the hotter surroundings that are further away. How we keep the cold surroundings from heating up, and how we create the cold surroundings in the first place, are relatively modern miracles of invention, which in Physics experiments are likely to be multi-stage exploits. To some considerable extent, as refrigeration technology improves, so improves Physics.
We can't totally eliminate thermal fluctuations of the EM field, insofar as the 3rd law of thermodynamics is empirically supported, so yes, the surroundings still affect the system of experimental interest, but less.
A: As an experimentalist, I will first address this last summary of your question:

How do you know the particle is still there?

Let us define the terms of the question:
Particle.
a) In particle physics we know a particle was there by the tracks it leaves in a bubble chamber.
b)By the signals it sets off as it passes and ionizes 
The measurements have shown us that we deal with very small dimensions in all quantities, mass, size etc.
We also have found that particles follow quantum dynamics and the solutions of the appropriate equations of motion.
Generally: can I trap one particle and "know" it is there? I have not done it, but it is being done billions of times a second at the accelerators. If I went to the trouble to design an experiment that has trapped a single proton in a magnetic configuration, I would know it was there from the radiation it would emmit as it oscillated in the magnetic trap. 
Usually though, because of the very small values accompanying the existence of a particle one deals with a flux of them at a time.
Now coherence. Coherence is the term describing the quantum mechanical solution of the equations of more than one particle, and refer to the phase differences between those particles : i.e. coherence means that those  phase differences remain constant . Described as quantum mechanical waves, the particles are "in step". If you only have one particle, as in my gedanken experiment above, the quantum mechanical solution is known and phases can only be defined with respect to the field. As long as energy is supplied to my proton this description will hold.
The "know the particle" phrase should become "know the particles" phrase.
Coherence is observed macroscopically:
in laser light
in superconducting magnets, over kilometers of wire length.
in superfluidity.
All these require zillions of particles and no question should arise if they are there or not. The answer by Peter Morgan addresses the question of stability of such systems.
Now I suspect you are asking the question from statements of coherence and the density matrix formulation. This has to do with the quantum mechanical statistical behaviour
 of many particles, so again, your one particle question does not compute.
You should maybe clarify in your head what you really want to learn about coherence. Maybe the density matrix formalism confuses you?
