Do coronal mass ejections dissipate or decay over time? What is the life cycle of a CME after it is ejected?  All accounts I've seen focus on impacts at Earth, but don't suggest how it evolves over time as it progresses away from the Sun.  Does it expand, or do its electromagnetic structures provide some form of structural integrity?  My prejudice is that the plasma would cool in space and convert into hydrogen and helium.
 A: 
Do coronal mass ejections dissipate or decay over time?

The short answer is yes.
The longer answer is that coronal mass ejections (CMEs) expand nearly as fast as they propagate so that by the time they reach Earthy they are nearly one astronomical unit (AU) in size.  Fast CMEs generate shocks, thus the magnetic cloud/obstacle of the CME acts as the shock piston (there is a full list of magnetic obstacles at https://wind.nasa.gov/ICMEindex.php).  As they propagate, they also undergo something called mass loading – the otherwise balloon-like shape of the CME becomes distorted in regions where its leading shock wave needs to heat and decelerate more plasma.
The public can actually request model runs of CMEs from NASA's Community Coordinated Modeling Center (CCMC) at Goddard Space Flight Center.  The results can include movies of the propagation of these phenomena.  I do not recall offhand to what radial distance the simulations will go, but I know that group has examined the evolution into the outer heliosphere.
There are also some nice visualization produced by the Scientific Visualization Studio (SVS) at Goddard, e.g., there is a nice example of an event that hit Mars at https://svs.gsfc.nasa.gov/4058.

What is the life cycle of a CME after it is ejected?

It takes a CME anywhere from ~18 hours to 4 days to reach Earth, i.e., at 1 AU.  Thus, it takes anywhere from ~75-400 days to get to 100 AU if we ignore deceleration.
If we define the leading edge of the shock as our point of reference, then multi-spacecraft studies have shown that they do decelerate as they propagate outwards.  This should be expected, since for shock waves to exist they need to dissipate energy (e.g., see https://physics.stackexchange.com/a/306184/59023).
As I stated at https://physics.stackexchange.com/a/139436/59023, https://physics.stackexchange.com/a/210097/59023, shock waves in space are collisionless – they rely on mechanisms other than binary particle collisions to convert the incident bulk flow kinetic energy into other forms (e.g., heat).  One of the mechanisms is particle reflection and acceleration.  This process takes energy/momentum away from the piston, thus decelerates the entire structure.

My prejudice is that the plasma would cool in space and convert into hydrogen and helium.

The plasma in the heliosphere stays nearly fully iononized all the way out to the termination shock, but it does cool down.  Interestingly, it does not cool adiabatically, as one might expect from an expanding gas.  Why the solar wind remains hotter than expected as it expands is one of the many big open questions in space plasma physics.
As an aside, the temperature inside magnetic clouds is, on average, much lower than the ambient solar wind.

Does it expand, or do its electromagnetic structures provide some form of structural integrity?

Yes, CMEs expand as they propagate as I stated above.
