I like that, especially the bootstrap around q1 drain. Cancels out what used to be called the "Miller Capacitance" in the old tube days. Starts to look like an oscillator though and what about the effect on noise figure ?. The 470r could be replaced with a current source to give a more balanced o/p drive
If I were trying to build something like this, I might cheat and download the manuals for hp and tek fet probes. They quite often used microcircuits for the input stage, but often included the internal schematic that might provide some pointers...
Regards,
Chris