I replaced the filter with Gerard's LC preselector design and replaced the radial RF choke with an SMT part.
System gain is now about 14.5 dB.
Input -60 dBm, FHS3 reads -66.5 dBm at the MAX2611 input, -45.8 dBm at MAX2611 output.
This seems much closer to the expected amplifier performance!
Filter loss still a little high. Filter performance seemed good with the tracking generator - 150 may have just been on the edge, but shouldn't have been 6.5 dB down. It's hard to say without being able to calibrate the 'through' test accurately. Could do with an SMA-SMA female joiner (built!).
Added a 25 dB pad (Pi, 56R, 440R, 56R)
Input -60 dBm, FHS3 reads -71 dBm at MAX2611 output = 25.2 dB attenuation.