What is perhaps a bit odd, is that the closer to the base of the output transistors I get, the less power there is. >2W at C17, 1.8W at C18, 1.3W at C39, 0.3W at the base of the output transistor. 1.3W on the collector of the output transistor, and approx 2W coming out.
That's normal. The impedance of an RF transistor is much lower than 50 ohms. The input network of the amplifier aims to not only divide the input power between the multiple transistors, but also transforms down to a lower impedance.
Similarly, on the output of the transistors the impedance is very low, and there is a matching network to bring the impedance up to 50 ohms.
What this means is, without knowing the exact input and output impedance of the transistors in question, the readings from your RF probe don't mean much. You can use it for comparative measurements, but that's about it.
Did you measure the DC voltage at the base of each transistor (with PTT applied and no RF input)? It should be around 0.7V. Also I don't know if you have the specs on what the bias current is supposed to be for this amplifier, but similarly you should measure the collector current (again, with PTT on but no RF).
Taken out of circuit, BJT RF power transistors can be tested pretty much like any other transistor. But since you've got a linear amplifier a basic check is to look for a sensible bias voltage at the bases, and current draw at the collectors, which if you get both would indicate that the transistors are at least working to some extent.
Do you have more documentation than just the schematic? There isn't even component part numbers indicated on there...