You can feed a 1420 MHz hydrogen-line horn from the rear panel, but a straight monopole pointing forward down the horn axis would generally be a poor arrangement.
The key issue is orientation. In the normal TE₁₀ mode of a rectangular hydrogen-line waveguide, the electric field is transverse to the direction the radio wave travels. Your receiving probe therefore needs to be parallel to that electric field, rather than parallel to the horn axis.
So:
A probe entering through the side/broad wall, projecting across the waveguide in the E-field direction → good conventional arrangement.
A probe mounted in the back wall and sticking straight forward toward the horn mouth → very weak TE₁₀ coupling and not recommended.
A coax connector through the back wall whose centre conductor is then bent through 90° inside the waveguide, so the active portion lies parallel to the TE₁₀ electric field → can work perfectly well.
For a WR-650-sized 1420 MHz waveguide, the free-space quarter wavelength is about 52.8 mm, while the guide wavelength is longer, roughly 274 mm, making a quarter-guide-wavelength spacing from a shorted back wall about 68 mm. In practice the exact probe length and position are best adjusted with your VNA.