To my mind, the hazard you're controlling for is metal. It's still metal unless your retailer has demanded otherwise. Other foreign matter contaminants are often badly detected in x-rays (whatever the glossy brochures say), the only exception I can think of is small stones in root vegetables but not a problem with your product.
I would write the HACCP plan with the hazard you're controlling for is metal, explain elsewhere how glass, hard plastic, bone etc are controlled for. The full hazard would be something like "persistence of metal contamination due to ineffective metal detection or x-ray". Then your control measure is effective metal detection or effective x-ray detection of metal.
In your validation, I'd validate both methods and prove that both are effective. Just remember to make sure that you buy test pieces appropriate to x-ray and train staff etc. Some of the controls are different (e.g. put the test card normally on top of pack, not centre of the aperture). You might have different test piece sizes so you'll need to justify both are acceptable. I normally use the FDA document on hard and sharp foods as a basis for that. That's it really. While some sites will use plastic and ceramic test pieces (the latter as a substitute for bone), the plastic used is often not the same density as plastic in food facilities which is often not detectable and cooked bones are often very hard to detect (especially small pieces). So they might do that and show it to the auditor but actually if you get them to slice off a piece of the plastic they're using or have a piece of bone returned from a consumer, both do not always detect.
An aside: I'm going to "big up" metal detectors still. I think the move to x-rays is understandable where you want that sensitivity improvement which is great for wet or multi component foods. BUT what metal detectors can do is detect multiple contaminants. I remember once having an issue where there was metal on metal contact on a line and we were getting small pieces of swarf in product. They were way below the limit of detection for x-ray but metal detection picked it up because the overall pack had a significantly changed signal in the magnetic field.