I would take the highest power density level allowed in one of those interference bands, turn that into two discreet theorectical tones spaced apart by your IF frequency, and use the simple IP3 calculation to see how big a third order product is generated in my IF frequency band. I would then compare this theoretical jammer tone in the IF to the desired IF signal level. The amount of front-end out-of-band filtering will, obviously, lower one or both of the tone levels and will give less interference in the IF.