I've done it both ways mentioned above. And unless your going to have a lot of click zones throughout I'd use David iz cool's suggestion. You can even make the sprites transparent, and it's takes a lot of the work out checking tons of rects.
The other way, is to have a BMP that has zones (sorta like a mouse map), that isn't displayed on the screen that you can check the pixel color for mouse clicks.
If you are going to check the mouse area, it's better for you to code some handling routines, and have a UDT keep track of image zones.
See which method you like best!