I would have thought that, for photometry, you would have little or no requirement to convert the image from RAW format. You are, after all, trying to determine the response curve of photosites to incoming photons.
What PixInsight can offer you is the ability to read in a RAW image and to then split that image into four new images - each corresponding to one 'colour' of the CFA. The ADU data is not 'modified' in any way - all that happens is that the 'positional reference' of the original photosite determines where, in the new image(s), and into which new image, the original ADU will now be placed.
At that point you can still examine data statistcally although, persoanlly, I would create a Preview to identify the 'central area' that you referred to, and then I would create a Crop Process Icon (note: not a Dynamic Crop) from that Preview, and save that Process to the Workspace. Then I would load in all the images that need to be examined and make an ImageCpntainer from these, and process its contents using the previously saved Crop Process. You can also use the Readout Tool to statistically examine the area under the probe (the mouse pointer) on each image - this may be good enough for you, especially if you set the probe 'size' to a suitably large array of pixels.
Hope this helps.
