For Scribus, you would set up a paragraph style (see: stylesheets) such as "figures - Header"; place a text box to overlap the image box, add the "A" and then mark up the "A" with the style. If you edit the style, it will update all instances of the style throughout the document. Using only stylesheets (within reason) is always the best way to proceed and helps ensure uniformity.
If you are going to mark up the images before placement, then you need to ensure that both (all) images are set up so that they are not going to be scaled in any way after placement. Then you choose a specific the size typeface, possibly specified in (pixels * output intent dpi). You will need to handle this across all images in the document and then you will need to edit each image file with a recalculated size if you need to change dimensions of the image or need to edit the typeface.
Sizing the text: if you KNOW you are placing images at 100% and that you are printing 300dpi/150lpi, you can deduce that a 1/8 inch high "A" will be 300*.125 = 37 pixels. If you resize the image, the text scales with it.
The Scribus way is the better way, but if you are doing complex charts with text within them you need to set up your images as uniformly as possible so that you are not scaling the type. Generally, even in vector formats and typefaces with specific units, the type will still be scaled relative to the viewport when you scale the placed image. (* but see Wrzlprmft answer regarding PDF+LaTex)