If you have only exons, the sequence between exons of same gene/transcript is not an Intron ? GTF files usually doesn't have intron coordinates, they provide only exon coords annotated with transcript/gene ids/names.

MAJIQ and rMATS calculates intron retention events but most likely they need a GTF file and hope the one you have should be enough.

If you want a simple calculation, you could follow MAJIQ principle. Divide each intron into equal sized bins, then calculate coverage in each bin, and make sure each bin has atleast certain coverate (5 raw reads or 0.1 tpm etc) and then take the average value of each bin as an intron retention estimate. If a sample doesnt have coverage across the intron, probably its not retained in that sample.

You can use gtf2leafcutter script from leafcutter program to get intron coordinates from a GTF file.