Gastrointestinal bleeding related to Imatinib therapy is rare after so many years . Bleeding is more commonly associated to Dasatinib or in patients with antocoagulation therapy and TKI ( as was this patient case).
This patient was under rivaroxaban treatment and had GI bleeding secondary to antral vascular ectasias , actually controlled after adrenaline treatment. Vascular ectasias is reported only in few people who take rivaroxaban, but in this case as the patient had associated it with Imatinib, perhaps some interaction between them, potentiated rivaroxaban, this is not well reported.
I would consider restarting the same dose after reconfirming no bleeding in a new FGC. I also suggest a complete ferrum profile lab and intense e.v. ferrotherapy, for Hb normalization if its possible. Monitore this bleeding event with frequent hemoglobin test.
Evaluate if definitely needs anticoagulation, if she does, it could be an option to change rivaroxaban to another drug you can measure its effect.
As she has stopped therapy more than 1 month ago, I suggest measuring transcript levels with a RQ-PCR BCR-ABL molecular test so as to confirm molecular status at present time.
I would try with Imatinib again, with close monitoring of hemoglobin to prevent GI bleeding, testing Hb frequenlty , I suggest not changing the TKI in the meantime.
If the event repeats, and the patients had persistent MR4,5 along the last 2-3 years you can consider stopping therapy with monthly molecular monitoring in IS.