This is a challenging issue as the data is sparse. First, there is some data to suggest that AYA have a worse prognosis than older patients (Pemmaraju et al. Haematologica. 2012 Jul;97(7):1029-35). The reason for this is unclear but I do not think it has anything to do with biology but rather with adherence. It would be important to work closely with family and patient and discuss this aspect. It also becomes particularly relevant in the setting of duration of therapy. Per the data on adherence by Marin et al., patients who take <90% of their dose have 0% chance of achieving undetectable transcript levels. Missing 10% of the dose is not much, it is only 3 days per month. This is relevant because I believe the treatment may not necessarily be 70 years as the option of treatment discontinuation is slowly evolving. Today it is successful in a minority (~40% treatment-free among ~40% eligible). But there is enough research ongoing to believe this will improve. So I would favor TKI over SCT. Which TKI is difficult to answer as the randomized trials did not include patients this age. But I think it is reasonable to extrapolate in the absence of data. I would thus use either higher-dose imatinib (at least 600mg or, in our hands, 800 mg although we do not know this is better than 5600) or a second generation (and dasatinib might make adherence easier for a young patient because of the schedule of administration).
I hope this helps.