I think this is the obvious way to go. The only issue is whether to do this in all patients or only Sokal high-risk patients. I'd say all patients, but stratify by Sokal score.
Basically such a trial will put you on a roughly even footing with the other 2nd-line agents. I would expect response rate and tolerability for pona to be somewhat better than the others have reported (although cross-trial comparisons are always hard), but obviously the others will always have a longer track record.
One hard question for such a trial is the dosing. I strongly suspect front line patients would be able to get away with a significantly lower dose than 45mg. That would improve tolerability, which is very significant for wide adoption. Perhaps some sort of flexi-dose schedule might work, starting patients low (perhaps 15mg) and upping the dose if they do not respond quickly.
Going head-to-head with the other 2nd-line drugs is tougher. You can't show efficacy improvements other than with a very large and long trial.
The latest research shows about 1/3 of patients will eventually fail Gleevec, so 2nd-line is a reasonable place to be. And if you can grab some of the high-risk 1st-line, so much the better.