Another ROSAT observation
Photographically it was a challenge: I had to do some serious image editing tricks to pull the trail out of the bright twilight background on the image below (on the unedited image, the trail is visible but very inconspicuous):
These high elevation (near 70 degrees) twilight passes are quickly moving earlier (and too early) in the evening for me: yesterday's was the last one I could expect to realistically observe. Passes at lower elevation (12-14 degrees) in late twilight will become visible for me after tomorrow and might allow me to observe it for a few more days later this week, until these passes move too early as well.
Using Alan Pickup's SatEvo software and the current 10.7cm solar flux, I get a projected decay at October 23. Harro Zimmer, using another model, gets October 24th. These predictions still have an uncertainty of a few days, so expect them to shift over the coming days, amongst others due to changing solar activity.