The folks who published this paper concluded that it was a false positive, and that the algorithm was picking them up because of the bisphosphonates being given in conjunction with the AI. It was in the paper just further down and hidden a bunch of medical lingo. I hope this eases your concerns.