I don't think the answer is as obvious as the other posters make it sound—I can't find any guidelines indicating an explicitly malicious coauthor should be enabled to tank a research study that they are no longer a part of. Imagine a scenario where six authors collaborate on a study that concludes guacamole causes oral cancer. A coauthor, who works for a guacamole company, refuses to participate in the publication process, in an attempt to suppress the results. They are breaking the agreement between authors, so your obligations to them change.
The coauthor has already committed scientific misconduct, with public evidence of their fraud, and has given you no reason to think they wouldn't do it again. Personally, I would absolutely not send the new manuscript to this person without getting a third party involved first. It's not necessary to allow yourself to be repeatedly defrauded by a bad actor while you hope—again, for no reason—that they suddenly begin to behave ethically.
Authorship disputes are common, though this one does seem exceptionally tricky. Is it possible to get your institution involved to settle this? If the coauthor quit the project but is still in school, I would hope there was an official mediation process for situations like this—I know of at least one person who was recommended for expulsion from undergraduate studies for more benign fraud than this. For what it's worth, Wiley, one of the largest academic publishers in the world, recommends "Authorship disputes will often need to be referred to institutions if the authors cannot resolve the dispute themselves," as does the Committee on Publication Ethics (COPE). I imagine this is usually regarding authors who want to be added to a publication, but it's still an authorship dispute.
One potential approach, particularly if your institution doesn't seem particularly interested in intervening, is to communicate (in writing!) with the coauthor about the situation and your desire to keep them as a coauthor. If they respond, great! Hopefully it will be a productive conversation, and they can offer you some reassurance that they won't steal your work. If they do not respond, and you're for some reason unable to contact them some other way, then your problem may actually become a little less tricky: Removing a coauthor who refuses to respond is a more common issue that journal editors may be more willing to deal with once the manuscript gets to them. If you submit the manuscript without them as a coauthor, the journal can then use this correspondence, and the retracted stolen results, to make a decision about the final author list.