It would be epic, New-York-Times-front-page-level misconduct.
Lesser misconduct, more relevant in this discussion perhaps, is selective citation (aka cherry picking). I say “lesser” because some kinds of selective citation are mostly obnoxious (citing friends, failing to cite competitors) while not usually rising to the level of “misconduct.” At our journal and our sister journals, we very rarely (as a matter of policy) require correction when the authors should have cited particular papers but didn’t. More serious misconduct would be selective citation that amounts to misrepresentation of the literature. For example, an author who cited only a single paper in 2004, since shown to be incorrect about X, in support of a claim of X, is not merely omitting citations–they are misrepresenting the literature. Such things are rare at our journal and very few can make it through peer review uncorrected, but that is actual misconduct IMO, subject to correction or retraction.
So, if this author has misrepresented the literature, making it appear to support a claim that is known to be false, they are obligated to correct the error once it has been pointed out. I am not taking a position on whether this has happened.
What is much more of a concern to me is that the author has engaged in public accusation while not engaging in professional correspondence about the science. If that is the case, then the author has, IMO, left scientific discourse and should not be given a prominent platform among scientists. This can of course be fixed, by engaging in professional correspondence, correcting errors, and seeking the company of more-qualified colleagues.