Sociopath
is a clinical term, it's just not in the DSM because it's not, per se, a disease or disorder. Being a sociopath doesn't, by itself, cause impairment as defined by the DSM. But that doesn't mean it's not a clinical term. Ask any psychiatrist or psychologist what the defining characteristics are of a sociopath and they could immediately recite the Hare Psycopathy Checklist.
And there are plenty of psychiatrists who treat sociopathy as such. It can cause the sort of impairment cognizable by the DSM, especially as children when the measure of impairment is different than as an adult (e.g. "unruly child", something that would be non-sense once they reach the age of majority). And in those cases while it might be formally treated under the rubric of Anti-Social Personality Disorder, clinicians understand that treatment (in as much as its treatable at all) requires understanding sociopaths as a distinct class with a very distinct pathology.
Sociopath in the vernacular does often connote a serial killer. Fair enough. But there's plenty of literature, for professionals and laymen, that discusses sociopathy in clinical, well-defined terms.
FWIW, AFAIU I don't think there's any real debate about whether sociopathy exists as a concrete human phenomenon. The debate is largely how big the umbrella--do we only include those who are "incurable" (i.e. neurobiologically lack the capacity for empathy, either at birth or from a combination of nature and nurture), or those which behave as-if they lack such capacity, but where the capacity is simply suppressed for other reasons. And of course, this begs the question of what, exactly, is empathy. Like sociopath, I think most agree that empathy is a thing (and much more concretely than, say, "love"). But is "cognitive empathy" a thing? And given these questions, how accurate can we ever hope to get in particular cases?
FWIW, here's a good, recent radio interview discussing a special clinic in Wisconsin that has seem good results treating sociopaths--not in making them stop being sociopaths who harm people, but at least in minimizing their harm to a level below that which is considered criminal.
https://www.npr.org/2017/05/24/529893128/scientists-develop-...
Of course, most sociopaths are already non-violent. Empathy seems to be but one evolutionary mechanism for checking violence and dampening violent tendencies. It's not the sole mechanism and might not even be the most important.