In treatment resistant patients, it should be considered an option though no where near a first-line.
In treating patients where several medications have failed, some doctors use a structure based algorithm. E.g. use a benzo-based antipsychotic (Zyprexa, Seroquel, Loxitane) vs others such as Abilify (which acts as an D2 partial agonist) vs others such as phenothiazines (e.g. Prolixin)
I've never seen this actually studied to see if such a treatment philosophy has any empiric data to back it up, but anecdotally I have seen it work in some patients. E.g. Prolixin worked and Trilafon worked but nothing else did (both are phenothiazines). Zyprexa worked but didn't completely stabilize the patient. The patient did better with each increased dosage of Zyprexa suggesting that perhaps the Zyprexa would've done more at higher dosages so the patient's 2nd atypical added on is similar in structure (Loxitane). The addition of Loxitane stabilized the patient.
Under this philosophy, Navane could be considered since it is not part of a structured family that is typically given out these days and given if the above structure classes have failed.
Which leads me to mentioning that IMHO, before Clozaril is tried, instead of starting it only after two antipsychotics have failed, to try at least several different structure classes. While Clozaril is effective, it is sometimes very difficult to implement in real world practice because of the frequent lab monitoring. I've seen several patients, for example, that are stabilized on a phenothiazine, and only on a phenothiazine, but several of the newer psychiatrists will not touch this medication because during their residency they were only trained to use atypicals and a small handful of typicals such as Haldol or Thorazine.