WRT the discrepancy: Keep in mind that we are talking about a group ability and group achievement test, so any interpretive discussion is in the context of the imprecision inherent in using this kind of group measure for diagnostic purposes (i.e., not recommended). That being said, this is just under a 1.5 SD difference between his ability and language achievement, which approaches significance as a discrepancy. (Math is pretty much in line with the ability measure.) My first question would be, do you see this IRL, as well (that his math is stronger than his reading/writing)? If this appears to represent a clinically significant difference (not just a fluke of testing), then I would be inclined to view it as more "real" than just an artifact of lack of exposure to higher-level instruction, mainly because it is in the language area, rather than the math area. Unlike for mathematics, most academically-advanced students are able to continue developing their language and vocabulary through outside reading, even when not formally instructed in their ZPD. So a gap in the language area suggests more of a need for follow-up than a gap in math. It is not a huge gap (again, assuming it is real), and is certainly above the mean.
And, btw, I would not describe all of his ITBS results as quite mediocre. 90th %ile in math is not too shabby (+1.3 SD).
FYI, this old study using the WAIS-III found negligible correlation between ITBS scores and FSIQ:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3666826/Interesting Belin-Blank presentation on use of CogAT and ITBS in gifted selection process (cut & paste into search engine). Doesn't recommend using the CogAT composite score, but the Verbal paired with Reading, and the QN paired with Mathematics:
faculty.education.uiowa.edu/docs/dlohman/icn_full_day_using_cogat.ppt?
