The reporting of this data set is a bit goofy. It has about 640,000 entries, each of which we are told has a location and a non-zero number of cases (examples were given of an entry with one case and an entry with two cases), but seemingly, though it's now been 11 days since the report about the leak appeard at r100.org and on Foreign Policy, no one has bothered adding up the column of case numbers.
I realize this may be a bit of an undertaking -- the leaked version might be on paper or as an image file, rather than in a spreadsheet program, and there might be multiple entries that refer to the same case that would have to be sorted out. If there aren't then 640,000 is a lower bound on the number of cases the data set contains, but between scanner, OCR software and spreadsheet programs, it ought to be reasonably easy for whoever actually has the leaked data set to give us something better than the number of entries.