I looked into their data for glioma (see page 51, table 4). They used 8 total studies. 3 studies showed glioma to be more likely in mobile phone users, 2 showed no difference, and 3 showed glioma less likely in mobile phone users. From this they concluded "probably carcinogenic for humans, in particular related to gliomas"?
This is a long paper and I didnt read it all, but I'm not sure I see how that conclusion is supported.