Ocular tuberculosis is difficult to diagnose but should be suspected when uveitis fails to respond to inflammation suppressive therapy. Interferon-gamma release assays (IGRAs) represent a substantial help to diagnose suspected ocular tuberculosis especially in non-endemic areas. Indocyanine green angiography (ICGA) is able to detect clinically silent choroiditis that, when associated with a positive IGRA test, should lead the clinician to suspect ocular tuberculosis, warranting specific therapy. The fact that IGRA tests can also react with some atypical strains of mycobacteria is not always known. We report here a case with resistant post-operative inflammation that presented with occult ICGA-detected choroiditis and a positive IGRA test that was most probably due to the non-tuberculous mycobacterium (NTM) Mycobacterium kansasii. A 66 year-old man presented with a resistant cystoid macular oedema (CMO) in his left eye after combined cataract and epiretinal membrane surgery. At entry, his best-corrected visual acuity (BCVA) was 0.5 for far and near OS. Intraocular inflammation measured by laser flare photometry was elevated in the left eye (54.4 ph/ms) and also in the right eye (50.9 ph/ms). Four subTenon's injections of 40 mg of triamcinolone did not produce any substantial improvement. Therefore a complete uveitis work-up was performed. Fluorescein angiography showed CMO OS and ICGA showed numerous hypofluorescent dots and fuzziness of choroidal vessels in both eyes. Among performed laboratory tests, the QuantiFERON®-TB Gold test was positive. After a pulmonological examination disclosing a right upper lobe infiltrate, the patient was started on a triple anti-tuberculous therapy. Bronchial aspirate, obtained during bronchoscopy, was Ziehl-positive and culture grew M. kansasii. Nine months later, BCVA OS increased to 1.0 and flare decreased to 40.2 ph/ms. The CMO OS resolved angiographically and did not recur with a macula still slightly thickened on OCT. Suspected ocular tuberculosis based on clinical findings and a positive IGRA test can, in rare instances, be due to atypical mycobacteria that also produce positive IGRA tests such as M. kansasii, M. szulgai, M. gordonae, M. flavescens and M. marinum. In our case failure to isolate the atypical mycobacterium would not have had negative therapeutic consequences, as M. kansasii is sensitive to the standard anti-tuberculous treatments, which is not the case with other NTMs.