OBJECTIVE : Computerised tomography (CT) scans of 30 patients, admitted with HIV-associated cryptococcal meningoencephalitis (CM), were retrospectively reviewed and the different neuroradiological findings categorised. In addition to the characterisation of the cohort, we looked at whether positive CT scans can indicate negative outcomes when compared with normal imaging. METHODS : We identified all patients admitted with CM to Kalafong Hospital in Pretoria, South Africa, over a 2-year period and selected those who underwent brain CT. Abnormal findings were divided into cryptococcal-related lesions and other pathological findings. Clinical data, as well as laboratory data and information about the outcomes were collected. RESULTS : Thirty-nine (44.8%) out of 87 patients had a CT done during the hospital admission, of which 30 were reviewed and independently reported by the authors. The majority of CT scans were non-contrasted (n = 21). Four patients (13.3%) had normal imaging. Amongst the 26 patients with abnormal CTs, we found 16 brain scans (53.3%) with changes most likely attributed to CM. Dilated Virchow–Robin (VR) spaces, found on eight scans (26.7%), were the most common CT finding related to neurocryptococcosis. Global cerebral atrophy, present in 17 patients (56.7%), was the prevailing generalised abnormality. The mortality of all patients who underwent imaging was similar (33.3%) to the mortality in the total cohort of patients with cryptococcal meningitis (31%). In the group with cryptococcal-related changes on imaging, the mortality was higher (53.3%) than in both groups and a subgroup of five patients with hydrocephalus showed 100% mortality. CONCLUSION : Computerised tomography brain imaging was performed in 44.8% of all patients admitted with CM into our hospital. More than half of the images showed cryptococcal-related pathological findings, of which dilated VR spaces were the most common. Only 13.3% of scans were normal. Mortality was higher in the patients with cryptococcal-related pathology (53.3% vs. 31%), with hydrocephalus being associated with a 100% mortality. No scan in our cohort showed any pathology requiring neurosurgical intervention or contraindicating the procedure of a lumbar puncture.