Objective: Our aim was to examine the 201Tl-SPECT scans in AIDS patients with focal CNS lesions to identify those studies with a false-positive or false-negative result to determine any potential pitfalls in interpretation as well as to suggest methods for technique optimization.

Methods: We retrospectively reviewed the charts of 162 AIDS patients with cerebral mass lesions on 201Tl-SPECT studies. One hundred sixty-one patients had CT examinations, of which 50 also had MR studies. One patient had MR imaging without CT. Those patients in whom the diagnosis by 201Tl-SPECT did not correspond with the known pathologic or clinically proved diagnosis were then singled out and their CT, MR, and 201Tl-SPECT studies were reviewed, including blinded interpretation of the 201Tl-SPECT scans alone and alongside the corresponding CT and MR examinations. Studies were examined for lesion morphology, size, location, enhancement pattern, and presence of necrosis. The review of the 201Tl-SPECT studies included both a qualitative approach (subjective analysis of the scans for areas of abnormally increased uptake) and a quantitative approach (comparison of lesion activity versus activity within a reference standard, such as the scalp).

Results: Sensitivity and specificity of 201Tl-SPECT in depicting lymphoma were 100% and 93%, respectively, based on the initial qualitative analysis. Fifty-one patients had positive 201Tl-SPECT results, of whom 43 were determined to have lymphoma (four by biopsy/autopsy, 39 by clinical and radiologic findings). Upon reevaluation with both a quantitative and qualitative approach, those studies initially interpreted as positive in patients without lymphoma (false positives) were found to be negative.

Conclusions: Brain 201Tl-SPECT is an effective study in the diagnosis of CNS lymphoma in AIDS patients. Specificity can be increased by routinely performing a quantitative analysis of all lesions.