Treatment with highly active antiviral drugs (HAART) has resulted in an overall decline in morbidity and mortality among HIV-infected patients with advanced disease . The widespread use of prophylactic regimens has changed the disease pattern and decreased the incidence of opportunistic infections . Vago et al.  reported on a large autopsy series that showed a significant reduction in the frequency of HIV-related central nervous system lesions in AIDS patients since the beginning of the HAART era.
In patients with HIV, Toxoplasma gondii is the most frequent infectious cause of focal brain lesions . Toxoplasmic encephalitis has a subacute onset with focal neurologic abnormalities frequently accompanied by headache, altered mental status, and fever .
Diffuse toxoplasmosis encephalitis should be considered in patients with anti-T gondii immunoglobulin G (IgG) antibodies and CD4 T-cell counts of <100/μL who present with unexplained neurologic disease.
Since the introduction of HAART, the incidence of toxoplasmic encephalitis has remained stable or decreased, CNS lymphoma has dramatically declined, PML has slightly increased,  and the incidence of HIV encephalopathy has increased . Although combination therapies have decreased overall mortality and the prevalence of CNS opportunistic infections, these therapies may be less effective for preventing the direct effects of HIV-1 on the brain. One of our patients presented with histologic features of AIDS encephalopathy. The majority of patients in our cohort were men with a median age of 40, consistent with other studies [15, 16]. These findings are compatible with our data , which is reported to be increasing in prevalence in the HAART era .
Thallium SPECT and PET scanning in the diagnosis of lymphoma and PCR analysis of the cerebrospinal fluid for the diagnosis of toxoplasmosis and JC virus have led to a shift in the use invasive diagnostic techniques (brain biopsy) to noninvasive diagnostic methods . The patient in our series that was diagnosed with lymphoma had a solitary enhancing lesion. Although CNS lymphoma generally tends to enhance, atypical characteristics may be observed on MR images of immunocompromised patients . Toxoplasma and lymphoma may have similar appearance on MR imaging. Erdag et al.  noted that lymphoma usually appears as a solitary lesion, unlike toxoplasmosis, which appears as a multifocal lesion. Only half of our patients with toxoplasmosis, however, had multifocal disease, similar to a previous report .
The majority of patients in our cohort that were diagnosed with PML had nonenhancing lesions, consistent with previous studies [18, 19]. Other studies, however, report enhancement on MR images of PML patients that may reach 42% , and the development of a mass effect and temporary enhancement on MR images in the early phase of treatment might represent positive predictive factors for prolonged survival .
A literature review revealed several descriptions of stereotactic brain biopsy in AIDS patients. The majority of these were reports on patients treated in the pre-HAART era [7–9, 16, 20–26]. In a recent study, Rosenow et al.  divided a large cohort of 246 patients that underwent brain biopsy into two diagnostic groups, before and after the HAART era. The authors concluded that the introduction of HAART led to a steep decrease in the number of biopsies performed annually. Biopsy led to an objective and definitive diagnosis in 96% of patients (48/50), in the patients in the HAART era. Definitive diagnoses were obtained for 91.3% of patients (179/196) in the pre-HAART era. The overall diagnostic yield of biopsy was 92.3% of the total patient cohort. Rosenow et al.  suggested that AIDS patients with intra-cerebral lesions should undergo toxoplasmosis titers, and those with negative toxoplasmosis titers should undergo early brain biopsy. In our study, the majority of the patients were treated in the HAART era, which may correlate with the relatively small cohort of patients that underwent stereotactic brain biopsy. The diagnostic yield of the biopsy was high, although lower than previously reported .
A total of 19 biopsies were performed in 16 patients; one procedure was aborted due to bleeding and three biopsies were nondiagnostic (15%). Significantly, two patients underwent repeated biopsies that were conclusive. Diagnoses were eventually obtained for 93.7% of patients (15/16). Based on the biopsy findings, the treatment modality was modified in 12 patients and led to clinical and radiologic improvement in 8 patients. Modification of the treatment following brain biopsy has also been reported [1, 8].
Hemorrhagic complications of biopsy are reported [16, 20, 23, 25]. The incidence of hemorrhage in our small cohort of patients was higher than that reported in a non-HIV group of patients treated in our department . Although the coagulation profile in our patients was within normal limits, there may have been undetectable coagulopathy in this group of patients. Anticoagulopathy regimens have been suggested by others  and pre-biopsy correction of thrombopenia less than 100,000/ml is suggested by Rosenow et al. . A larger cohort of patients is required to establish more definitive conclusions.