Diagnosis of AIDS-related focal brain lesions: a decision-making analysis based on clinical and neuroradiologic characteristics combined with polymerase chain reaction assays in CSF

Neurology. 1997 Mar;48(3):687-94. doi: 10.1212/wnl.48.3.687.

Abstract

Objective: To identify disease patterns in AIDS-related focal brain lesions (FBL) and to design a decision-making strategy for differential diagnosis.

Design: Prospective study. Probabilities of CNS disorders were calculated using Bayes' theorem according to clinical variables (mass effect at CT or MRI, Toxoplasma serology, anti-Toxoplasma prophylaxis) and to the results of polymerase chain reaction (PCR) assays.

Patients: 136 consecutive HIV-infected patients with a definitive diagnosis of FBL-causing disorder observed from 1991 to 1995 in a single clinical setting.

Interventions: Patients underwent empiric anti-Toxoplasma therapy. After 3 weeks, patients with progressive/stable disease underwent brain biopsy. In 66 patients Epstein-Barr virus (EBV)-DNA, JC virus (JCV)-DNA, and T gondii-DNA amplification was performed by PCR in CSF. Diagnostic criteria were histopathologic examination of bioptic or autoptic tissue specimens for all disorders and complete/partial resolution of FBL after empiric therapy for toxoplasmic encephalitis (TE).

Results: Neuroradiologic characteristics did not discriminate between TE and primary CNS lymphoma (PCNSL). Probability of TE was 0.87 in Toxoplasma-seropositive patients with mass effect who were not receiving anti-Toxoplasma prophylaxis, but only 0.59 if prophylaxis was performed. In seronegative patients with mass effect, the likelihood of PCNSL was 0.74. If EBV-DNA or T gondii-DNA tests were positive, the probability of PCNSL or TE increased to more than 0.96. The absence of T gondii-DNA did not exclude the possibility of a TE diagnosis. Among FBL without mass effect, the probability of progressive multifocal leukoencephalopathy (PML) was 0.81; this increased to 0.99 if JCV-DNA testing was positive. Sensitivity of brain biopsy was 93%, with a perioperative morbidity of 12% and a mortality of 2%.

Conclusions: Due to the low diagnostic capability of clinical variables, PCR amplifications in CSF, especially for EBV-DNA and for JCV-DNA, represent, in most cases, an essential step in the differential diagnosis of AIDS-related FBL. This is particularly true in patients with FBL without mass effect or with mass effect and who are either seronegative or undergoing anti-Toxoplasma prophylaxis. Brain biopsy remains a necessary procedure in EBV-DNA-positive cases and in seronegative patients with FBL displaying a mass effect. Positive JCV-DNA testing may obviate the need for brain biopsy in patients with FBL without mass effect. An advanced diagnostic strategy based on combined clinical criteria and PCR tests may allow rapid and accurate identification of patients for prompt brain biopsy or specific therapy.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • AIDS Dementia Complex / cerebrospinal fluid
  • AIDS Dementia Complex / diagnosis*
  • AIDS-Related Opportunistic Infections / cerebrospinal fluid
  • AIDS-Related Opportunistic Infections / diagnosis*
  • Adult
  • Animals
  • Bayes Theorem
  • Biopsy
  • Cerebrospinal Fluid / microbiology
  • DNA, Protozoan / cerebrospinal fluid
  • DNA, Viral / cerebrospinal fluid
  • Diagnosis, Differential
  • Encephalitis / cerebrospinal fluid
  • Encephalitis / diagnosis*
  • Female
  • Humans
  • Leukoencephalopathy, Progressive Multifocal / cerebrospinal fluid
  • Leukoencephalopathy, Progressive Multifocal / diagnosis*
  • Lymphoma / cerebrospinal fluid
  • Lymphoma / diagnosis*
  • Magnetic Resonance Imaging
  • Male
  • Polymerase Chain Reaction
  • Prospective Studies
  • Regression Analysis
  • Sensitivity and Specificity
  • Serologic Tests
  • Tomography, X-Ray Computed
  • Toxoplasma / isolation & purification
  • Toxoplasmosis, Cerebral / cerebrospinal fluid
  • Toxoplasmosis, Cerebral / diagnosis*

Substances

  • DNA, Protozoan
  • DNA, Viral