Evaluation of Brain Lesions in HIV-infected Patients for Diagnosis of Primary Central Nervous System Lymphoma

Epstein Barr Virus (EBV)-associated primary central nervous system lymphoma (PCNSL) remains a major problem among AIDS patients. The clinical presentation is often clinically indistinguishable from toxoplasmic encephalitis. The method of choice for establishing the definitive diagnosis is brain biopsy. This procedure can be associated with a significant morbidity and mortality, and therefore less invasive means of diagnosing cerebral mass lesions have been studied. Currently, an accepted standard of care for HIV-infected patients that present with signs and symptoms of focal brain lesions is to empirically treat for toxoplasmic encephalitis. Brain biopsy is often deferred until there is demonstration of lack of clinical response or progression on empiric therapy. As a result, treatment initiation is frequently delayed. During this time it is not unusual for further clinical deterioration to occur before appropriate therapies can be initiated. Frequently, the alternative approaches then become a question of appropriate palliation rather than curative intent therapy. Less invasive diagnostic tests to assist in the diagnosis have been investigated. Based on the finding that essentially 100% of HIV-related PCNSL are EBV-associated, the detection of EBV DNA by PCR amplification in the cerebrospinal fluid (CSF) has demonstrated clinical usefulness in the diagnosis, as has the use of neuroradiologic imaging to detect the malignancy. Prior studies have demonstrated that the use of a combination of neuroradiologic, immunologic, and clinical variables in the workup of focal brain lesions in HIV-infected patients to be quite accurate in identifying patients in need of brain biopsy, but a diagnostic algorithm that incorporates the combination of the most sensitive and specific tests in a timely manner has not yet been explored. This study seeks to evaluate an algorithm for the workup of HIV infected patients with focal brain lesions so as to expedite the diagnosis and subsequent treatment of PCNSL. The goals of the study are to 1) determine the specificity, sensitivity, and positive predictive value of a diagnostic algorithm that entails the use of the combination of EBV detection in the CSF and FDG-PET scanning to diagnose PCNSL; 2) evaluate the time to response to anti-toxoplasmic encephalitis therapy, and 3) evaluate the sensitivities and specificities of FDG-PET and (201)Tl-SPECT scanning in identifying PCNSL. Up to one hundred HIV-infected patients with history of at least one focal brain lesion will be screened for enrollment. All patients will be treated empirically for toxoplasmic encephalitis until an alternative diagnosis is confirmed. All enrolled patients will be treated concurrently with antiretroviral therapy. Patients identified to have PCNSL will be referred to the NCI Treatment of PCNSL Protocol for further treatment if the study is open for enrollment.