Utility of FDG-PETCT and magnetic resonance spectroscopy in differentiating between cerebral lymphoma and non-malignant CNS lesions in HIV-infected patients

Autor: Benjamin Taylor, Saye Khoo, Peter J Julyan, C Hogan, Thomas Westwood, A Bonington, Bernadette M Carrington, Glyn Coutts, Suzanne C Bonington
Rok vydání: 2013
Předmět:
Zdroj: European Journal of Radiology. 82:e374-e379
ISSN: 0720-048X
Popis: Background and purpose In HIV infected patients, MRI cannot reliably differentiate between central nervous system (CNS) lymphoma and non-malignant CNS lesions, particularly cerebral toxoplasmosis (CTOX). This study prospectively investigates the utility of FDG PET-CT and magnetic resonance spectroscopy (MRS) in discriminating CNS lymphoma from non-malignant CNS lesions in HIV infected patients, and assesses the ability of FDG PET-CT to guide the use of early brain biopsy. Methods 10 HIV patients with neurological symptoms and contrast enhancing lesions on MRI were commenced on anti-toxoplasmosis therapy before undergoing FDG PET-CT and MRS. Brain biopsies were sought in those with FDG PET-CT suggestive of CNS lymphoma, and in those with a negative FDG PET-CT scan who failed to respond to therapy. Final diagnosis was based on histology or treatment response. Results Two patients were confirmed to have CNS lymphoma and FDG PET-CT was consistent with this diagnosis in both. Six patients had cerebral toxoplasmosis in all of whom FDG PET-CT was consistent with non-malignant disease. One patient had progressive multifocal leukoencephalopathy (PML), FDG PET-CT was equivocal. One patient had a haemorrhagic brain metastasis and FDG PET-CT wrongly suggested non-malignant disease. MRS was performed successfully in eight subjects: three results were suggestive of CNS lymphoma (one true positive, two false positive), four suggested CTOX (two false negative, two true negative), one scan was equivocal. Conclusion FDG PET-CT correctly identified all cases of CNS lymphoma and CTOX, supporting its use in this situation. MRS was unhelpful in our cohort.
Databáze: OpenAIRE