eMedicine Specialties > Radiology > Brain/Spine

Progressive Multifocal Leukodystrophy: Imaging

Author: Manohar Aribandi, MD, Academic Chief, Section of Neuroradiology, Department of Radiology, Geisinger Medical Center
Coauthor(s): Anil Kumar Aribandi, MBBS, MD, MRCP, Specialist Registrar, Department of Hematology, Manchester Royal Infirmary, UK
Contributor Information and Disclosures

Updated: Oct 14, 2009

Radiography

Findings

Radiographs have no role in the diagnosis of PML.

Computed Tomography

Findings

CT scans usually show several bilateral, asymmetric hypoattenuating foci of various sizes without mass effect or enhancement. The lesions may involve the periventricular white matter, subcortical white matter, or both. Subcortical U-fiber involvement results in lesions having a lateral scalloped margin that follows the gray matter–white matter junction.

Degree of Confidence

Although lesions may be seen on CT scanning, MRI offers superior sensitivity in the detection and characterization of the lesions. The diagnosis may be suspected with CT, but MRI is needed for a more confident exclusion of other differential considerations.

False Positives/Negatives

Artifacts; chronic, ischemic white-matter changes; and chronic infarcts may mimic true white-matter lesions of PML.

Magnetic Resonance Imaging

Findings

Findings on T1- and T2-weighted and fluid-attenuated inversion recovery images

MRI has far greater sensitivity than other studies in detecting the lesions of PML and in defining their extent of involvement. On T2W images, lesions appear hyperintense and typically involve the periventricular and subcortical white matter, having a characteristic scalloped lateral margin when they involve the subcortical white matter. Lesions are more conspicuously visualized on fluid-attenuated inversion recovery (FLAIR) images, appearing hyperintense against a background of suppressed CSF signal intensity.

Lesions appear hypointense and well demarcated on T1-weighted (T1W) images, though they may be isointense in the initial phase of the disease. Although the disease may involve any part of the brain, lesions typically occur in the parieto-occipital lobes. Lesions appear to start in the subcortical white matter before extending to the periventricular white matter. Mass effect is infrequently described and usually minimal and correlated with shorter survival when seen on initial studies.

Distribution of lesions

A multifocal distribution pattern is seen. This pattern may be unilateral, but more often, it is bilateral and asymmetric. Lesions may start small, but they eventually enlarge and coalesce. Posterior-fossa involvement is common and is seen in up to one third of patients. Synchronous lesions are usually seen in the supratentorial compartment, though in 10% of patients, lesions are confined to the structures of the posterior fossa. PML may also appear to involve the basal ganglia and deep gray-matter nuclei because of white-matter fibers in these structures. When present, involvement of the gray matter is a secondary finding. In rare cases, lesions are single and can occur anywhere in the brain, including the brainstem.

Findings on contrast-enhanced images

The lesions typically do not enhance and do not have mass effect; however, some reports describe lesions with faint peripheral enhancement or diffuse enhancement with mass effect, especially in the early stages. Enhancement could suggest a relatively good immune response and hence an improved prognosis.

Findings on diffusion-weighted images

On diffusion-weighted images (DWI), lesions can show restricted diffusion, though this is uncommon. The extent of abnormal diffusion appears to be correlated with the speed of clinical progression. Areas with DWI abnormality may correspond to areas that are actively infected by the virus at the time of imaging, but this has been uncommonly reported.

Prognostic utility of MRI in PML

Post et al studied the prognostic utility of MRI in PML.¹⁷ None of the MRI variables was predictive of patient survival, with the exception of mass effect, which was usually minimal, infrequent, and associated with decreased survival. Serial MRI studies showed progression of disease in 1-24 weeks, and a more rapid change was seen in many patients in just 9 weeks.

The authors suggested that findings of increasing atrophy, increasing confluence and extent of white-matter lesions, spread of disease across the corpus callosum, and increasing hypointensity of the lesions on follow-up T1W MRIs may be poor prognostic indicators or as failure of response to therapy. Stabilization or a decrease in the size of the lesion, clinical improvement, and loss JC viral detection on CSF PCR testing may indicate a response to therapy. However, improvements on MRI can lag clinical improvement by 2-6 months or even indicate temporary worsening.

In PML patients undergoing HAART, MRIs can show initial worsening in the first few months followed by stabilization and regression by 12 months. HAART consists of a combination of 3 or 4 anti-HIV drugs from the following classes: nucleoside reverse transcriptase inhibitors, non-nucleoside reverse transcriptase inhibitors, and protease inhibitors.

Thurnher et al reported 2 patients with long-term responses to HAART, apparent increased mass effect and enhancement before eventual improvement, and atrophy of the involved brain parenchyma.¹¹ T1W MRIs but not FLAIR MRIs showed hypointensity. Two nonresponders initially had extensive changes, without enhancement or mass effect after the start of therapy. Their T1W MRIs also showed hypointense areas that were hyperintense on FLAIR images.

This increasing hypointensity on T1W MRIs might have been due to increasing demyelination, which appears hyperintense on FLAIR images. This finding might indicate a worsened prognosis. Increased T1W hypointensity in responders was due to gliosis and necrosis seen in burned-out lesions; this finding is hypointense on FLAIR images.

Berger et al¹⁸ found enhancement in 8.9% of short-term survivors in contrast to 50% in long-term survivors. Enhancement suggests an improved immune response and hence an improved prognosis.

In a study by Collazos et al,¹⁹ contrast enhancement was seen on images in patients with PML treated with HAART only when their CD4⁺ count increased. In these patients, the PML lesions changed from enhancing to nonenhancing on follow-up MRI after 3-6 months of therapy. Larger studies than this are needed to clearly identify the MRI changes after therapy and to confirm the utility of MRI in assessing the patient's prognosis and response to therapy.

Magnetic resonance spectroscopy

Evaluation of the PML lesions with magnetic resonance (MR) spectroscopy reveals reduced N -acetylaspartate (NAA) and creatine levels, increased choline levels, and an excess of lipids and sometimes of myo-inositol. In some cases, lactate is present. Cell membrane and myelin breakdown is the presumed cause of elevated choline levels, neuronal loss leads to decreased NAA concentrations, and glial-cells proliferation elevates myo-inositol values. The mild elevation of lactate and lipid levels may be due to the activity of macrophages and the breakdown products of myelin.

The elevation of choline and myo-inositol values is seen in the early phase of the disease. In the later phase all the metabolites are decreased. These metabolic abnormalities are not specific for PML and may be similar in other lesions complicating HIV disease.

Magnetization transfer imaging

PML lesions appear to have strongly reduced magnetization transfer ratios. According to Ernst et al,²⁰ these features may help in distinguishing lesions from white-matter lesions of HIV leukoencephalopathy. Large prospective studies are needed to assess the utility of newer MR techniques, such as MR spectroscopy and magnetization transfer imaging, in the diagnosis and follow-up of PML.

Degree of Confidence

MRI has far greater sensitivity than that of other studies in depicting the lesions of PML and in defining their extent of involvement. Although MRI results may suggest the diagnosis of PML in the appropriate clinical setting, typical features of PML are often nonspecific, and other differential considerations, such as ischemic changes, HIV leukoencephalopathy, and gliotic changes from previous trauma, must be considered. When the lesions of PML are atypically enhancing or when they have mass effect, other lesions, such as those due to toxoplasmosis, lymphoma, or other intracranial masses should be considered.

The presence of multiple pathologies in immunocompromised conditions adds to the complexity of image interpretation. Infection coexisting with other opportunistic infection may be responsible for some atypical manifestations of the disease and for an apparent response to some form of therapy in some patients.

False Positives/Negatives

It is unusual for a normal variant to mimic PML lesions on MRI.

PML must be distinguished from HIV leukoencephalopathy. PML tends to be multifocal, with bilateral, asymmetric, and predominantly subcortical involvement. In comparison, HIV leukoencephalopathy tends to produce lesions that are usually diffuse, bilateral, and symmetric; these predominantly affect the periventricular white matter.

Diffuse cortical atrophy and ventricular dilatation are not predominant findings of PML and may be helpful in distinguishing from HIV leukoencephalopathy, which is usually associated with substantial atrophy. Lesions appear well defined and hypointense on T1W images, whereas they appear isointense and poorly defined in HIV leukoencephalopathy. Clinical correlation may be helpful, as patients with PML have progressive focal motor and sensory neurologic deficits, whereas those with HIV leukoencephalopathy present with global cognitive changes and dementia.

Nuclear Imaging

Findings

Nuclear medicine studies do not play a major role in the diagnosis of PML. However, Port et al described a case of PML in a patient with AIDS in whom MRIs of the lesion showed enhancement and mass effect, with increased thallium-201 uptake on single photon emission CT (SPECT).¹² O'Mally et al²¹ had reported a case of PML with no uptake on²⁰¹ Th SPECT.

Degree of Confidence

On thallium SPECT, photopenic lesions can be caused by any previous insult to the brain parenchyma, and hence, the differential considerations include a wide variety of causes. The uncommon observation of increased radiotracer uptake may be seen with lymphoma and is atypically found in some infections.

False Positives/Negatives

No anatomic variants have been described to mimic the lesions of PML, but cold lesions (eg, arachnoid cyst, porencephalic cyst) can conceivably appear as photopenic lesions similar to the lesions of PML.

Angiography

Findings

Angiography has no role in the diagnosis of PML. However, Nelson et al²² reported angiographic findings in 6 patients with PML. In 4, angiograms showed abnormal parenchymal blush of the affected region in the early to mid arterial phase and persisting into the venous phase. Associated arteriovenous shunting was also present. In only 1 of the 4 patients did MRI demonstrate enhancement.

On pathologic evaluation, intense perivascular inflammatory cellular infiltrates, angiogenesis, and gliosis were found in the patients with angiographic abnormalities, whereas changes were minimal in patients who did not have angiographic abnormalities.

Degree of Confidence

The findings of vascular blush and arteriovenous shunting on angiography are nonspecific and can be seen with infectious, neoplastic, vascular, and ischemic etiologies.

False Positives/Negatives

No known anatomic variants mimic the lesions of PML.

• Search for CME/CE on This Topic »

• Guidelines Issued for Management of Opportunistic Infections Among HIV-Exposed Children