Magnetization transfer imaging in progressive multifocal leukoencephalopathy

Progressive multifocal leukoencephalopathy (PML) is a subacute demyelinating disorder of the CNS caused by the JC polyomavirus. PML usually manifests clinically with visual field defects, cognitive impairment, and hemiparesis. There are no systemic features. The CSF is usually normal, although mild pleocytosis, increased protein, and elevated pressure may occur in severe cases. ^[1] Serum and CSF antibody titers for JC virus are unreliable because they are present in nearly 80% of adults. ^[2] MRI typically demonstrates multifocal areas of increased signal intensity on T sub 2-weighted images, with predominant involvement of the subcortical parieto-occipital lobes.

Definitive diagnosis of PML requires brain biopsy. Microscopic examination reveals multiple patches of demyelination with numerous enlarged bizarre astrocytes. Oligodendrocyte nuclear inclusions, which may be basophilic or eosinophilic, are the hallmark pathologic features. Mild chronic inflammation or central necrosis are occasionally noted as well. ^[3] In situ hybridization techniques confirm the presence of JC virus.

Magnetization transfer imaging (MTI), a technique that enables evaluation of the structural integrity of macromolecular structures such as myelin, has been successfully used in animal models of Wallerian degeneration ^[4] and experimental allergic encephalomyelitis. ^[5] Moreover, magnetization transfer ratio (MTR) may assess quantitatively the presence and extent of disease in MS in humans. ^[6,7] MTI may increase our sensitivity to the presence of demyelination and aid in subcategorizing white matter abnormalities ^[8] and may therefore offer potential diagnostic advantages when applied to the demyelinating patches that are characteristic of PML.

We report a patient with biopsy-proven PML who we serially imaged with conventional MRI and MTI throughout the course of her illness. There was a profound and homogeneous reduction in the MTR in the regions of abnormal white matter.

Case report.

A 66-year-old right-handed woman with a history of breast cancer and cutaneous T-cell lymphoma (CTCL) presented with progressive visual loss. Breast carcinoma was treated with resection and subsequent radiation therapy 6 years ago without further recurrence. CTCL was also diagnosed 6 years ago and was treated with photophoresis every 4 weeks. Therapy with electron beam radiation was added to this regimen 1 year before the development of neurologic symptoms.

Over several weeks, she developed trouble with facial recognition. She complained of difficulty with the localization of an object within her visual field, but could easily see the details of the object once it was found.

On examination, she was fully oriented with normal attention, memory, and language function. Visual acuity was 20/100 OD and 20/200 OS. On confrontation visual field examination, she was able to count fingers superiorly but had dense defects inferiorly. Pupils were briskly reactive without an afferent pupillary defect. Funduscopic examination was unremarkable. Extraocular movements were full and conjugate, although tracking was impaired. Opticokinetic nystagmus response was absent. Motor strength and tone, sensation, reflexes, and coordination were normal. Gait was hesitant due to the visual deficits but revealed no significant abnormality. Her general physical examination demonstrated multiple widespread scaly erythematous plaques and a surgical scar on the left breast but was otherwise normal.

T₂-weighted contrast was obtained using a fast spin echo (FSE) technique (GE Medical Systems, Milwaukee, WI). Thirty interleaved slices with 5 mm thickness were acquired using the following parameters: TR 2700 ms, TE(eff) 18/85 ms, ETL 16, Matrix 356 x 192, 1 NEX, and a 22 x 16-cm field of view for a total 4:30 imaging time. On examination of the long-echo-time FSE images (Figure 1), high signal abnormality was detected in the white matter of the parietal and occipital lobes bilaterally. Patchy T₂ signal abnormalities were also observed in the periventricular white matter of the frontal lobes and the right middle cerebellar peduncle. There was no evidence of enhancement with gadolinium. CSF was normal, including cytologic evaluation. The patient was empirically treated with IV methylprednisolone 1,000 mg/d for 3 days without improvement in her vision. She declined brain biopsy at that time.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Axial T₂-weighted MRI demonstrated multifocal areas of high signal abnormality throughout the parieto-occipital white matter.

During the next 3 months, her visual acuity worsened to light perception in both eyes. She was also disoriented, her memory was impaired, and she had difficulty with repetition. MRI was repeated and showed progression of the previous findings and a new focal T₂ signal abnormality in the left centrum semiovale. A brain biopsy specimen from the right occipital lobe revealed oligodendrocytes with hyaline nuclear inclusions and bizarre giant astrocytes characteristic of PML. In addition, there was widespread gliosis and disintegration of white matter with infiltration by macrophages. In situ hybridization probes for JC virus were positive.

She was treated with subcutaneous alpha-interferon three million units daily. She continued to decline during the next 6 months. She developed repetitive behaviors, including frequent counting aloud. She produced little meaningful speech and did not follow commands. She was cortically blind. Strength in the arms was normal, but the legs did not move and there were multiple joint contractures. MRI showed further progression of the white matter abnormalities and worsening cerebral atrophy. The alpha-interferon was discontinued, and the patient was referred for hospice care.

Magnetization transfer ratios and imaging.

Magnetization transfer (MT) data were acquired using an MT-prepared three-dimensional gradient echo pulse sequence provided with the vascular imaging package (GE Medical Systems). Twenty-eight slices with 5 mm thickness were acquired with the following parameters: TR 106 ms, TE 5 ms, flip angle 12 degrees, matrix 256 x 128, 1 NEX, and field of view 22 cm for imaging time of 7:17 per scan. One scan was obtained with the standard MT saturation pulse turned off, and a second scan was obtained with MT saturation. MTR was calculated as the fractional reduction of signal with MT saturation, normalized by the control scan. ^[5] The MTR for six regions of interest (mean +/- SD) was 20.0 +/- 1.0% at initial presentation, 23.1 +/- 3.5% 3 months later, and 18.9 +/- 2.9% approximately 9 months after initial evaluation. There was no significant difference between or among these values (by paired t tests or ANOVA). These ratios were significantly lower than MTR values of 33.7 +/- 1.7% obtained from normal control subjects (p < 0.01, n = 11; McGowan, unpublished data). Data were further displayed as a semiquantitative contour plot (Figure 2), which characterized the extensive abnormalities defined by MTR.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Contour plot of MTR thresholds. (A) Contour lines reveal the extent of disease, demarcating areas with MTR less than 28% (arrows). (B) Contour lines surround the most severely affected regions, those with MTR less than 23% (arrows). CSF spaces are marked with x.

Discussion.

Magnetization transfer appears to be a useful technique for evaluating PML. The images are qualitatively similar to T₂-weighted MRI, suggesting that MTI provides at least a comparable level of sensitivity. MTI has already demonstrated superior sensitivity over conventional MRI for MS. ^[7] Moreover, the quantitative analysis of MTR may offer additional insight into the differential diagnosis of acquired widespread white matter disease. Diagnostic considerations such as periventricular ischemic disease, MS, PML, lymphoma, and radiation encephalopathy might be identified by distinct MTI features, even though they often appear radiographically similar on standard MRI. The lesions of MS tend to have greater variability in the MTR among plaques, possibly as a consequence of their temporal heterogeneity. ^[6-8] Ischemic lesions cause only modest reductions in MTR of approximately 15%, much of which is attributed to increased water content. ^[9,10] Lymphoma and radiation changes have not yet been evaluated with MTI. In our case, the lesions of PML had very low MTR without significant differences between individual foci of disease. The dramatic and homogeneous reduction in the MTR within the abnormal white matter correlated with the underlying severe and synchronous derangement of myelin structure, which was confirmed by biopsy.

Magnetization transfer did not reveal a significant change in MTR during the 9 months of follow-up in this patient, despite continued clinical deterioration and increasing numbers of demyelinating foci on T₂-weighted MRI. The marked decline in MTR may have preceded the clinical and MRI features and was therefore not evident in this case, but this supposition requires additional investigation. Further study of PML is necessary to confirm our findings and may lead to earlier and less invasive antemortem diagnosis. Sequential evaluation with quantitative MTI may be a valuable adjunct in the assessment of response to currently available and newer putative therapeutic interventions.