Susceptibility-Weighted Imaging of Intravascular Lymphoma of the Central Nervous System

Intravascular lymphoma (IVL) is characterized by infiltration and proliferation of malignant B cells within small–blood vessel lumens. The central nervous system (CNS) is the most common primary site, and heterogeneous presentation results in diagnostic delay such that 60% of CNS presentations are diagnosed post mortem.¹ Magnetic resonance imaging (MRI) patterns are consistent with microangiopathy, including infarctlike lesions, white matter lesions, leptomeningeal enhancement, masslike lesions, and hyperintense pontine lesions. However, half of patients with neurologic involvement have not shown specific brain lesions using traditional MRI sequences.² Given the vascular pathophysiology of IVL, susceptibility-weighted imaging (SWI) sequences to detect microhemorrhage may have increased sensitivity; however, their diagnostic utility is poorly characterized. This case series describes clinical and radiographic characteristics of 6 patients with CNS IVL, emphasizing SWI.

We conducted a case series using medical record data from a convenience sample of 6 patients with biopsy-confirmed CNS IVL between June 2013 and February 2020. All patients underwent brain MRIs, including diffusion-weighted imaging, precontrast and postcontrast T1-weighted imaging, T2-weighted imaging, SWI, and fluid-attenuated inversion recovery (FLAIR) imaging. No software was used for data analysis.

Four of the 6 patients were female; the median (IQR) age at diagnosis was 67 (55-72) years. Three patients were diagnosed by brain biopsy, 2 by skin biopsy, and 1 by autopsy. All patients had cerebrospinal fluid analysis; only 1 demonstrated a modest pleocytosis, 3 had unremarkable results of serum flow cytometry, and all had at least 1 negative result on flow cytometry.

Abnormal susceptibility was seen in all patients, occurring in both supratentorial and infratentorial gray and white matter, often punctate or gyriform, inconsistently associated with T2 hyperintensity and/or reduced diffusion (Figure). Abnormal susceptibility ranged in appearance and severity, from innumerable subcortical microhemorrhages (case 1), lobar macrohemorrhage (case 6), and a few punctate foci of microhemorrhage on SWI not appreciable on subsequent gradient echo sequences T2*-weighted imaging (case 2).

All patients demonstrated abnormal white matter T2 or FLAIR hyperintensities. Five patients demonstrated reduced diffusion, predominantly in small-vessel patterns. Four showed abnormal enhancements, often subtle, parenchymal, and involving gray or white matter.

Among 6 patients with biopsy-proven CNS IVL, all patients had abnormal susceptibility identified on SWI and abnormal white matter hyperintensities on T2 or FLAIR. Other MRI sequences less consistently demonstrated abnormalities.

Abnormal SWI findings in patients with CNS IVL are not well characterized, and associations have been limited to single-case reports.^3,4 In this series, areas of abnormal susceptibility did not consistently correspond with T2 hyperintensity or reduced diffusion, which was consistent with prior evidence and suggests microhemorrhage or intravascular thrombosis as a separate and potentially predominant imaging finding.⁴ This theory may partially account for the historical insensitivity of MRI in detecting CNS IVL.² Widespread use of SWI in MRI protocols has increased detection of blood products and could contribute to improved diagnosis of CNS IVL if reliably adopted in its diagnostic evaluation.⁵ The superiority of SWI over gradient echo sequence T2*-weighted imaging in detecting microhemorrhage arises from enhancements, including 3-dimensional technique or multiple-echo time technique, to heighten sensitivity in comparison with gradient echo or T2* sequences. This is consistent with case 2 in our series, which demonstrated abnormal SWI lesions not appreciable on subsequent T2*-weighted sequences.

The small sample size and lack of controls limits this study, preventing estimation of sensitivity and specificity of SWI for IVL. Additionally, some patients were imaged at advanced stages of disease; however, this is typical of CNS IVL diagnosis.

In conclusion, patients with CNS IVL may demonstrate characteristic patterns of microhemorrhage and intravascular thrombosis on SWI. Although SWI abnormalities are seen in many neurologic conditions, their presence in the appropriate clinical context should prompt IVL consideration and subsequent tissue sampling, including blind skin biopsy as an alternative to brain biopsy. Such strategies may facilitate earlier diagnosis and improve outcomes.⁶

Accepted for Publication: October 8, 2021.

Published Online: November 29, 2021. doi:10.1001/jamaneurol.2021.4391

Corresponding Author: Megan B. Richie, MD, Department of Neurology, University of California, San Francisco, 505 Parnassus Ave, Box 0114, San Francisco, CA 94143 (megan.richie@ucsf.edu).

Author Contributions: Dr Richie had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Richie, Guterman.

Acquisition, analysis, or interpretation of data: Richie, Shah, Cha.

Drafting of the manuscript: Richie, Guterman, Cha.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Richie, Shah.

Administrative, technical, or material support: Cha.

Supervision: Richie.

Conflict of Interest Disclosures: Dr Guterman reported funding from the National Institute of Neurological Disorders and Stroke (grant 1K23NS116128-01), the National Institute on Aging (grant 5R01AG056715), and the American Academy of Neurology outside the submitted work; receiving personal compensation from JAMA Neurology and Remo Health outside the submitted work; and receiving personal compensation from Marinus Pharmaceuticals Inc outside the submitted work. Dr Shah reported serving as an unaffiliated neurotrauma consultant for the National Football League and on the board of editors and as a writer of continuing medical education quizzes for JAMA Neurology. No other disclosures were reported.