Seven-tesla phase imaging of acute multiple sclerosis lesions: a new window into the inflammatory process

Abstract

Objective: In multiple sclerosis (MS), accurate, in vivo characterization of dynamic inflammatory pathological changes occurring in newly forming lesions could have major implications for understanding disease pathogenesis and mechanisms of tissue destruction. Here, we investigated the potential of ultrahigh-field magnetic resonance imaging (MRI; 7T), particularly phase imaging combined with dynamic contrast enhancement, to provide new insights in acute MS lesions.

Methods: Sixteen active MS patients were studied at 7T. Noncontrast, high-resolution T2* magnitude and phase scans, T1 scans before/after gadolinium contrast injection, and dynamic contrast-enhanced (DCE) T1 scans were acquired. T2*/phase features and DCE pattern were determined for acute and chronic lesions. When possible, 1-year follow-up 7T MRI was performed.

Results: Of 49 contrast-enhancing lesions, 44 could be analyzed. Centrifugal DCE lesions appeared isointense or hypointense on phase images, whereas centripetal DCE lesions showed thin, hypointense phase rims that clearly colocalized with the initial site of contrast enhancement. This pattern generally disappeared once enhancement resolved. Conversely, in 43 chronic lesions also selected for the presence of hypointense phase rims, the findings were stable over time, and the rims were typically thicker and darker. These considerations suggest different underlying pathological processes in the 2 lesion types.

Interpretation: Ultrahigh-field MRI and, especially, phase contrast, are highly sensitive to tissue changes in acute MS lesions, which differ from the patterns seen in chronic lesions. In acute lesions, the hypointense phase rim reflects the expanding inflammatory edge and may directly correspond to inflammatory byproducts and sequelae of blood-brain barrier opening.

Figure 1

(A) 7T T2*/phase features and dynamic contrast enhancement (DCE) changes in a medial posterior frontal juxtacortical open-ring enhancing lesion in a 36-year-old woman with RRMS (EDSS 3, disease duration 3 years; patient #5). Over time, the leakage of contrast follows a centripetal pattern (DCE #1, DCE #2 and MPRAGE images acquired, respectively, approximately 1, 12, and 25 min after contrast injection). A hypointense rim is clearly visible on phase (the rim thickness is shown with red lines) and T2* images; perilesional edema, outside the rim, is evident only on T2* (red arrow). The fifth panel shows the superimposed phase rim (indicated by white lines) and initial locus of enhancement as detected on the DCE scans (delimited by red lines). (B) Relationship of small blood vessels to a centripetal lesion on phase and T2* images (three contiguous 1.0 mm slices) in a 48-year-old woman with SPMS (EDSS 5.5, disease duration 20 years, patient #7). The central vein (white arrow), which traverses the lesion and is best seen on the middle slice, drained into the cortex (not shown). Prominent peripheral veins (cyan arrows) pass through the lesion and are visualized on both phase and T2* images.

Figure 2

7T T2*/phase features and contrast enhancement dynamics in centripetal (A, B) and centrifugal (C) enhancing lesions. The hypointense phase rim is clearly visible in A (Supplementary Video 1) and B, but whereas it is clearly seen on T2* in A, it is virtually invisible on T2* in B. In (C), a centrifugal lesion (Supplementary Video 4) is subtly but homogeneously hypointense on phase. The area of phase hypointensity, delimited by cyan dashes, is smaller than the area of T2* hyperintensity (indicated by white dashes). In (E), stable phase and T2* features, including a thick rim, in a chronic lesion at baseline and 1.3 years later.

Figure 3

7T T2*/phase features at baseline and approximately one year later in 6 MS lesions, which are identified and detailed further in Table 2. Cyan arrows indicate the lesion of interest. At follow up, the appearance of a new lesion adjacent to the lesion of interest is shown with purple arrows (lesions #4, #5 and #6). Lesion #9 was not visible on baseline T2* and entirely disappeared at follow-up (cyan circles).