Peritumoral Signal on Postcontrast FLAIR Images: Description and Proposed Biomechanism in Vestibular Schwannomas

DISCUSSION

This study assessed the incidence of peritumoral signal on postcontrast FLAIR images in a cohort of patients with vestibular schwannoma. We found that a previously undescribed peritumoral halo was observed in a high proportion of cases, typically measuring about 1 mm in maximum thickness. This halo is difficult to detect without the use of coregistered/fused images or image subtraction. Asymmetrically elevated postcontrast FLAIR signal was also present in both the fundus and ipsilateral cochlea, confirming the findings of prior studies.

The most logical explanation for the observed findings is that the hyperintense halo represents extratumoral leakage of gadolinium, because it was not demonstrated on any precontrast FLAIR images. The presence of halos on postcontrast FLAIR images, but not postcontrast T1WI, could be due to ≥1 factor. It is possible that leakage is better visualized on FLAIR images because of superior sensitivity to the T1-shortening effect of gadolinium.^15,16 It is also possible that the FLAIR images better depicted gadolinium leakage because they were acquired later in the imaging protocol, allowing interval leaching of contrast into the peritumoral space.

Because the halos were nonuniform in appearance and not circumferential peritumoral rings, we hypothesize that the early gadolinium leakage becomes trapped in a peritumoral space created by the enveloping arachnoid or within cysts and/or adhesions. As seen during microsurgical tumor dissection, VSs are surrounded by a thin, irregular, arachnoid layer that is variably adherent to the tumor surface. Additionally, peritumoral cysts have been well-documented previously,¹⁷ and vestibular schwannomas specifically are known to exist in an inflammatory microenvironment.¹⁸ Thus gadolinium may have become caught in the veil of the peritumoral arachnoid or peritumoral cysts before leaking further into the adjacent CSF. Elevated signal in the ipsilateral fundus and cochlea suggests that additional leakage does occur, possibly in a more diluted manner than the halo effect.

It is not clear why the leaked contrast seems to be locally transported and/or resorbed in the ipsilateral cochlea, even in cases in which the IAC is not occluded. It is possible that even partial occlusion of the IAC (or underlying anatomic constraints of the IAC structure) leads to somewhat stagnant flow of leaked gadolinium into the adjacent CSF, which then, preferentially, is concentrated in the adjacent cochlea rather than diffusing more broadly into the CSF. Altogether, the precise pathophysiology of the observed findings is likely even more complex than what we have outlined, and it is possible that the halo sign is not specific to VS. Further studies using larger cohorts and a greater variety of tumors should help to develop a more comprehensive model.

The concept of VS contrast extravasation is not new. VSs have previously been hypothesized to leak gadolinium into the adjacent structures, most notably the ipsilateral cochlea. Yamazaki et al,¹⁹ for example, found that ipsilateral cochleae have higher signal intensity on postcontrast FLAIR. In that study, the authors demonstrated a convincing increase in intracochlear signal intensity on postcontrast FLAIR images compared with precontrast images, and gadolinium leakage into the IAC fundus. This increased ipsilateral cochlear signal on postcontrast FLAIR signal has been assessed elsewhere.²⁰ The authors contended that the intracochlear signal was suggestive of increased permeability of the blood-labyrinthine barrier in the setting of a VS.¹⁹ Most interesting, Fig 3 in the article of Yamazaki et al also seems to show a peritumoral halo on postcontrast FLAIR images, though neither fused nor subtraction images were shown. The finding was not discussed by the authors.

Pathologically elevated signal is known to be present in the ipsilateral cochlea of patients with VSs on precontrast FLAIR imaging. Unlike the cochlear enhancement observed by Yamazaki et al,¹⁹ this intrinsic T2/FLAIR hyperintensity in the cochlea is attributed to elevated well-documented protein in the perilymph.^21,22 Historically, labyrinthine taps to test for proteinaceous fluid were used as a diagnostic test for VSs.²³ Although the cause of increased perilymphatic protein remains unknown, the mechanism favored by many authors is the blockage of neuroaxonal protein transportation, related to compression of the vestibulocochlear nerve.²¹

Conceivably, the detection of elevated protein seen on FLAIR and gadolinium in the fundus and labyrinth may be related to the tumor obstructing the IAC and creating a relatively stagnant small reservoir of CSF. In this setting, low levels of leached protein or contrast from the tumor may accumulate to sufficient levels to be detected. This possibility may explain the relative gradient seen in FLAIR and gadolinium, with higher concentrations often occurring in the fundus adjacent to the tumor compared with the labyrinth. In contrast, any tumor or contrast leaked medially into the relatively large CPA cistern will naturally dilute and recirculate in the ∼125 mL of CSF in the remaining subarachnoid space, which turns over approximately 3 times daily. Prior literature has identified elevated levels of circulating CSF protein in individuals with small vestibular schwannomas and associated communicating hydrocephalus, implicating protein sludging as a mechanism for reduced CSF resorption.²⁴

Finally, a very recent publication proposed that peritumoral “rim” enhancement on FLAIR could help differentiate meningiomas from malignant dural-based tumors.²⁵ The rim of peritumoral contrast shown in that study is very similar to the halo we describe in the current findings. The authors of that study opined that the observed findings were due to the dual blood supply (meningeal and pial) of meningiomas. This supply led to relative intratumoral signal loss on FLAIR (due to T2 shortening in the setting of high gadolinium concentrations) and relatively hyperintense signal in the pial-supplied tumor capsule. The results of the current study, however, seem to refute the findings of this recent publication. Most important, vestibular schwannomas are nonencapsulated tumors and do not have a dual blood supply.²⁶ Also, as previously stated, the halos observed in the current study are irregular and do not appear to conform to a smooth capsule. Last, enhancement of a tumor capsule would not explain the more diffuse gradient that seems to occur in our cohort, in which gadolinium appears to accumulate in the fundus and cochlea. Together, we believe that these findings are evidence of gadolinium leakage, as stated above.

The results of this current study leave many questions unanswered, and future studies will be needed for further exploration. For example, the incidence and/or appearance of peritumoral halos may vary among patients with sporadic VSs compared with VSs related to neurofibromatosis type 2, or in untreated-versus-treated tumors. It may also be useful to determine whether the presence of a peritumoral halo predicts the stability or growth of a tumor. At present, these topics are outside of the scope of the current study.

This study has several limitations, including being a retrospective review. The goal of defining the radiologic finding of a peritumoral halo is somewhat limited in scope, and the results of this study are preliminary and based on a small cohort. Additional studies using large sample sizes will be needed to confirm these findings, as well as to assess the clinical relevance. In addition, many of the VSs analyzed in this study were relatively small. Future studies will also be needed to address whether tumor size has an impact on the thickness and/or appearance of a peritumoral halo. Also, the precontrast FLAIR images that were available for review were part of separate MR imaging examinations. Thus, no precontrast FLAIR imaging was available for direct comparison with postcontrast FLAIR in this study. Finally, the proposed mechanisms for the observed peritumoral halo are hypothetical and based on limited prior studies. It is possible that the peritumoral signal on postcontrast FLAIR images simply represents an imaging artifact. Also, because the average halo thickness was around 1 mm, the precision of the reported measurements may be questionable, given the pixel size on standard PACS calipers.