In this study, we primarily evaluated two imaging signs on MRI, T2W hypointense ring and T2-FLAIR mismatch, which have significant value in distinguish LGGs and HGGs, together with IDH mutation status. The assessment of these two signs demonstrated a high consistency among readers (κ = 0.85 and 0.84, respectively). In addition, we constructed multivariate models to predict HGGs and IDHwt using T2W hypointense ring, no-T2-FLAIR mismatch and age. Both models, based on the combination of multiple parameters, displayed excellent prediction ability and outperformed models using a single parameter.

The appearance of the T2W hypointense ring of glioma might vary between different patients, appearing as a thin or thick ring, intact or incomplete loops, just like a superior or inferior arc. Schwartz et al. reported that 69% of gliomas had hypointense borders, 26% of which were rims, and 74% were arcs [18]. Our findings of T2W hypointense ring in gliomas were consistent with their results (69.08%). But the exact causes of this sign with various shapes should be further investigated pathologically. As highly heterogeneous tumors, the extensive genetic variations and microenvironmental biochemistry of gliomas are the underlying causes of the different MRI presentations. It is generally recognized in WHO Grade II and III glioma subtypes (astrocytoma and oligodendroglioma) and secondary glioblastoma, while IDH mutation is not found in any pilocytic astrocytomas of WHO grade I. This indicates that these tumors occur through different mechanisms, and pilocytic astrocytoma rarely undergoes malignant transformation [20, 21]. In this study, in order to eliminate its influence, pilocytic astrocytomas (WHO Grade I) were not included.

When Schwartz et al. compared the position of the enhancing ring on T1WI with the hypointense borders (rims or arcs) on T2WI, they found that the T2 border of most gliomas only partially corresponded to the hyperintense ring [18]. The reason may be that in all intracranial lesions, the circular enhancement represents blood-brain barrier (BBB) destruction to varying degrees, while the T2W hypointense ring may originate from the tumor itself, compressed white matter, granulation tissue, and paramagnetic free radicals produced by macrophages [22]. During clinical work, some patients cannot undergo enhanced MRI examination because of the contraindications, such as severe renal insufficiency, allergic constitution, asthma, etc. In this instance, if we can preliminarily judge the grade of glioma and IDH mutation status based on routine MRI images without contrast agents, it will be an excellent direction for patient treatment and prognosis. In this study, some HGGs showed no enhancement after injection of contrast agent and some gliomas with rosette enhancement were not completely consistent with the T2W hypointense ring, indicating that the T2W hypointense ring of gliomas might be an imaging indicator independent of the enhanced state, with clinical application value.

To our knowledge, there were few studies on evaluating the T2W hypointense ring sign in MRI of glioma to date. This sign is generally located between the parenchyma of the tumor and the edema. We speculate that it may be an indicator of tumor glial hyperplasia and invasion into the peripheral normal brain tissues. At present, the radiological definition of the tumor margins in clinical practice is usually determined by comparing the boundary of enhanced lesions on post-gadolinium T1W images [23, 24]. However, previous studies had found that tumor cells usually existed in peritumoral edema of glioblastoma multiforme, and the actual tumor edge could extend several centimeters beyond the edge of tumor parenchyma detected by microscope image analysis [25]. This study couldn’t directly and intuitively judge the boundary of glioma invasion, but if the T2W hypointense ring was confirmed as tumor glial hyperplasia, it might be possible to use this sign to predict whether glioma was invading around and the direction of invasion.

Another possibility is that the deposition of hemosiderin accounted for the formation of the T2W hypointense ring. A previous study revealed that hypointense rings in necrotic glioblastoma were incomplete, irregular, and commonly found at the inner aspect of the contrast-enhanced border [26]. This study speculated the hypointense rings in glioblastoma might resulted from the random accumulation of hemorrhage products at the edge of the necrotic cavity [26]. Schwartz et al. also reported that T2W hypointense ring sign might correspond to paramagnetic free radicals produced by macrophages, hemosiderin-containing macrophages, or fiber edges [18]. Glioblastoma is a highly invasive tumor with blood vessels rapidly generated to support tumor cells growth. Compared to normal capillaries, these neovasculatures have immature vessel walls and wider endothelial spaces, which determine that the intravascular fluid is more likely to leak outside the vessels, leading to peritumoral edema and hemorrhage [27]. In our study, most gliomas with T2W hypointense ring signs were highly invasive HGGs that could easily invade and cause rupture and bleeding of abundant neovasculatures, subsequently leading to hemosiderin deposition.

The ADC value is negatively correlated with cell proliferation indices [28]. Previous studies confirmed that compared with LGGs, HGGs had a lower ADC value [10, 29]. In this study, we also attempted to analyze the ADC values of the T2W hypointense ring. The results showed that compared with LGGs, HGGs had lower ADC values of the T2W hypointense ring. However, this finding did not reach a significant statistical difference, which might be owing to the small sample size of LGGs with T2W hypointense rings (n = 4). Moreover, if hemosiderin deposition existed within the T2W hypointense ring, this might lead to a large deviation of ADC values due to magnetic susceptibility. On the other hand, most of the T2W hypointense rings in our cases were thin and some rings showed uneven signals, which made it a great challenge to accurately draw ROIs for ADC measurement.

Recently, studies had reported on the signs of T2-FLAIR mismatch [15,16,17, 19, 30]. Patel et al. investigated the T2-FLAIR mismatch marker in two groups of LGG data sets and found this sign could identify IDHmut-Noncodel glioma in both sets with a PPV of 100%, but its NPV was only 54% and 76% [17]. This was further confirmed by the research of Broen et al. [16]. Another meta-analysis, including 12 studies and 1053 patients, had a combination of a specificity of 100%, but a sensitivity of only 42% [30]. At present, the pathophysiological basis of the T2-FLAIR mismatch sign remains unclear. Patel et al. also showed that in IDHmut-Noncodel gliomas with T2-FLAIR mismatch, the protein level in the mTOR pathway was significantly increased [17]. Another possible explanation was that the T2-FLAIR mismatch sign might reflect tumor cell structure [30]. Unlike previous studies, we focused on the “no-T2-FLAIR mismatch” sign to predict HGGs and IDHwt gliomas, and the results were statistically significant.

The following are the limitations of this study. This was a single-center retrospective study and the sample size of patients with LGGs was relatively small. To further validate our findings, a multicenter prospective study should be conducted in the future. Additionally, the nature of the T2W hypointense ring in glioma should be clarified in the future.