T2/T2-FLAIR mismatch sign as a predictive biomarker for anaplastic astrocytoma genetic profile

Abstract

The aim of the study was to assess T2/T2-FLAIR mismatch phenomenon as a predictor of particular genetic profile in the anaplastic astrocytoma group, including those tumors demonstrating contrast enhancement on MRI.Materials and methods. It is a retrospective study. All MR images were anonymized.MRI analysis. We studied 242 MRIs of patients with anaplastic astrocytomas (AA) and anaplastic oligodendrogliomas (AO) who were surgically treated at Burdenko Neurosurgery Center from 01.01.2017 to 31.12.2019. Among 242 patients we identified 23 (9.5%) whose MRI fulfilled the criteria for T2/FLAIR mismatch sign. The images were studied by 3 experienced neuroradiologist and patients were allocated to the T2/FLAIR mismatch group only upon consensus. Readers evaluated T2WI and FLAIR sequences of each MRI examination. They determined thr following characteristics of the tumours, using a binary scoring system for each: 1) presence or absence of homogeneous signal intensity on T2WI; 2) presence or absence of complete/near-complete hyperintence signal on T2WI, and relatively hypointence on FLAIR except for a hyperintense peripheral rim; 3) margins of lesion sharp or indistinct; 4) presence or absence of peritumoral edema. Discordant results were resolved by consensus with more experienced neuroradiologist.Histological examination and molecular markers. Histological examination was carried out by 2 qua lified pathologists after staining the preparations with hematoxylin and eosin and calculating the index of proliferative activity. The final diagnosis was established on the basis of a combination of morphological and molecular genetic studies. The material for the study was 242 biopsies from patients operated on at Burdenko Neurosurgery Center with anaplastic astrocytomas and anaplastic oligodendrogliomas WHO Grade III, fixed in 10% neutral formalin and embedded in Histomax (Leica) paraffin. Among 23 patients with T2 / FLAIR mismatch in 4 archival cases was performed an immunohistochemical study with Anti-IDH1 R132H (clone H09) (dianova) antibodies to establish the mutational status of the IDH-1 gene; in 19 remaining cases the IDH1 R132H mutation was studied by real-time PCR using self-selected primers and probes. One of the samples showed the wild type IDH1 R132H and it was further investigated by Sanger sequencing to determine the mutational status of the IDH1 and IDH2 genes using direct primers.Determination of 1p / 19q co-deletion was performed by in situ fluorescence hybridization using a ZytoLight Glioma 1p / 19q Probe Set (ZytoVision).Study results. T2/T2-FLAIR mismatch sign was found in 23 patients with AA, in 8 cases tumors demonstrated contrast enhancement, including 3 of them with substantial enhancement. The mean age in the T2/T2-FLAIR mismatch group was lower than in the main group (34 vs. 42.7 years), as well as percentage of patients with contrastenhancing tumors (36.4% vs. 55.8%). The proportion of tumors with substantial contrast enhancement was similar in both groups (37.5% vs. 46.7%).DiscussionS.H. Patel et al. in their pioneer paper demonstrated 12% frequency of T2/FLAIR mismatch sign in the experimental group (125 patients) and 17% in validation group (60 patients). M.P.G. Broen et al. present with a higher rate of the sign – 25%. S. Deguchi et al. have demonstrated even higher frequency of T2/FLAIR mismatch among IDHmutant grade II astrocytomas -45%. Our results revealed 9.5% rate of this sign.The first article declaring less than 100% specificity of T2/FLAIR mismatch sign, after great success and acknowledgement of this predictive diagnostic method, was published by T.A. Juratli et al. 2019: the sign was observed in 73% of patients (versus 12% in Patel’s group). Herewith, 29% of patients with T2/FLAIR mismatch sign demonstrated both IDH-1 mutation and 1p/19q co-deletion. The reason for this phenomenon were less strict inclusion criteria in Juratli’s study. Moreover, authors did not specify the number of patients with contrast-enhancing tumors, although according to R. Jain et al., all these tumors must be excluded from the study. It is noteworthy, that even with these non-strict inclusion criteria, T2/FLAIR mismatch sign has demonstrated 100% predictive value for IDH-1 mutation in Jurartli’s study (none of the patients presented with IDH-wild type tumour). In the meta-analysis performed by Goyal et al. 2019, based on several T2/FLAIR mismatch studies with 746 patients, authors demonstrated high specificity (98.5%) and low sensitivity (33.7%) of this sign for detecting IDH-1+ and 1p/19q- cooperation. The main conclusion of this paper is that T2/FLAIR mismatch sign has high specificity and low sensitivity for detecting IDH-1 mutation, but not 1p/19q co-deletion, although there might be exceptions from this rule.Recent studies devoted to the T2/FLAIR mismatch sign included mostly grade II gliomas and a small amount of AA and AO, emphasizing that those tumors did not demonstrate contrast enhancement M.P.G. Broen, S.H. Patel. Moreover, this characteristic is referred as a necessary qualification for T2/FLAIR mismatch sign, since contrastenhancing parts of the tumor may overlap area of interest and confound MRI signal in other sequences.Results of our study convincingly demonstrate possibility of T2/FLAIR mismatch sign in grade III gliomas with contrast enhancement. We observed this phenomenon in 8 cases out of 22 (36.4%). Moreover, among those8 patients, 3 presented with intensive contrast enhancement. Comparing main group with control one, we revealed, that mean age in the group with T2/FLAIR mismatch sign was lower (34 vs. 42.7). and among 8 patients with contrast enhancement only 1 was older than 40 years. Tumors from the main group demonstrated contrast enhancement significantly on a more frequent basis (55.8% vs. 36.4%), meanwhile rate of intensive contrast enhancement was similar (46.7% vs. 37.5%). Thus, we demonstrated correlation between tumour grade and contrast enhancement and T2/FLAIR mismatch sign.Earlier studies made attempts to predict tumor mutational status upon MRI: for instance, frontal lobe tumours, not spreading to the midline structures more often demonstrate IDH mutation Z.L. Sun, A. Lai. Moreover, it is wellknown fact, that IDH+ gliomas present with more distinct borders, more homogeneous MR-signal and less frequently demonstrate contrast enhancement A. Lai, G. Reyes-Botero. However, all these characteristics are very subjective and only advisory. More explicit results can be revealed by using MR-perfusion P. Kickingereder, MR-spectroscopy (2-hydroxyglutarat) M.I. de la Fuente and PET B. Suchorska 2017. However, abovementioned methods are not so widespread, unlike MRI, which is capable to predict genetic profile (IDH and 1p/19q status) with almost 100% accuracy. Unconditioned weakness of this method is qualitative, not quantitative his nature, sometimes demanding discussion and still not unambiguous. Probably, future achievements in neuroradiology will allow to perform quantitative analysis of MR-signal and formalize T2/FLAIR mismatch sign.It is difficult to overestimate T2/FLAIR mismatch sign for diagnosis, surgery planning and overall treatment strategy. These aspects were disputed for grade II gliomas Sofietti, A.S. Jakola, M.M.J. Wijnenga. Anaplastic tumours possess worse prognosis and this additional information might be of extreme use. T. Kawaguchi et al. evaluated correlation between radical resection and prognosis of treatment for AA tumours with and without IDHmutation: it turned out, that for IDH-negative tumors radical surgery did not significantly affect overall survival (although these groups demonstrated different OS). On the contrary, radical resection significantly affected OS for IDH+ gliomas. Our study has demonstrated capability T2/FLAIR mismatch sign detection for anaplastic gliomas regardless of tumour contrast enhancement.It is still unclear why not all IDH-positive 1p/19q-negative gliomas demonstrate T2/FLAIR mismatch sign. These “exceptions” were documented earlier by S.H. Patel, M.P.G. Broen for grade II gliomas and in our study for grade III gliomas: only 23 out of 26 patients with above-mentioned molecular profile presented with T2/FLAIR mismatch sign. S.H. Patel et al. speculate about activation of pathways by mTOR protein, which is involved in IDH-positive gliomas malignant change H. Wakimoto, but lack of data precludes authors from statistically significant conclusions.Conclusion. Study results confirm the hypothesis of relevance of T2/T2-FLAIR mismatch sign for anaplastic atrocytomas with contrast enhancement on MRI as highly specific biomarker for tumor genetic profile. In some cases information provided by MRI in this group of patients may improve preoperative diagnostic and affect treatment strategy.\xa0