The development of a combined clinico-radiomics model for predicting post-operative recurrence in atypical meningiomas: a multicenter study.

PURPOSE: Atypical meningiomas could manifest early recurrence after surgery and even adjuvant radiotherapy. We aimed to construct a clinico-radiomics model to predict post-operative recurrence of atypical meningiomas based on clinicopathological and radiomics features. MATERIALS AND METHODS: The study cohort was comprised of 224 patients from two neurosurgical centers. 164 patients from center I were divided to the training cohort for model development and the testing cohort for internal validation. 60 patients from center II were used for external validation. Clinicopathological characteristics, radiological semantic, and radiomics features were collected. A radiomic signature was comprised of four radiomics features. A clinico-radiomics model combining the radiomics signature and clinical characteristics was constructed to predict the recurrence of atypical meningiomas. RESULTS: 1920 radiomics features were extracted from the T1 Contrast and T2-FLAIR sequences of patients in center I. The radiomics signature was able to differentiate post-operative patients into low-risk and high-risk groups based on tumor recurrence (P < 0.001). A clinic-radiomics model was established by combining age, extent of resection, Ki-67 index, surgical history and the radiomics signature for recurrence prediction in atypical meningiomas. The model achieved a good prediction performance with the integrated AUC of 0.858 (0.802-0.915), 0.781 (0.649-0.912) and 0.840 (0.747-0.933) in the training, internal validation and external validation cohort, respectively. CONCLUSIONS: The present study established a radiomics signature and a clinico-radiomics model with a favorable performance in predicting tumor recurrence for atypical meningiomas.

Distinct clinical outcome of microcystic meningioma as a WHO grade 1 meningioma subtype.

OBJECTIVE: To evaluate the clinicopathological characteristics, radiology, and long-term outcomes of microcystic meningiomas (MM) and compare it with other subtypes of meningiomas managed at a single neurosurgical center. METHODS: A total of 87 consecutive patients who underwent surgical resection and were diagnosed as MM between 2005 and 2016 were enrolled for analysis. Clinicopathological, radiology, and prognostic information was collected and analyzed. Progression free survival (PFS) was compared with 659 patients with other subtypes of WHO grade 1 meningiomas and 167 patients with atypical meningiomas treated during the same period. RESULTS: Fifty six females and 31 males with MM were analyzed. Peri-tumor brain edema was frequent on T2 WI (85%).12 patients (13.8%) experienced tumor progression during the mean follow-up of 101.66 ± 40.92 months. The median PFS was unavailable, and the 5, 10, and 15 year progression-free rates were 96.9%, 84.0%, and 73.9%, respectively. Univariate COX analysis demonstrated skull base location and higher Ki-67 index as significant negative prognostic factors for PFS (P < 0.05); multivariate analysis identified tumor location and Ki-67 index as independent factors (P < 0.01), as well. Of note, the PFS of MM was worse than other WHO grade 1 subtypes (P < 0.001), but better than atypical meningiomas (P < 0.001), and the PFS differences were retained even when the analysis was limited to the patients receiving GTR (P < 0.05). CONCLUSION: The PFS of MM was worse than other WHO grade 1 subtypes and better than atypical meningiomas. Skull base location and higher Ki-67 index were independent negative prognostic factors in MM.

Loss of H3K27me3 expression enriches in recurrent grade 1&2 meningiomas and maintains as a biomarker stratifying progression risk.

PURPOSE: To determine if loss of H3K27me3 could predict higher risk of re-recurrence in recurrent meningiomas. METHODS: A retrospective, single-center cohort study was performed for patients who underwent resection of recurrent grade 1 (N = 132) &2 (N = 32) meningiomas from 2009 to 2013. Association of H3K27me3 staining and clinical parameters was analyzed. Additionally, H3K27me3 staining was performed from 45 patients whose tumors recurred and were resected during the follow-up, to evaluate H3K27me3 change during tumor progression. Survival analysis was performed as well. RESULTS: Loss of H3K27me3 expression was observed in 83 patients, comprising 63 grade 1 (47.7%) and 20 grade 2 patients (62.5%). Both grade 1 (p < 0.001) and grade 2 recurrent meningiomas (p < 0.001) had a higher frequency of H3K27me3 loss, compared to de novo meningiomas. 8 of 27 tumors with retained H3K27me3 lost H3K27me3 during re-recurrence (29.6%), while no gain of H3K27me3 was observed in progressive disease from 18 tumors with H3K27me3 loss. Loss of H3K27me3 expression was associated with an earlier re-recurrence in recurrent meningiomas grade 1 and 2 (p < 0.001), and was an independent prognostic factor for PFS in recurrent grade 1 meningiomas (p = 0.005). CONCLUSION: Compared to primary meningiomas, recurrent meningiomas more predominantly had loss of H3K27me3 expression, and further loss can occur during the progression of recurrent tumors. Our results further demonstrated that loss of H3K27me3 predicted shorter PFS in recurrent grade 1 and grade 2 meningiomas. Our work thus supports clinical testing of H3K27me3 in recurrent meningiomas WHO grade 1 and 2.

Predicting meningioma grades and pathologic marker expression via deep learning.

OBJECTIVES: To establish a deep learning (DL) model for predicting tumor grades and expression of pathologic markers of meningioma. METHODS: A total of 1192 meningioma patients from two centers who underwent surgical resection between September 2018 and December 2021 were retrospectively included. The pathological data and post-contrast T1-weight images for each patient were collected. The patients from institute I were subdivided into training, validation, and testing sets, while the patients from institute II served as the external testing cohort. The fine-tuned ResNet50 model based on transfer learning was adopted to classify WHO grade in the whole cohort and predict Ki-67 index, H3K27me3, and progesterone receptor (PR) status of grade 1 meningiomas. The predictive performance was evaluated by the accuracy and loss curve, confusion matrix, receiver operating characteristic curve (ROC), and area under curve (AUC). RESULTS: The DL prediction model for each label achieved high predictive performance in two cohorts. For WHO grade prediction, the area under the curve (AUC) was 0.966 (95%CI 0.957-0.975) in the internal testing set and 0.669 (95%CI 0.643-0.695) in the external validation cohort. The AUC in predicting Ki-67 index, H3K27me3, and PR status were 0.905 (95%CI 0.895-0.915), 0.773 (95%CI 0.760-0.786), and 0.771 (95%CI 0.750-0.792) in the internal testing set and 0.591 (95%CI 0.562-0.620), 0.658 (95%CI 0.648-0.668), and 0.703 (95%CI 0.674-0.732) in the external validation cohort, respectively. CONCLUSION: DL models can preoperatively predict meningioma grades and pathologic marker expression with favorable predictive performance. CLINICAL RELEVANCE STATEMENT: Our DL model could predict meningioma grades and expression of pathologic markers and identify high-risk patients with WHO grade 1 meningioma, which would suggest a more aggressive operative intervention preoperatively and a more frequent follow-up schedule postoperatively. KEY POINTS: WHO grades and some pathologic markers of meningioma were associated with therapeutic strategies and clinical outcomes. A deep learning-based approach was employed to develop a model for predicting meningioma grades and the expression of pathologic markers. Preoperative prediction of meningioma grades and the expression of pathologic markers was beneficial for clinical decision-making.

Association of frequent NF2 mutations with spinal location predominance and worse outcomes in psammomatous meningiomas.

OBJECTIVE: Psammomatous meningiomas (PMs) are a rare histological subtype of meningioma but are rather frequent in spinal meningiomas. The authors aimed to analyze the incidence, clinical features, molecular alterations, long-term outcomes, and prognostic factors of PMs. METHODS: In total, 151 patients with PMs were included in this study. Clinical characteristics, molecular alterations, and progression-free survival (PFS) were analyzed in PMs. Clinical characteristics were compared between PMs and other WHO grade 1 meningiomas. Targeted sequencing of meningioma-relevant genes was performed to determine the molecular alterations in PMs. RESULTS: PMs accounted for 1.34% of all meningiomas. Clinically, spinal location (p < 0.001) and female predominance (p < 0.001) were statistically significant in PMs when compared with the other grade 1 subtypes. Radiologically, calcification was frequently found in PMs (88.24%). Genetically, NF2 was the most frequently mutated gene in PMs (59.7%), followed by TRAF7 and AKT1. Ten patients experienced recurrence during the long-term follow-up. Multivariate analysis demonstrated that age (p = 0.009), extent of resection (p < 0.001), Ki-67 index (p = 0.007), and NF2 status (p < 0.001) were independent prognostic factors in the cohort of PMs. Interestingly, NF2 mutation was detected in all (48/48) spinal PMs (SPMs) but in only 38.46% (35/91) of cranial PMs (CPMs), revealing a significant difference (p < 0.001). The mean Ki-67 index (p = 0.044) and proportion of PMs with PR-positive expression (p = 0.048) were significantly higher in SPMs than in CPMs. The frequent NF2 mutations are associated with spinal location predominance and worse PFS in PMs. CONCLUSIONS: Female sex and spinal location predominance were statistically significant in PMs. NF2 mutation was an independent predictor for worse PFS of PMs. Of note, NF2 mutation was detected in all SPMs but in only 38.46% of CPMs, revealing a significant difference.

Genetic characterization and mutational profiling of foramen magnum meningiomas: a multi-institutional study.

OBJECTIVE: Foramen magnum (FM) meningiomas pose significant surgical challenges and have high morbidity and mortality rates. This study aimed to investigate the distribution of clinically actionable mutations in FM meningiomas and identify clinical characteristics associated with specific mutational profiles. METHODS: The authors conducted targeted next-generation sequencing of 62 FM meningiomas from three international institutions, covering all relevant meningioma genes (AKT1, KLF4, NF2, POLR2A, PIK3CA, SMO, TERT promoter, and TRAF7). Patients with a radiation-induced meningioma or neurofibromatosis type 2 (NF2) were excluded from the study. Additionally, patient and tumor characteristics, including age, sex, radiological features, and tumor location, were retrospectively collected and evaluated. RESULTS: The study cohort consisted of 46 female and 16 male patients. Clinically significant driver mutations were detected in 58 patients (93.5%). The most commonly observed alteration was TRAF7 mutations (26, 41.9%), followed by AKT1E17K mutations (19, 30.6%). Both mutations were significantly associated with an anterolateral tumor location relative to the brainstem (p = 0.0078). NF2 mutations were present in 11 cases (17.7%) and were associated with posterior tumor location, in contrast to tumors with TRAF7 and AKT1E17K mutations. Other common mutations in FM meningiomas included POLR2A mutations (8, 12.9%; 6 POLR2AQ403K and 2 POLR2AH439_L440del), KLF4K409Q mutations (7, 11.3%), and PIK3CA mutations (4, 6.5%; 2 PIK3CAH1047R and 2 PIK3CAE545K). POLR2A and KLF4 mutations exclusively occurred in female patients and showed no significant association with specific tumor locations. All tumors harboring AKT1E17K and POLR2A mutations displayed meningothelial histology. Ten tumors exhibited intratumoral calcification, which was significantly more frequent in NF2-mutant compared with AKT1-mutant FM meningiomas (p = 0.047). CONCLUSIONS: These findings provide important insights into the molecular genetics and clinicopathological characteristics of FM meningiomas. The identification of specific genetic alterations associated with tumor location, volume, calcification, histology, and sex at diagnosis may have implications for personalized treatment strategies in the future.

Favorable Long-Term Outcomes of Chordoid Meningioma Compared With the Other WHO Grade 2 Meningioma Subtypes.

BACKGROUND: WHO grade 2 meningiomas, including atypical, chordoid, and clear cell subtypes, form a heterogenous group of meningiomas with varying aggressiveness and clinical behavior. OBJECTIVE: To demonstrate the differences of clinical-histopathological characteristics and long-term outcomes among these 3 subtypes. METHODS: A total of 609 consecutive patients diagnosed with WHO grade 2 meningiomas (543 atypical meningiomas [AMs], 36 chordoid meningiomas [CMs], and 30 clear cell meningiomas [CCMs]) from 2010 to 2018 were enrolled in this study. We compared the clinical-histopathological characteristics and long-term outcomes in these 3 subtypes and assessed survival differences among the subtypes. Targeted panel sequencing of meningioma-relevant genes was performed in the cases of CM. RESULTS: The patients with CCM were significantly younger than those with AM ( P < .001) and CM ( P = .016). CMs were more likely to receive gross total resection than AMs and CCMs ( P = .033). The Ki-67 index was lower ( P < .001) while the progesterone receptors-positive rate was higher ( P = .034) in CM than in AM and CCM. Importantly, survival analysis demonstrated that CM had better progression-free survival ( P = .022) and overall survival ( P = .0056) than non-CM tumors. However, the PFS of CM was still worse than WHO grade 1 meningiomas ( P < .001). Alterations in NF2 (20.6%) and KMT2C (26.5%) were associated with poorer PFS in CM ( P = .013 for NF2 ; P = .021 for KMT2C ). CONCLUSION: Patients with CM had better long-term postoperative outcomes than the other WHO grade 2 subtypes. A lower Ki-67 index, higher PR status, higher extent of resection, and lower frequency of NF2 alteration might contribute to favorable clinical outcomes of CM.

Progesterone receptor expression and prediction of benefits of adjuvant radiotherapy in de novo atypical meningiomas after gross-total resection.

OBJECTIVE: Benefits of adjuvant radiotherapy (ART) after gross-total resection (GTR) of de novo atypical meningiomas (AMs) are controversial, and factors predictive of radiotherapy benefits in patients with de novo AMs after GTR are unknown. The authors aimed to evaluate the benefits of ART and explore potential factors sensitizing AMs to ART. METHODS: A total of 231 consecutive patients who were pathologically diagnosed with de novo AMs and treated with GTR (Simpson class I-III resections) from 2010 to 2018 were enrolled in the study. Clinicopathological and prognostic information was collected and analyzed. Univariate and multivariate Cox analyses were used to evaluate prognostic predictors and compare the response to radiotherapy. Propensity score matching (PSM) was used to balance the confounding bias in subgroups. RESULTS: A total of 138 patients (59.74%) received ART. Progesterone receptor (PR) expression was positive in 157 patients (67.97%). During the mean follow-up period of 76.25 months, 65 patients (28.14%) experienced recurrence and 38 (16.45%) died of tumor progression. For disease-specific survival (DSS), ART was a better prognostic factor via univariate (p = 0.003) and multivariate (p = 0.025) analyses. For progression-free survival (PFS), univariate Cox analysis showed that ART improved PFS (p = 0.013), but multivariate analysis did not (p = 0.068). Positive PR expression (p = 0.019), age 53.5 years or younger (p = 0.012), and Ki-67 7.5% or lower (p = 0.025) were independent prognostic predictors for better PFS. In the subcohort analysis, the beneficial impact of ART was observed in the PR-negative cohort (p = 0.002) but not in the PR-positive cohort (p = 0.86). The heterogeneity analysis demonstrated that the PR-negative cohort was more sensitive to ART than the PR-positive cohort (p = 0.036). ART was not found to be associated with better PFS in younger patients (≤ 53.5 years, p = 0.14), older patients (> 53.5 years, p = 0.085), those with a Ki-67 index ≤ 7.5% (p = 0.068), or those with a Ki-67 > 7.5% (p = 0.13). The contrasting effects of ART in the PR-negative versus PR-positive cohorts remained true even after PSM, confirming that PR-negative, but not PR-positive, de novo AMs benefited from ART after GTR. CONCLUSIONS: ART was an independent prognostic factor for DSS of patients with de novo AMs treated with GTR (p = 0.025), but not for PFS (p = 0.068). Negative PR expression was a radiosensitive biomarker on PFS for de novo AM patients after GTR.

Application of MRI-Based Radiomics in Preoperative Prediction of NF2 Alteration in Intracranial Meningiomas.

Purpose: This study aimed to investigate the feasibility of predicting NF2 mutation status based on the MR radiomic analysis in patients with intracranial meningioma. Methods: This retrospective study included 105 patients with meningiomas, including 60 NF2-mutant samples and 45 wild-type samples. Radiomic features were extracted from magnetic resonance imaging scans, including T1-weighted, T2-weighted, and contrast T1-weighted images. Student's t-test and LASSO regression were performed to select the radiomic features. All patients were randomly divided into training and validation cohorts in a 7:3 ratio. Five linear models (RF, SVM, LR, KNN, and xgboost) were trained to predict the NF2 mutational status. Receiver operating characteristic curve and precision-recall analyses were used to evaluate the model performance. Student's t-tests were then used to compare the posterior probabilities of NF2 mut/loss prediction for patients with different NF2 statuses. Results: Nine features had nonzero coefficients in the LASSO regression model. No significant differences was observed in the clinical features. Nine features showed significant differences in patients with different NF2 statuses. Among all machine learning algorithms, SVM showed the best performance. The area under curve and accuracy of the predictive model were 0.85; the F1-score of the precision-recall curve was 0.80. The model risk was assessed by plotting calibration curves. The p-value for the H-L goodness of fit test was 0.411 (p> 0.05), which indicated that the difference between the obtained model and the perfect model was statistically insignificant. The AUC of our model in external validation was 0.83. Conclusion: A combination of radiomic analysis and machine learning showed potential clinical utility in the prediction of preoperative NF2 status. These findings could aid in developing customized neurosurgery plans and meningioma management strategies before postoperative pathology.

m6A Regulator Expression Segregates Meningiomas Into Biologically Distinct Subtypes.

BACKGROUND: Meningiomas are the most common primary intracranial tumors in adults. According to the 2021 World Health Organization (WHO) classification of central nervous system tumors, approximately 80% of meningiomas are WHO grade 1, that is, histopathologically benign, whereas about 20% are WHO grade 2 or grade 3, showing signs of atypia or malignancy. The dysregulation of N6-methylation (m6A) regulators is associated with disorders of diverse critical biological processes in human cancer. This study aimed to explore whether m6A regulator expression was associated with meningioma molecular subtypes and immune infiltration. METHODS: We evaluated the m6A modification patterns of 160 meningioma samples based on 19 m6A regulators and correlated them with immune infiltration characteristics. Novel molecular subtypes were defined based on prognostic hub gene expression. RESULTS: Two meningioma clusters were identified based on the expression of 19 m6A regulators. In cluster 1, 607 differentially expressed genes (DEGs) were upregulated and 519 were downregulated. A total of 1,126 DEGs comprised three gene expression modules characterized by turquoise, blue, and gray. Functional annotation suggested that the turquoise module was involved in Wnt-related and other important cancer-related pathways. We identified 32 hub genes in this module by constructing a protein-protein interaction network. The meningioma samples were divided into two molecular subtypes. EPN1, EXOSC4, H2AX, and MZT2B not only showed significant differences between meningioma molecular subtypes but also had the potential to be the marker genes of specific meningioma subtypes. CONCLUSION: m6A regulator gene expression may be a novel prognostic marker in meningioma.