Performance of Deep learning models for automatic histopathological grading of Meningiomas: a systematic review and meta-analysis

Mahsa Asadi Anar^1*Parsia Noori Mirtaheri²Matin Akhbari³Farnaz Najafi⁴Hoda Mehrabi⁵Ali Babapour⁶Zahra Rahimian⁷Amirhossein Rigi⁸Saeid Rahbarbaghbani³Hesam Mobaraki³Sanaz Masoumi⁹Danial Nouri⁸Tarlan Mirzohreh¹⁰Seyyed Kiarash Sadat RAfiei⁸Zahra Golkar¹¹Yasaman Asadollah Salmanpour¹²Ali Vesali Mahmoud¹³Mohammad Sadra Gholami Chahkand¹⁴

• ¹College of Medicine, University of Arizona, Tucson, United States
• ²Iran University of Medical Sciences, Tehran, Tehran, Iran
• ³Istanbul Yeni Yüzyıl University, Istanbul, Türkiye
• ⁴Islamic Azad University of Medical Sciences, Tehran, Tehran, Iran
• ⁵Arak University of Medical Sciences, Arak, Markazi, Iran
• ⁶Tabari institute of higher education, Tehran, Alborz, Iran
• ⁷Shiraz University of Medical Sciences, Shiraz, Fars, Iran
• ⁸Shahid Beheshti University of Medical Sciences, Tehran, Tehran, Iran
• ⁹Tehran University of Medical Sciences, Tehran, Tehran, Iran
• ¹⁰Tabriz University of Medical Sciences, Tabriz, Iran
• ¹¹Isfahan University of Medical Sciences, Isfahan, Isfahan, Iran
• ¹²Science and Research Branch, Islamic Azad University, Tehran, Tehran, Iran
• ¹³Buali Sina university, hamedan, Iran
• ¹⁴Golestan University of Medical Sciences, Gorgan, Golestan, Iran

Accurate preoperative grading of meningiomas is crucial for selecting the most suitable treatment strategies and predicting patient outcomes. Traditional MRI-based assessments are often insufficient to distinguish between low-and high-grade meningiomas reliably. Deep learning (DL) models have emerged as promising tools for automated histopathological grading using imaging data. This systematic review and meta-analysis aimed to comprehensively evaluate the diagnostic performance of deep learning (DL) models for meningioma grading.This study was conducted in accordance with the PRISMA-DTA guidelines and was prospectively registered on the Open Science Framework. A systematic search of PubMed, Scopus, and Web of Science was performed up to March 2025. Studies using DL models to classify meningiomas based on imaging data were included. A random-effects meta-analysis was used to pool sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC). A bivariate randomeffects model was used to fit the summary receiver operating characteristic (SROC) curve. Study quality was assessed using the Newcastle-Ottawa Scale, and publication bias was evaluated using Egger's test.Twenty-seven studies involving 13130 patients were included. The pooled sensitivity was 92.31% (95% CI: 92.1-92.52%), specificity 95.3% (95% CI: 95.11-95.48%), and accuracy 97.97% (95% CI: 97.35-97.98%), with an AUC of 0.97 (95% CI: 0.96-0.98). The bivariate SROC curve demonstrated excellent diagnostic performance, characterized by a relatively narrow 95% confidence interval despite moderate to high heterogeneity (I² = 79.7%, p < 0.001).DL models demonstrate high diagnostic accuracy for automatic meningioma grading and could serve as valuable clinical decision-support tools.