EUVIMEDEUVIMED entdecken
Crossref

Deep learning MGMT promoter methylation status prediction in IDH-wildtype glioblastoma using multiparametric MRI: Benchmark vs. real‑world performance

Augusto Leone, Elena Filimonova, Francesco Carbone, Matteo Zoli, Jamil Rzaev, Mariya Schukina, Alessandro Carretta, Andrea Bianconi, Fabio Cofano, Alberto Morello, Giuseppe Di Perna, Daniele Armocida und 19 weitere

Neurosurgical Review · 16.11.2026 · 49 (1)

DOI: 10.1007/s10143-026-04566-6 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Neurosurgical Review, online erschienen am 16.11.2026. Verfasst von Augusto Leone und 30 weiteren Autor:innen. Beteiligte Einrichtung: Department of Neurosurgery, Städtisches Klinikum Karlsruhe, Karlsruher Neurozentrum, Karlsruhe, Germany. Hauptthemen laut MEDLINE: Glioblastoma, Deep Learning, Brain Neoplasms, DNA Repair Enzymes. Die Arbeit nennt 55 Literaturangaben. Gefördert von: Ministry of Economic Development of the Russian Federation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Preoperative determination of O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status is of major prognostic and predictive relevance in IDH-wildtype glioblastoma, yet currently depends on invasive tissue sampling. This study investigates deep learning–based strategies for MGMT prediction from multiparametric MRIs, with particular emphasis on robustness and generalizability under heterogeneous, multicenter conditions. A total of 949 glioblastoma cases were analyzed, integrating public benchmark data from the Brain Tumor Segmentation Challenge (BraTS) with multi-institutional private cohorts acquired across diverse scanners and imaging protocols. Four complementary experimental settings were designed to compare reproduced literature baselines, private-only training, whole-brain end-to-end models, and segmentation-guided region-of-interest (ROI) pipelines. Tumor delineation was performed using an nnU-Net framework with expert quality control, and multiple three-dimensional convolutional neural network architectures were evaluated. Model performance was highest on the homogeneous BraTS benchmark, where a whole-brain DenseNet121 achieved an area under the receiver operating characteristic curve of 0.80. However, marked performance deterioration was observed when models were evaluated on independent held-out multicenter test set. In contrast, segmentation-guided ROI-based approaches showed a more stable performance profile in this preliminary independent held-out multicenter test. Explicit lesion localization was associated with reduction in false-positive classifications, suggesting a potential benefit for performance stability across heterogeneous datasets, although validation in external cohorts from institutions not represented during training is required. Collectively, these results demonstrate that upstream methodological choices, especially tumor segmentation and ROI definition, remain critical determinants of performance in MRI-based radiogenomic modeling. Future studies should emphasize prospective multicenter validation, domain-shift harmonization, multimodal data integration, and uncertainty-aware deep learning to enhance robustness and interpretability.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

GlioblastomaDeep LearningBrain NeoplasmsDNA Repair EnzymesDNA Modification MethylasesTumor Suppressor ProteinsDNA MethylationMultiparametric Magnetic Resonance ImagingHumansPromoter Regions, GeneticConvolutional Neural NetworksFemaleBenchmarkingIsocitrate DehydrogenaseMaleMiddle Aged

Stichwörter der Autor:innen

Convolutional neural networksDeep learningGlioblastomaIDH-wildtypeMGMT promoter methylationVirtual biopsy

Substanzen

DNA Repair EnzymesDNA Modification MethylasesTumor Suppressor ProteinsMGMT protein, humanIsocitrate Dehydrogenase

Autor:innen und Institutionen

  • Department of Neurosurgery, Städtisches Klinikum Karlsruhe, Karlsruher Neurozentrum, Karlsruhe, Germany.
  • Federal Neurosurgical Center, 132/1 Nemirovicha-Danchenko St, Novosibirsk, 630087, Russia.
  • The Artificial Intelligence Research Center of Novosibirsk State University, 1 Pirogov St, Novosibirsk, 630090, Russia.
  • Department of Neurosurgery, Städtisches Klinikum Karlsruhe, Karlsruher Neurozentrum, Karlsruhe, Germany. francesco.carbone@unifg.it.
  • Division of Neurosurgery, University Hospital of Foggia, Policlinico "Riuniti", Foggia, Italy. francesco.carbone@unifg.it.
  • Department of Biomedical and Neuromotor Sciences (DIBINEM), "Alma Mater Studiorum" University of Bologna, Bologna, Italy.

Förderung

  • Ministry of Economic Development of the Russian Federation

Zitieren

APA

Leone, A., Filimonova, E., Carbone, F., Zoli, M., Rzaev, J., Schukina, M., Carretta, A., Bianconi, A., Cofano, F., Morello, A., Di Perna, G., Armocida, D., Spetzger, U., Roumia, S., Di Napoli, V., Fochi, N. P., Lau, R., Internò, V., Giordano, G., …, & Colamaria, A. (2026). Deep learning MGMT promoter methylation status prediction in IDH-wildtype glioblastoma using multiparametric MRI: Benchmark vs. real‑world performance. Neurosurgical Review, 49(1). https://doi.org/10.1007/s10143-026-04566-6

Vancouver

Leone A, Filimonova E, Carbone F, Zoli M, Rzaev J, Schukina M, et al. Deep learning MGMT promoter methylation status prediction in IDH-wildtype glioblastoma using multiparametric MRI: Benchmark vs. real‑world performance. Neurosurgical Review. 2026;49(1). doi:10.1007/s10143-026-04566-6

BibTeX
@article{Leone2026,
  author = {Leone, Augusto and Filimonova, Elena and Carbone, Francesco and Zoli, Matteo and Rzaev, Jamil and Schukina, Mariya and Carretta, Alessandro and Bianconi, Andrea and Cofano, Fabio and Morello, Alberto and Di Perna, Giuseppe and Armocida, Daniele and Spetzger, Uwe and Roumia, Safwan and Di Napoli, Veronica and Fochi, Nicola Pio and Lau, Ruth and Internò, Valeria and Giordano, Guido and Curcio, Antonello and Rustici, Arianna and Angileri, Flavio and Mazzatenta, Diego and Signorelli, Francesco and Porta, Camillo and Garbossa, Diego and Luu, Minh Sao Khue and Benedichuk, Margaret and Shakoor, Ahsan and Tuchinov, Bair and Colamaria, Antonio},
  title = {Deep learning MGMT promoter methylation status prediction in IDH-wildtype glioblastoma using multiparametric MRI: Benchmark vs. real‑world performance},
  journal = {Neurosurgical Review},
  year = {2026},
  volume = {49},
  number = {1},
  doi = {10.1007/s10143-026-04566-6}
}
Kostenlos starten →