Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
P18.10.B WHICH MENINGIOMAS BENEFIT MOST FROM METHYLATION PROFILING?
0
Zitationen
8
Autoren
2023
Jahr
Abstract
Abstract BACKGROUND methylation profiling combined with copy number assay provide a better prognostication than WHO grade alone. Profiling all meningiomas is financially unsustainable in most centres, therefore pathologists and clinicians ask for guidance which meningiomas to analyse. Here, we analysed how the WHO grades resolve into methylation classes and risk strata, to provide guidance on which meningiomas to profile MATERIAL AND METHODS From 2018-2023 a subset of meningiomas diagnosed within our institution, and all referred meningiomas underwent methylation profiling with copy number assay. Using the classifier 12.6 we determined how the 3 CNS WHO grades (in short G1, G2, G3), resolve into the methylation classes (MC) benign, intermediate, or malignant, or the risk strata low, intermediate, and high [PMID 36377252, 34618539] RESULTS Of 1408 locally diagnosed meningiomas, 309 (22%) underwent profiling and 393 meningiomas were referred. Of these 702 meningiomas, 545 had a calibrated score of >0.90. Of 253 G1 meningiomas 231 (91%) allocated to MC benign, and 22 (9%) into MC intermediate. Instead, of the 214 G2 meningiomas, 124 (58%) allocated to MC benign, 77 (36%) to MC intermediate and 13 (6%) into MC malignant. The 17 G3 meningiomas distributed approximately equally into the 3 MC’s. Similarly, when analysing the allocation to integrated risk scores, of the 214 G2 meningiomas, 71 (33 %) were low, 106 (50%) intermediate, and 37 (17 %) high risk. None of the 16 G3 meningiomas had a low risk score, whilst 5 had an intermediate, and 11 a high risk score. When stratifying G2 meningiomas by grading criterium (brain invasion, mitotic count, histological type) there was a similar risk score distribution across the 3 criteria. CONCLUSION G2 meningiomas show the most notable discrepancy between grade and methylation class or risk score, and thus should preferentially undergo methylation profiling. Instead, most G1 meningiomas have a low, and most G3 meningiomas a high risk score.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.339 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.211 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 7.614 Zit.
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.776 Zit.
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)
2018 · 5.478 Zit.