Künstliche Intelligenz in der Neuroradiologie

ZUSAMMENFASSUNG

Die Nutzung Künstlicher Intelligenz (KI) in der Neuroradiologie bietet vielversprechende Perspektiven für die Diagnose und Verlaufsbeurteilung neurologischer Erkrankungen. Dabei hat es in den letzten Jahren insbesondere Fortschritte im Bereich der Segmentierung, aber auch der Clinical Decision Support (CDS) Systeme gegeben. Die Vorteile der KI-basierten Bildsegmentierung liegen in ihrer Geschwindigkeit, Genauigkeit und Reproduzierbarkeit im Vergleich zur manuellen Analyse durch Radiologen. Dies ermöglicht eine effizientere Auswertung großer Datenmengen und die Quantifizierung von Gewebestrukturen, z. B. für eine bessere Beurteilung des Therapieverlaufs.

Ein weiterer Entwicklungsfokus von KI-Algorithmen liegt im Bereich der klinischen Entscheidungsunterstützung (CDS). Maschinelles Lernen ermöglicht komplexe medizinische Szenarien zu analysieren und prädiktive Modelle abzuleiten. Klinische Untersuchungen hierzu gibt es beispielsweise in der Notfall- und Schlaganfallbildgebung. Trotz erster positiver Ergebnisse in klinischen Studien bestehen weiterhin Herausforderungen für den klinischen Einsatz von KI-basiertem CDS, v. a. in Bezug auf deren Erklär- und Interpretierbarkeit.

ABSTRACT

The use of artificial intelligence (AI) in neuroradiology offers promising perspectives for the diagnosis and disease course assessment of neurological diseases. In this regard, there has been progress in recent years, particularly in the field of image segmentation, but also in clinical decision support (CDS) systems. The advantages of AI-based image segmentation are its speed, accuracy, and reproducibility compared to manual analysis by radiologists. This allows for more efficient analysis of large amounts of data and enables quantification of tissue volumes, e. g., for more objective tracking of disease progression and better therapy monitoring.

Another development focus of AI algorithms is in CDS. Machine learning enables complex medical scenarios to be analyzed and predictive models to be derived. Clinical research on this exists in emergency and stroke imaging, for example. However, regulatory and explainability challenges remain in implementing AI-based CDS in clinical practice.

Schlüsselwörter

Künstliche Intelligenz - Neuroradiologie - Bildgebung - klinische Entscheidungsunterstützung - Segmentierung

Key words

Artificial intelligence - neuroradiology - imaging - clinical decision support - segmentation

Publication History

Article published online:
04 September 2023

• Literatur

• 1 Opfer R. et al Atlas based brain volumetry: How to distinguish regional volume changes due to biological or physiological effects from inherent noise of the methodology. Magn Reson Imaging 2016; 34: 455-461
• 2 Ronneberger O. et al U-Net: Convolutional Networks for Biomedical Image Segmentation. In: Medical Image Computing and Computer-Assisted Intervention MICCAI 2015 18th International Conference. 2015
• 3 Isensee F. et al nnU-Net: a self-configuring method for deep learning- based biomedical image segmentation. Nat Methods 2021; 18: 203-211
• 4 Antonelli M. et al The Medical Segmentation Decathlon. Nat Commun 2022; 13: 4128
• 5 LeCun Y. et al Deep learning. Nature 2015; 521: 436-444
• 6 Thompson A. et al Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol 2018; 17: 162-173
• 7 Brownlee W. et al Early imaging predictors of long-term outcomes in relapse-onset multiple sclerosis. Brain J Neurol 2019; 142: 2276-2287
• 8 Wattjes M. et al 2021 MAGNIMS-CMSC-NAIMS consensus recommendations on the use of MRI in patients with multiple sclerosis. Lancet Neurol 2021; 20: 653-670
• 9 Schmidt P. et al An automated tool for detection of FLAIR-hyperintense white-matter lesions in Multiple Sclerosis. NeuroImage 2012; 59: 3774-3783
• 10 Brugnara G. et al Automated volumetric assessment with artificial neural networks might enable a more accurate assessment of disease burden in patients with multiple sclerosis. Eur Radiol 2020; 30: 2356-2364
• 11 Li H. et al Fully convolutional network ensembles for white matter hyperintensities segmentation in MR images. NeuroImage 2018; 183: 650-665
• 12 Commowick O. et al Objective Evaluation of Multiple Sclerosis Lesion Segmentation using a Data Management and Processing Infrastructure. Sci Rep 2018; 08: 13650
• 13 Carass A. et al Longitudinal multiple sclerosis lesion segmentation: Resource and challenge. NeuroImage 2017; 148: 77-102
• 14 Eichinger P. et al A novel imaging technique for better detecting new lesions in multiple sclerosis. J Neurol 2017; 264: 1909-1918
• 15 Schmidt P. et al Automated segmentation of changes in FLAIR-hyperintense white matter lesions in multiple sclerosis on serial magnetic resonance imaging. NeuroImage Clin 2019; 23: 101849
• 16 McKinley R. et al Automatic detection of lesion load change in Multiple Sclerosis using convolutional neural networks with segmentation confidence. NeuroImage Clin 2020; 25: 102104
• 17 Wen P. et al Updated response assessment criteria for high-grade gliomas: response assessment in neuro-oncology working group. J Clin Oncol 2010; 28: 1963-1972
• 18 Menze B. et al The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS). IEEE Trans Med Imaging 2015; 34: 1993-2024
• 19 Baid U. et al The RSNA-ASNR-MICCAI BraTS 2021 Benchmark on Brain Tumor Segmentation and Radiogenomic Classification. 2021 DOI: 10.48550/arXiv.2107.02314
• 20 Kofler F. et al Are we using appropriate segmentation metrics? Identifying correlates of human expert perception for CNN training beyond rolling the DICE coefficient. Mach Learn Biomed Imaging 2023; 02: 27-71
• 21 Kickingereder P. et al Automated quantitative tumour response assessment of MRI in neuro-oncology with artificial neural networks: a multicentre, retrospective study. Lancet Oncol 2019; 20: 728-740
• 22 Vollmuth P. et al Artificial intelligence (AI)-based decision support improves reproducibility of tumor response assessment in neuro-oncology: An international multi-reader study. Neuro-Oncol 2023; 25: 533-543
• 23 Kofler F. et al BraTS Toolkit: Translating BraTS Brain Tumor Segmentation Algorithms Into Clinical and Scientific Practice. Front Neurosci 2020; 14: 125
• 24 Buchner J. et al Development and external validation of an MRI-based neural network for brain metastasis segmentation in the AURORA multicenter study. Radiother. Oncol J Eur Soc Ther Radiol Oncol 2023; 178: 109425
• 25 Billot B. et al Robust machine learning segmentation for large-scale analysis of heterogeneous clinical brain MRI datasets. Proc Natl Acad Sci USA 2023; 120: e2216399120
• 26 Dieckmeyer M. et al Effect of MRI acquisition acceleration via compressed sensing and parallel imaging on brain volumetry. Magma NYN 2021; 34: 487-497
• 27 Nakamura K. et al Diurnal fluctuations in brain volume: Statistical analyses of MRI from large populations. NeuroImage 2015; 118: 126-132
• 28 Hedderich D. et al Normative brain volume reports may improve differential diagnosis of dementing neurodegenerative diseases in clinical practice. Eur Radiol 2020; 30: 2821-2829
• 29 Hedderich D. et al Artificial intelligence tools in clinical neuroradiology: essential medico-legal aspects. Neuroradiology 2023; 65: 1091-1099
• 30 Chilamkurthy S. et al Deep learning algorithms for detection of critical findings in head CT scans: a retrospective study. Lancet 2018; 392: 2388-2396
• 31 Murray N. et al Artificial intelligence to diagnose ischemic stroke and identify large vessel occlusions: a systematic review. J Neurointerventional Surg 2020; 12: 156-164
• 32 Matsoukas S. et al Artificial intelligence-assisted software significantly decreases all workflow metrics for large vessel occlusion transfer patients, within a large spoke and hub system. Cerebrovasc Dis Extra 2023; 13: 41-46
• 33 Pati S. et al Federated learning enables big data for rare cancer boundary detection. Nat Commun 2022; 13: 7346
• 34 Korevaar S. et al Failure to Achieve Domain Invariance With Domain Generalization Algorithms: An Analysis in Medical Imaging. IEEE Access 2023; 11: 39351-39372