Comparative analysis of machine learning algorithms for computer-assisted reporting based on fully automated cross-lingual RadLex mappings

• Verfasst von: Maros, Máté E. [VerfasserIn]
• Verfasst von: Cho, Chang Gyu [VerfasserIn]
• Verfasst von: Junge, Andreas Georg [VerfasserIn]
• Verfasst von: Kämpgen, Benedikt [VerfasserIn]
• Verfasst von: Saase, Victor [VerfasserIn]
• Verfasst von: Siegel, Fabian [VerfasserIn]
• Verfasst von: Trinkmann, Frederik [VerfasserIn]
• Verfasst von: Ganslandt, Thomas [VerfasserIn]
• Verfasst von: Groden, Christoph [VerfasserIn]
• Verfasst von: Wenz, Holger [VerfasserIn]
• Titel: Comparative analysis of machine learning algorithms for computer-assisted reporting based on fully automated cross-lingual RadLex mappings
• Verf.angabe: Máté E. Maros, Chang Gyu Cho, Andreas G. Junge, Benedikt Kämpgen, Victor Saase, Fabian Siegel, Frederik Trinkmann, Thomas Ganslandt, Christoph Groden and Holger Wenz
• E-Jahr: 2021
• Jahr: 09 March 2021
• Umfang: 18 S.
• Fussnoten: Gesehen am 12.08.2021
• Titel Quelle: Enthalten in: Scientific reports
• Ort Quelle: [London] : Macmillan Publishers Limited, part of Springer Nature, 2011
• Jahr Quelle: 2021
• Band/Heft Quelle: 11(2021), Artikel-ID 5529, Seite 1-18
• ISSN Quelle: 2045-2322
• DOI: doi:10.1038/s41598-021-85016-9
• Datenträger: Online-Ressource
• Sprache: eng
• K10plus-PPN: 1766546102
• K10plus-PPN:
  Universitätsbibliothek
• Lokale URL UB: Zum Volltext

Abstract

Computer-assisted reporting (CAR) tools were suggested to improve radiology report quality by context-sensitively recommending key imaging biomarkers. However, studies evaluating machine learning (ML) algorithms on cross-lingual ontological (RadLex) mappings for developing embedded CAR algorithms are lacking. Therefore, we compared ML algorithms developed on human expert-annotated features against those developed on fully automated cross-lingual (German to English) RadLex mappings using 206 CT reports of suspected stroke. Target label was whether the Alberta Stroke Programme Early CT Score (ASPECTS) should have been provided (yes/no:154/52). We focused on probabilistic outputs of ML-algorithms including tree-based methods, elastic net, support vector machines (SVMs) and fastText (linear classifier), which were evaluated in the same 5 × fivefold nested cross-validation framework. This allowed for model stacking and classifier rankings. Performance was evaluated using calibration metrics (AUC, brier score, log loss) and -plots. Contextual ML-based assistance recommending ASPECTS was feasible. SVMs showed the highest accuracies both on human-extracted- (87%) and RadLex features (findings:82.5%; impressions:85.4%). FastText achieved the highest accuracy (89.3%) and AUC (92%) on impressions. Boosted trees fitted on findings had the best calibration profile. Our approach provides guidance for choosing ML classifiers for CAR tools in fully automated and language-agnostic fashion using bag-of-RadLex terms on limited expert-labelled training data.