Diagnostic surveillance of high-grade gliomas: towards automated change detection using radiology report classification

DĂ©tails

ID Serval
serval:BIB_4653B6A7BD8C
Type
Actes de conférence (partie): contribution originale à la littérature scientifique, publiée à l'occasion de conférences scientifiques, dans un ouvrage de compte-rendu (proceedings), ou dans l'édition spéciale d'un journal reconnu (conference proceedings).
Collection
Publications
Institution
Titre
Diagnostic surveillance of high-grade gliomas: towards automated change detection using radiology report classification
Titre de la conférence
Machine Learning and Principles and Practice of Knowledge Discovery in Databases. ECML PKDD 2021.
Auteurâž±eâž±s
Di Noto Tommaso, Atat Chirine, Teiga Eduardo Gamito, Hegi Monika, Hottinger Andreas, Bach Cuadra Meritxell, Hagmann Patric, Richiardi Jonas
Editeur
Springer, Cham
ISBN
978-3-030-93732-4
Statut Ă©ditorial
Publié
Date de publication
01/01/2022
Peer-reviewed
Oui
Volume
1525
SĂ©rie
Communications in Computer and Information Science
Pages
423-436
Langue
anglais
Résumé
Natural Language Processing (NLP) on electronic health records (EHRs) can be used to monitor the evolution of pathologies over time to facilitate diagnosis and improve decision-making. In this study, we designed an NLP pipeline to classify Magnetic Resonance Imaging (MRI) radiology reports of patients with high-grade gliomas. Specifically, we aimed to distinguish reports indicating changes in tumors between one examination and the follow-up examination (treatment response/tumor progression versus stability). A total of 164 patients with 361 associated reports were retrieved from routine imaging, and reports were labeled by one radiologist. First, we assessed which embedding is more suitable when working with limited data, in French, from a specific domain. To do so, we compared a classic embedding techniques, TF-IDF, to a neural embedding technique, Doc2Vec, after hyperparameter optimization for both. A random forest classifier was used to classify the reports into stable (unchanged tumor) or unstable (changed tumor). Second, we applied the post-hoc LIME explainability tool to understand the decisions taken by the model. Overall, classification results obtained in repeated 5-fold cross-validation with TF-IDF reached around 89% AUC and were significantly better than those achieved with Doc2Vec (Wilcoxon signed-rank test, 𝑃=0.009). The explainability toolkit run on TF-IDF revealed some interesting patterns: first, words indicating change such as progression were rightfully frequent for reports classified as unstable; similarly, words indicating no change such as not were frequent for reports classified as stable. Lastly, the toolkit discovered misleading words such as T2 which are clearly not directly relevant for the task. All the code used for this study is made available.
Création de la notice
18/02/2022 16:27
DerniĂšre modification de la notice
08/06/2022 5:36
Données d'usage