RT Journal Article
SR Electronic
T1 Image-Based Search in Radiology: Identification of Brain Tumor Subtypes within Databases Using MRI-Based Radiomic Features
JF American Journal of Neuroradiology
JO Am. J. Neuroradiol.
FD American Society of Neuroradiology
DO 10.3174/ajnr.A8805
A1 von Reppert, Marc
A1 Chadha, Saahil
A1 Willms, Klara
A1 Avesta, Arman
A1 Maleki, Nazanin
A1 Zeevi, Tal
A1 Lost, Jan
A1 Tillmanns, Niklas
A1 Jekel, Leon
A1 Merkaj, Sara
A1 Lin, MingDe
A1 Hoffmann, Karl-Titus
A1 Aneja, Sanjay
A1 Aboian, Mariam S.
A1 for the IBSR Consortium
YR 2025
UL http://www.ajnr.org/content/early/2025/06/19/ajnr.A8805.abstract
AB BACKGROUND AND PURPOSE: Existing neuroradiology reference materials do not cover the full range of primary brain tumor presentations, and text-based medical image search engines are limited by the lack of consistent structure in radiology reports. To address this, an image-based search approach is introduced here, leveraging an institutional database to find reference MRIs visually similar to presented query cases.MATERIALS AND METHODS: Two hundred ninety-five patients (mean age and standard deviation, 51 ± 20 years) with primary brain tumors who underwent surgical and/or radiotherapeutic treatment between 2000 and 2021 were included in this retrospective study. Semiautomated convolutional neural network–based tumor segmentation was performed, and radiomic features were extracted. The data set was split into reference and query subsets, and dimensionality reduction was applied to cluster reference cases. Radiomic features extracted from each query case were projected onto the clustered reference cases, and nearest neighbors were retrieved. Retrieval performance was evaluated by using mean average precision at k, and the best-performing dimensionality reduction technique was identified. Expert readers independently rated visual similarity by using a 5-point Likert scale.RESULTS: t-Distributed stochastic neighbor embedding with 6 components was the highest-performing dimensionality reduction technique, with mean average precision at 5 ranging from 78%–100% by tumor type. The top 5 retrieved reference cases showed high visual similarity Likert scores with corresponding query cases (76% ‘similar’ or ‘very similar’).CONCLUSIONS: We introduce an image-based search method for exploring historical MR images of primary brain tumors and fetching reference cases closely resembling queried ones. Assessment involving comparison of tumor types and visual similarity Likert scoring by expert neuroradiologists validates the effectiveness of this method.A/Oastrocytoma and oligodendroglioma WHO CNS grades 2–3CNNconvolutional neural networkG/Aglioblastoma and astrocytoma WHO CNS grade 4ICCintraclass correlation coefficientmAP@kmean average precision at kMENmeningiomaPApilocytic astrocytomaPCAprincipal component analysisPHATEpotential of heat-diffusion for affinity-based trajectory embeddingt-SNEt-distributed stochastic neighbor embeddingT1CET1 contrast-enhancedUMAPuniform manifold approximation and projectionWHOWorld Health Organization