000302889 001__ 302889
000302889 005__ 20250720021536.0
000302889 0247_ $$2doi$$a10.1016/j.compbiomed.2025.110640
000302889 0247_ $$2pmid$$apmid:40639009
000302889 0247_ $$2ISSN$$a0010-4825
000302889 0247_ $$2ISSN$$a1879-0534
000302889 0247_ $$2altmetric$$aaltmetric:179347685
000302889 037__ $$aDKFZ-2025-01429
000302889 041__ $$aEnglish
000302889 082__ $$a570
000302889 1001_ $$0P:(DE-He78)e35f6a9bd89b1c66d107df8a2325a758$$aDenner, Stefan$$b0$$eFirst author$$udkfz
000302889 245__ $$aLeveraging foundation models for content-based image retrieval in radiology.
000302889 260__ $$aAmsterdam [u.a.]$$bElsevier Science$$c2025
000302889 3367_ $$2DRIVER$$aarticle
000302889 3367_ $$2DataCite$$aOutput Types/Journal article
000302889 3367_ $$0PUB:(DE-HGF)16$$2PUB:(DE-HGF)$$aJournal Article$$bjournal$$mjournal$$s1752752309_14022
000302889 3367_ $$2BibTeX$$aARTICLE
000302889 3367_ $$2ORCID$$aJOURNAL_ARTICLE
000302889 3367_ $$00$$2EndNote$$aJournal Article
000302889 500__ $$a#EA:E230#LA:E230#
000302889 520__ $$aContent-based image retrieval (CBIR) has the potential to significantly improve diagnostic aid and medical research in radiology. However, current CBIR systems face limitations due to their specialization to certain pathologies, limiting their utility. On the other hand, several vision foundation models have been shown to produce general-purpose visual features. Therefore, in this work, we propose using vision foundation models as powerful and versatile off-the-shelf feature extractors for content-based image retrieval. Our contributions include: (1) benchmarking a diverse set of vision foundation models on an extensive dataset comprising 1.6 million 2D radiological images across four modalities and 161 pathologies; (2) identifying weakly-supervised models, particularly BiomedCLIP, as highly effective, achieving a P@1 of up to 0.594 (P@3: 0.590, P@5: 0.588, P@10: 0.583), comparable to specialized CBIR systems but without additional training; (3) conducting an in-depth analysis of the impact of index size on retrieval performance; (4) evaluating the quality of embedding spaces generated by different models; and (5) investigating specific challenges associated with retrieving anatomical versus pathological structures. Despite these challenges, our research underscores the vast potential of foundation models for CBIR in radiology, proposing a shift towards versatile, general-purpose medical image retrieval systems that do not require specific tuning. Our code, dataset splits and embeddings are publicly available here.
000302889 536__ $$0G:(DE-HGF)POF4-315$$a315 - Bildgebung und Radioonkologie (POF4-315)$$cPOF4-315$$fPOF IV$$x0
000302889 588__ $$aDataset connected to CrossRef, PubMed, , Journals: inrepo02.dkfz.de
000302889 650_7 $$2Other$$aContent-based image retrieval
000302889 650_7 $$2Other$$aFoundation models
000302889 650_7 $$2Other$$aMedical imaging
000302889 650_7 $$2Other$$aSelf-supervised learning
000302889 650_7 $$2Other$$aSupervised learning
000302889 650_7 $$2Other$$aWeakly-supervised learning
000302889 7001_ $$0P:(DE-He78)c1fcef80eab3d1e4fc187faece1a439c$$aZimmerer, David$$b1$$udkfz
000302889 7001_ $$0P:(DE-He78)95f361c74f433d336bfd0a95bc9b0eba$$aBounias, Dimitrios$$b2$$udkfz
000302889 7001_ $$0P:(DE-He78)d52d4217d38d20b78d1bc8014e2b0c35$$aBujotzek, Markus$$b3$$udkfz
000302889 7001_ $$0P:(DE-He78)d2bf7126723ea8f6005ba141ea3c3e2c$$aXiao, Shuhan$$b4$$udkfz
000302889 7001_ $$0P:(DE-He78)166c110dab6977cb48587308422952ff$$aStock, Raphael$$b5$$udkfz
000302889 7001_ $$0P:(DE-He78)4854a5d7f6e812324fd74320396f4178$$aKausch, Lisa$$b6
000302889 7001_ $$0P:(DE-He78)2529b97355581f2d933fcfd7908d9ed4$$aSchader, Philipp$$b7$$udkfz
000302889 7001_ $$aPenzkofer, Tobias$$b8
000302889 7001_ $$0P:(DE-He78)04a0b5a49db132d8f00cee326cb743ca$$aJäger, Paul$$b9
000302889 7001_ $$0P:(DE-He78)33c74005e1ce56f7025c4f6be15321b3$$aMaier-Hein, Klaus$$b10$$eLast author$$udkfz
000302889 773__ $$0PERI:(DE-600)1496984-1$$a10.1016/j.compbiomed.2025.110640$$gVol. 196, no. Pt A, p. 110640 -$$nPt A$$p110640$$tComputers in biology and medicine$$v196$$x0010-4825$$y2025
000302889 909CO $$ooai:inrepo02.dkfz.de:302889$$pVDB
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)e35f6a9bd89b1c66d107df8a2325a758$$aDeutsches Krebsforschungszentrum$$b0$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)c1fcef80eab3d1e4fc187faece1a439c$$aDeutsches Krebsforschungszentrum$$b1$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)95f361c74f433d336bfd0a95bc9b0eba$$aDeutsches Krebsforschungszentrum$$b2$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)d52d4217d38d20b78d1bc8014e2b0c35$$aDeutsches Krebsforschungszentrum$$b3$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)d2bf7126723ea8f6005ba141ea3c3e2c$$aDeutsches Krebsforschungszentrum$$b4$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)166c110dab6977cb48587308422952ff$$aDeutsches Krebsforschungszentrum$$b5$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)4854a5d7f6e812324fd74320396f4178$$aDeutsches Krebsforschungszentrum$$b6$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)2529b97355581f2d933fcfd7908d9ed4$$aDeutsches Krebsforschungszentrum$$b7$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)04a0b5a49db132d8f00cee326cb743ca$$aDeutsches Krebsforschungszentrum$$b9$$kDKFZ
000302889 9101_ $$0I:(DE-588b)2036810-0$$6P:(DE-He78)33c74005e1ce56f7025c4f6be15321b3$$aDeutsches Krebsforschungszentrum$$b10$$kDKFZ
000302889 9131_ $$0G:(DE-HGF)POF4-315$$1G:(DE-HGF)POF4-310$$2G:(DE-HGF)POF4-300$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$aDE-HGF$$bGesundheit$$lKrebsforschung$$vBildgebung und Radioonkologie$$x0
000302889 9141_ $$y2025
000302889 915__ $$0StatID:(DE-HGF)0420$$2StatID$$aNationallizenz$$d2024-12-18$$wger
000302889 915__ $$0StatID:(DE-HGF)0100$$2StatID$$aJCR$$bCOMPUT BIOL MED : 2022$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0200$$2StatID$$aDBCoverage$$bSCOPUS$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0300$$2StatID$$aDBCoverage$$bMedline$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0600$$2StatID$$aDBCoverage$$bEbsco Academic Search$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0030$$2StatID$$aPeer Review$$bASC$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0199$$2StatID$$aDBCoverage$$bClarivate Analytics Master Journal List$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)1050$$2StatID$$aDBCoverage$$bBIOSIS Previews$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0160$$2StatID$$aDBCoverage$$bEssential Science Indicators$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)1030$$2StatID$$aDBCoverage$$bCurrent Contents - Life Sciences$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)1190$$2StatID$$aDBCoverage$$bBiological Abstracts$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0113$$2StatID$$aWoS$$bScience Citation Index Expanded$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)0150$$2StatID$$aDBCoverage$$bWeb of Science Core Collection$$d2024-12-18
000302889 915__ $$0StatID:(DE-HGF)9905$$2StatID$$aIF >= 5$$bCOMPUT BIOL MED : 2022$$d2024-12-18
000302889 9202_ $$0I:(DE-He78)E230-20160331$$kE230$$lE230 Medizinische Bildverarbeitung$$x0
000302889 9201_ $$0I:(DE-He78)E230-20160331$$kE230$$lE230 Medizinische Bildverarbeitung$$x0
000302889 9200_ $$0I:(DE-He78)E230-20160331$$kE230$$lE230 Medizinische Bildverarbeitung$$x0
000302889 980__ $$ajournal
000302889 980__ $$aVDB
000302889 980__ $$aI:(DE-He78)E230-20160331
000302889 980__ $$aUNRESTRICTED