Performance of AI approaches for COVID-19 diagnosis using chest CT scans: the impact of architecture and dataset = Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und Datensätzen
AI is emerging as a promising tool for diagnosing COVID-19 based on chest CT scans. The aim of this study was the comparison of AI models for COVID-19 diagnosis. Therefore, we: (1) trained three distinct AI models for classifying COVID-19 and non-COVID-19 pneumonia (nCP) using a large, clinically re...
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , |
|---|---|
| Dokumenttyp: | Article (Journal) |
| Sprache: | Englisch |
| Veröffentlicht: |
29. April 2025
|
| In: |
RöFo
Year: 2025, Pages: [1-14] |
| ISSN: | 1438-9010 |
| DOI: | 10.1055/a-2577-3928 |
| Online-Zugang: | Verlag, kostenfrei, Volltext: https://doi.org/10.1055/a-2577-3928 Verlag, kostenfrei, Volltext: http://www.thieme-connect.de/DOI/DOI?10.1055/a-2577-3928 
|
| Verfasserangaben: | Astha Jaiswal, Philipp Fervers, Fanyang Meng, Huimao Zhang, Dorottya Móré, Athanasios Giannakis, Jasmin Wailzer, Andreas Michael Bucher, David Maintz, Jonathan Kottlors, Rahil Shahzad, Thorsten Persigehl |
MARC
| LEADER | 00000naa a22000002c 4500 | ||
|---|---|---|---|
| 001 | 1940056411 | ||
| 003 | DE-627 | ||
| 005 | 20251103113618.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 251103s2025 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1055/a-2577-3928 |2 doi | |
| 035 | |a (DE-627)1940056411 | ||
| 035 | |a (DE-599)KXP1940056411 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 33 |2 sdnb | ||
| 100 | 1 | |a Jaiswal, Astha |e VerfasserIn |0 (DE-588)1113020458 |0 (DE-627)867024933 |0 (DE-576)476817447 |4 aut | |
| 245 | 1 | 0 | |a Performance of AI approaches for COVID-19 diagnosis using chest CT scans |b the impact of architecture and dataset = Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und Datensätzen |c Astha Jaiswal, Philipp Fervers, Fanyang Meng, Huimao Zhang, Dorottya Móré, Athanasios Giannakis, Jasmin Wailzer, Andreas Michael Bucher, David Maintz, Jonathan Kottlors, Rahil Shahzad, Thorsten Persigehl |
| 246 | 3 | 1 | |a Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und Datensätzen |
| 264 | 1 | |c 29. April 2025 | |
| 300 | |b Illustrationen | ||
| 300 | |a 14 | ||
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Gesehen am 03.11.2025 | ||
| 520 | |a AI is emerging as a promising tool for diagnosing COVID-19 based on chest CT scans. The aim of this study was the comparison of AI models for COVID-19 diagnosis. Therefore, we: (1) trained three distinct AI models for classifying COVID-19 and non-COVID-19 pneumonia (nCP) using a large, clinically relevant CT dataset, (2) evaluated the models’ performance using an independent test set, and (3) compared the models both algorithmically and experimentally. In this multicenter multi-vendor study, we collected n=1591 chest CT scans of COVID-19 (n=762) and nCP (n=829) patients from China and Germany. In Germany, the data was collected from three RACOON sites. We trained and validated three COVID-19 AI models with different architectures: COVNet based on 2D-CNN, DeCoVnet based on 3D-CNN, and AD3D-MIL based on 3D-CNN with attention module. 991 CT scans were used for training the AI models using 5-fold cross-validation. 600 CT scans from 6 different centers were used for independent testing. The models’ performance was evaluated using accuracy (Acc), sensitivity (Se), and specificity (Sp). The average validation accuracy of the COVNet, DeCoVnet, and AD3D-MIL models over the 5 folds was 80.9%, 82.0%, and 84.3%, respectively. On the independent test set with n=600 CT scans, COVNet yielded Acc=76.6%, Se=67.8%, Sp=85.7%; DeCoVnet provided Acc=75.1%, Se=61.2%, Sp=89.7%; and AD3D-MIL achieved Acc=73.9%, Se=57.7%, Sp=90.8%. The classification performance of the evaluated AI models is highly dependent on the training data rather than the architecture itself. Our results demonstrate a high specificity and moderate sensitivity. The AI classification models should not be used unsupervised but could potentially assist radiologists in COVID-19 and nCP identification. | ||
| 700 | 1 | |a Fervers, Philipp |e VerfasserIn |4 aut | |
| 700 | 1 | |a Meng, Fanyang |e VerfasserIn |4 aut | |
| 700 | 1 | |a Zhang, Huimao |e VerfasserIn |4 aut | |
| 700 | 1 | |a Móré, Dorottya |d 1991- |e VerfasserIn |0 (DE-588)1247638219 |0 (DE-627)1782202951 |4 aut | |
| 700 | 1 | |a Giannakis, Athanasios |e VerfasserIn |0 (DE-588)1227892640 |0 (DE-627)1749093278 |4 aut | |
| 700 | 1 | |a Wailzer, Jasmin |e VerfasserIn |4 aut | |
| 700 | 1 | |a Bucher, Andreas Michael |e VerfasserIn |4 aut | |
| 700 | 1 | |a Maintz, David |e VerfasserIn |4 aut | |
| 700 | 1 | |a Kottlors, Jonathan |e VerfasserIn |4 aut | |
| 700 | 1 | |a Shahzad, Rahil |e VerfasserIn |4 aut | |
| 700 | 1 | |a Persigehl, Thorsten |d 1974- |e VerfasserIn |0 (DE-588)124998011 |0 (DE-627)369634780 |0 (DE-576)294611355 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t RöFo |d Stuttgart [u.a.] : Thieme, 1975 |g (2025), Seite [1-14] |h Online-Ressource |w (DE-627)324825021 |w (DE-600)2031079-1 |w (DE-576)107736748 |x 1438-9010 |7 nnas |a Performance of AI approaches for COVID-19 diagnosis using chest CT scans the impact of architecture and dataset = Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und Datensätzen |
| 773 | 1 | 8 | |g year:2025 |g pages:[1-14] |g extent:14 |a Performance of AI approaches for COVID-19 diagnosis using chest CT scans the impact of architecture and dataset = Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und Datensätzen |
| 856 | 4 | 0 | |u https://doi.org/10.1055/a-2577-3928 |x Verlag |x Resolving-System |z kostenfrei |3 Volltext |
| 856 | 4 | 0 | |u http://www.thieme-connect.de/DOI/DOI?10.1055/a-2577-3928 |x Verlag |z kostenfrei |3 Volltext |
| 951 | |a AR | ||
| 992 | |a 20251103 | ||
| 993 | |a Article | ||
| 994 | |a 2025 | ||
| 998 | |g 1227892640 |a Giannakis, Athanasios |m 1227892640:Giannakis, Athanasios |d 50000 |e 50000PG1227892640 |k 0/50000/ |p 6 | ||
| 999 | |a KXP-PPN1940056411 |e 4795724261 | ||
| BIB | |a Y | ||
| SER | |a journal | ||
| JSO | |a {"relHost":[{"title":[{"title_sort":"RöFo","title":"RöFo","subtitle":"Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren"}],"language":["ger"],"recId":"324825021","disp":"Performance of AI approaches for COVID-19 diagnosis using chest CT scans the impact of architecture and dataset = Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und DatensätzenRöFo","type":{"bibl":"periodical","media":"Online-Ressource"},"note":["Gesehen am 20.03.2023"],"part":{"extent":"14","text":"(2025), Seite [1-14]","pages":"[1-14]","year":"2025"},"titleAlt":[{"title":"RöFo vereinigt mit Aktuelle Radiologie"},{"title":"Fortschritte auf dem Gebiete der Röntgenstrahlen und der Nuklearmedizin"},{"title":"Fortschritte auf dem Gebiete der Röntgenstrahlen und der neuen bildgebenden Verfahren"}],"pubHistory":["122.1975 -"],"id":{"issn":["1438-9010"],"eki":["324825021"],"doi":["10.1055/s-00000066"],"zdb":["2031079-1"]},"origin":[{"publisherPlace":"Stuttgart [u.a.]","publisher":"Thieme","dateIssuedKey":"1975","dateIssuedDisp":"1975-"}],"physDesc":[{"extent":"Online-Ressource"}]}],"physDesc":[{"noteIll":"Illustrationen","extent":"14 S."}],"id":{"eki":["1940056411"],"doi":["10.1055/a-2577-3928"]},"origin":[{"dateIssuedDisp":"29. April 2025","dateIssuedKey":"2025"}],"name":{"displayForm":["Astha Jaiswal, Philipp Fervers, Fanyang Meng, Huimao Zhang, Dorottya Móré, Athanasios Giannakis, Jasmin Wailzer, Andreas Michael Bucher, David Maintz, Jonathan Kottlors, Rahil Shahzad, Thorsten Persigehl"]},"titleTranslated":[{"translated":"Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und Datensätzen"}],"language":["eng"],"recId":"1940056411","note":["Gesehen am 03.11.2025"],"type":{"bibl":"article-journal","media":"Online-Ressource"},"title":[{"title":"Performance of AI approaches for COVID-19 diagnosis using chest CT scans","subtitle":"the impact of architecture and dataset = Leistungsfähigkeit von KI-Methoden zur COVID-19-Diagnose mittels Thorax-CT: Der Einfluss von KI-Architektur und Datensätzen","title_sort":"Performance of AI approaches for COVID-19 diagnosis using chest CT scans"}],"person":[{"role":"aut","roleDisplay":"VerfasserIn","display":"Jaiswal, Astha","given":"Astha","family":"Jaiswal"},{"role":"aut","roleDisplay":"VerfasserIn","display":"Fervers, Philipp","given":"Philipp","family":"Fervers"},{"family":"Meng","given":"Fanyang","roleDisplay":"VerfasserIn","display":"Meng, Fanyang","role":"aut"},{"role":"aut","display":"Zhang, Huimao","roleDisplay":"VerfasserIn","given":"Huimao","family":"Zhang"},{"display":"Móré, Dorottya","roleDisplay":"VerfasserIn","role":"aut","family":"Móré","given":"Dorottya"},{"roleDisplay":"VerfasserIn","display":"Giannakis, Athanasios","role":"aut","family":"Giannakis","given":"Athanasios"},{"family":"Wailzer","given":"Jasmin","roleDisplay":"VerfasserIn","display":"Wailzer, Jasmin","role":"aut"},{"given":"Andreas Michael","family":"Bucher","role":"aut","roleDisplay":"VerfasserIn","display":"Bucher, Andreas Michael"},{"display":"Maintz, David","roleDisplay":"VerfasserIn","role":"aut","family":"Maintz","given":"David"},{"role":"aut","display":"Kottlors, Jonathan","roleDisplay":"VerfasserIn","given":"Jonathan","family":"Kottlors"},{"role":"aut","roleDisplay":"VerfasserIn","display":"Shahzad, Rahil","given":"Rahil","family":"Shahzad"},{"family":"Persigehl","given":"Thorsten","roleDisplay":"VerfasserIn","display":"Persigehl, Thorsten","role":"aut"}]} | ||
| SRT | |a JAISWALASTPERFORMANC2920 | ||