Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5)
Abstract Real-world retinal optical coherence tomography (OCT) scans are available in abundance in primary and secondary eye care centres. They contain a wealth of information to be analyzed in retrospective studies. The associated electronic health records alone are often not enough to generate a h...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-11-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-46626-7 |
| _version_ | 1797630294388375552 |
|---|---|
| author | Oliver Leingang Sophie Riedl Julia Mai Gregor S. Reiter Georg Faustmann Philipp Fuchs Hendrik P. N. Scholl Sobha Sivaprasad Daniel Rueckert Andrew Lotery Ursula Schmidt-Erfurth Hrvoje Bogunović |
| author_facet | Oliver Leingang Sophie Riedl Julia Mai Gregor S. Reiter Georg Faustmann Philipp Fuchs Hendrik P. N. Scholl Sobha Sivaprasad Daniel Rueckert Andrew Lotery Ursula Schmidt-Erfurth Hrvoje Bogunović |
| author_sort | Oliver Leingang |
| collection | DOAJ |
| description | Abstract Real-world retinal optical coherence tomography (OCT) scans are available in abundance in primary and secondary eye care centres. They contain a wealth of information to be analyzed in retrospective studies. The associated electronic health records alone are often not enough to generate a high-quality dataset for clinical, statistical, and machine learning analysis. We have developed a deep learning-based age-related macular degeneration (AMD) stage classifier, to efficiently identify the first onset of early/intermediate (iAMD), atrophic (GA), and neovascular (nAMD) stage of AMD in retrospective data. We trained a two-stage convolutional neural network to classify macula-centered 3D volumes from Topcon OCT images into 4 classes: Normal, iAMD, GA and nAMD. In the first stage, a 2D ResNet50 is trained to identify the disease categories on the individual OCT B-scans while in the second stage, four smaller models (ResNets) use the concatenated B-scan-wise output from the first stage to classify the entire OCT volume. Classification uncertainty estimates are generated with Monte-Carlo dropout at inference time. The model was trained on a real-world OCT dataset, 3765 scans of 1849 eyes, and extensively evaluated, where it reached an average ROC-AUC of 0.94 in a real-world test set. |
| first_indexed | 2024-03-11T11:06:18Z |
| format | Article |
| id | doaj.art-3332971330e442398e301671d3f110df |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-03-11T11:06:18Z |
| publishDate | 2023-11-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-3332971330e442398e301671d3f110df2023-11-12T12:12:55ZengNature PortfolioScientific Reports2045-23222023-11-0113111310.1038/s41598-023-46626-7Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5)Oliver Leingang0Sophie Riedl1Julia Mai2Gregor S. Reiter3Georg Faustmann4Philipp Fuchs5Hendrik P. N. Scholl6Sobha Sivaprasad7Daniel Rueckert8Andrew Lotery9Ursula Schmidt-Erfurth10Hrvoje Bogunović11Department of Ophthalmology and Optometry, Medical University of ViennaDepartment of Ophthalmology and Optometry, Medical University of ViennaDepartment of Ophthalmology and Optometry, Medical University of ViennaDepartment of Ophthalmology and Optometry, Medical University of ViennaDepartment of Ophthalmology and Optometry, Medical University of ViennaDepartment of Ophthalmology and Optometry, Medical University of ViennaInstitute of Molecular and Clinical Ophthalmology BaselNIHR Moorfields Biomedical Research Centre, Moorfields Eye Hospital NHS Foundation TrustBioMedIA, Imperial College LondonClinical and Experimental Sciences, Faculty of Medicine, University of SouthamptonDepartment of Ophthalmology and Optometry, Medical University of ViennaDepartment of Ophthalmology and Optometry, Medical University of ViennaAbstract Real-world retinal optical coherence tomography (OCT) scans are available in abundance in primary and secondary eye care centres. They contain a wealth of information to be analyzed in retrospective studies. The associated electronic health records alone are often not enough to generate a high-quality dataset for clinical, statistical, and machine learning analysis. We have developed a deep learning-based age-related macular degeneration (AMD) stage classifier, to efficiently identify the first onset of early/intermediate (iAMD), atrophic (GA), and neovascular (nAMD) stage of AMD in retrospective data. We trained a two-stage convolutional neural network to classify macula-centered 3D volumes from Topcon OCT images into 4 classes: Normal, iAMD, GA and nAMD. In the first stage, a 2D ResNet50 is trained to identify the disease categories on the individual OCT B-scans while in the second stage, four smaller models (ResNets) use the concatenated B-scan-wise output from the first stage to classify the entire OCT volume. Classification uncertainty estimates are generated with Monte-Carlo dropout at inference time. The model was trained on a real-world OCT dataset, 3765 scans of 1849 eyes, and extensively evaluated, where it reached an average ROC-AUC of 0.94 in a real-world test set.https://doi.org/10.1038/s41598-023-46626-7 |
| spellingShingle | Oliver Leingang Sophie Riedl Julia Mai Gregor S. Reiter Georg Faustmann Philipp Fuchs Hendrik P. N. Scholl Sobha Sivaprasad Daniel Rueckert Andrew Lotery Ursula Schmidt-Erfurth Hrvoje Bogunović Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5) Scientific Reports |
| title | Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5) |
| title_full | Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5) |
| title_fullStr | Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5) |
| title_full_unstemmed | Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5) |
| title_short | Automated deep learning-based AMD detection and staging in real-world OCT datasets (PINNACLE study report 5) |
| title_sort | automated deep learning based amd detection and staging in real world oct datasets pinnacle study report 5 |
| url | https://doi.org/10.1038/s41598-023-46626-7 |
| work_keys_str_mv | AT oliverleingang automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT sophieriedl automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT juliamai automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT gregorsreiter automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT georgfaustmann automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT philippfuchs automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT hendrikpnscholl automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT sobhasivaprasad automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT danielrueckert automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT andrewlotery automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT ursulaschmidterfurth automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 AT hrvojebogunovic automateddeeplearningbasedamddetectionandstaginginrealworldoctdatasetspinnaclestudyreport5 |