Bart Liefers1, Paul Taylor2, Abdulrahman Alsaedi3, Clare Bailey4, Konstantinos Balaskas5, Narendra Dhingra6, Catherine A Egan7, Filipa Gomes Rodrigues7, Cristina González Gonzalo8, Tjebo F C Heeren9, Andrew Lotery10, Philipp L Müller11, Abraham Olvera-Barrios9, Bobby Paul12, Roy Schwartz13, Darren S Thomas2, Alasdair N Warwick14, Adnan Tufail9, Clara I Sánchez15. 1. A-eye Research Group, Diagnostic Image Analysis Group, Department of Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, The Netherlands; Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, The Netherlands. Electronic address: b.liefers@nhs.net. 2. Institute of Health Informatics, University College London (UCL), London, UK. 3. College of Medicine, Imam Mohammad Ibn Saud Islamic University, Riyadh, Saudi Arabia; Moorfields Eye Hospital NHS Foundation Trust, London. 4. University Hospitals Bristol NHS Foundation Trust, Bristol. 5. Moorfields Ophthalmic Reading Centre, Moorfields Eye Hospital NHS Foundation Trust, London; National Institute for Health Research Biomedical Research Centre at Moorfields Eye Hospital National Health Service (NHS) Foundation Trust, UCL Institute of Ophthalmology, London. 6. Mid Yorkshire Hospitals NHS Trust, Southampton. 7. Moorfields Eye Hospital NHS Foundation Trust, London; National Institute for Health Research Biomedical Research Centre at Moorfields Eye Hospital National Health Service (NHS) Foundation Trust, UCL Institute of Ophthalmology, London. 8. A-eye Research Group, Diagnostic Image Analysis Group, Department of Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, The Netherlands; Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, The Netherlands. 9. Moorfields Eye Hospital NHS Foundation Trust, London; Institute of Ophthalmology, University College London, London, UK. 10. Faculty of Medicine, University of Southampton, Southampton. 11. Moorfields Eye Hospital NHS Foundation Trust, London; Institute of Ophthalmology, University College London, London, UK; Department of Ophthalmology, University of Bonn, Bonn, Germany. 12. Barking, Havering and Redbridge University Hospitals NHS Trust, Romford, United Kingdom. 13. Moorfields Eye Hospital NHS Foundation Trust, London. 14. Moorfields Eye Hospital NHS Foundation Trust, London; Institute of Cardiovascular Science, University College London, London. 15. A-eye Research Group, Diagnostic Image Analysis Group, Department of Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, The Netherlands; Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, The Netherlands; Department of Ophthalmology, Radboud University Medical Center, Nijmegen, the Netherlands.
Abstract
PURPOSE: We sought to develop and validate a deep learning model for segmentation of 13 features associated with neovascular and atrophic age-related macular degeneration (AMD). DESIGN: Development and validation of a deep-learning model for feature segmentation. METHODS: Data for model development were obtained from 307 optical coherence tomography volumes. Eight experienced graders manually delineated all abnormalities in 2712 B-scans. A deep neural network was trained with these data to perform voxel-level segmentation of the 13 most common abnormalities (features). For evaluation, 112 B-scans from 112 patients with a diagnosis of neovascular AMD were annotated by 4 independent observers. The main outcome measures were Dice score, intraclass correlation coefficient, and free-response receiver operating characteristic curve. RESULTS: On 11 of 13 features, the model obtained a mean Dice score of 0.63 ± 0.15, compared with 0.61 ± 0.17 for the observers. The mean intraclass correlation coefficient for the model was 0.66 ± 0.22, compared with 0.62 ± 0.21 for the observers. Two features were not evaluated quantitatively because of a lack of data. Free-response receiver operating characteristic analysis demonstrated that the model scored similar or higher sensitivity per false positives compared with the observers. CONCLUSIONS: The quality of the automatic segmentation matches that of experienced graders for most features, exceeding human performance for some features. The quantified parameters provided by the model can be used in the current clinical routine and open possibilities for further research into treatment response outside clinical trials.
PURPOSE: We sought to develop and validate a deep learning model for segmentation of 13 features associated with neovascular and atrophic age-related macular degeneration (AMD). DESIGN: Development and validation of a deep-learning model for feature segmentation. METHODS: Data for model development were obtained from 307 optical coherence tomography volumes. Eight experienced graders manually delineated all abnormalities in 2712 B-scans. A deep neural network was trained with these data to perform voxel-level segmentation of the 13 most common abnormalities (features). For evaluation, 112 B-scans from 112 patients with a diagnosis of neovascular AMD were annotated by 4 independent observers. The main outcome measures were Dice score, intraclass correlation coefficient, and free-response receiver operating characteristic curve. RESULTS: On 11 of 13 features, the model obtained a mean Dice score of 0.63 ± 0.15, compared with 0.61 ± 0.17 for the observers. The mean intraclass correlation coefficient for the model was 0.66 ± 0.22, compared with 0.62 ± 0.21 for the observers. Two features were not evaluated quantitatively because of a lack of data. Free-response receiver operating characteristic analysis demonstrated that the model scored similar or higher sensitivity per false positives compared with the observers. CONCLUSIONS: The quality of the automatic segmentation matches that of experienced graders for most features, exceeding human performance for some features. The quantified parameters provided by the model can be used in the current clinical routine and open possibilities for further research into treatment response outside clinical trials.
Authors: Kyung Jun Choi; Jung Eun Choi; Hyeon Cheol Roh; Jun Soo Eun; Jong Min Kim; Yong Kyun Shin; Min Chae Kang; Joon Kyo Chung; Chaeyeon Lee; Dongyoung Lee; Se Woong Kang; Baek Hwan Cho; Sang Jin Kim Journal: Sci Rep Date: 2021-11-04 Impact factor: 4.379
Authors: LeAnne H Young; Jongwoo Kim; Mehmet Yakin; Henry Lin; David T Dao; Shilpa Kodati; Sumit Sharma; Aaron Y Lee; Cecilia S Lee; H Nida Sen Journal: Transl Vis Sci Technol Date: 2022-07-08 Impact factor: 3.048
Authors: Konstantinos Balaskas; S Glinton; T D L Keenan; L Faes; B Liefers; G Zhang; N Pontikos; R Struyven; S K Wagner; A McKeown; P J Patel; P A Keane; D J Fu Journal: Sci Rep Date: 2022-09-16 Impact factor: 4.996
Authors: Marlene Saßmannshausen; Sarah Thiele; Charlotte Behning; Maximilian Pfau; Matthias Schmid; Sérgio Leal; Ulrich F O Luhmann; Robert P Finger; Frank G Holz; Steffen Schmitz-Valckenberg Journal: Transl Vis Sci Technol Date: 2022-03-02 Impact factor: 3.283