Applications of Deep Learning
Objective: To develop a generative adversarial network (GAN) to segment major blood vessels from retinal flat-mount images from oxygen-induced retinopathy (OIR) and demonstrate the utility of these GAN-generated vessel segmentations in quantifying vascular tortuosity. Design: Development and validat...
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-01-01
|
| Series: | Ophthalmology Science |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666914523000702 |
| _version_ | 1827793522049155072 |
|---|---|
| author | Jimmy S. Chen, MD Kyle V. Marra, MD, PhD Hailey K. Robles-Holmes, BS Kristine B. Ly, BS Joseph Miller, MS Guoqin Wei, PhD Edith Aguilar, MD Felicitas Bucher, MD, PhD Yoichi Ideguchi, BS Aaron S. Coyner, PhD Napoleone Ferrara, MD J. Peter Campbell, MD, MPH Martin Friedlander, MD, PhD Eric Nudleman, MD, PhD |
| author_facet | Jimmy S. Chen, MD Kyle V. Marra, MD, PhD Hailey K. Robles-Holmes, BS Kristine B. Ly, BS Joseph Miller, MS Guoqin Wei, PhD Edith Aguilar, MD Felicitas Bucher, MD, PhD Yoichi Ideguchi, BS Aaron S. Coyner, PhD Napoleone Ferrara, MD J. Peter Campbell, MD, MPH Martin Friedlander, MD, PhD Eric Nudleman, MD, PhD |
| author_sort | Jimmy S. Chen, MD |
| collection | DOAJ |
| description | Objective: To develop a generative adversarial network (GAN) to segment major blood vessels from retinal flat-mount images from oxygen-induced retinopathy (OIR) and demonstrate the utility of these GAN-generated vessel segmentations in quantifying vascular tortuosity. Design: Development and validation of GAN. Subjects: Three datasets containing 1084, 50, and 20 flat-mount mice retina images with various stains used and ages at sacrifice acquired from previously published manuscripts. Methods: Four graders manually segmented major blood vessels from flat-mount images of retinas from OIR mice. Pix2Pix, a high-resolution GAN, was trained on 984 pairs of raw flat-mount images and manual vessel segmentations and then tested on 100 and 50 image pairs from a held-out and external test set, respectively. GAN-generated and manual vessel segmentations were then used as an input into a previously published algorithm (iROP-Assist) to generate a vascular cumulative tortuosity index (CTI) for 20 image pairs containing mouse eyes treated with aflibercept versus control. Main Outcome Measures: Mean dice coefficients were used to compare segmentation accuracy between the GAN-generated and manually annotated segmentation maps. For the image pairs treated with aflibercept versus control, mean CTIs were also calculated for both GAN-generated and manual vessel maps. Statistical significance was evaluated using Wilcoxon signed-rank tests (P ≤ 0.05 threshold for significance). Results: The dice coefficient for the GAN-generated versus manual vessel segmentations was 0.75 ± 0.27 and 0.77 ± 0.17 for the held-out test set and external test set, respectively. The mean CTI generated from the GAN-generated and manual vessel segmentations was 1.12 ± 0.07 versus 1.03 ± 0.02 (P = 0.003) and 1.06 ± 0.04 versus 1.01 ± 0.01 (P < 0.001), respectively, for eyes treated with aflibercept versus control, demonstrating that vascular tortuosity was rescued by aflibercept when quantified by GAN-generated and manual vessel segmentations. Conclusions: GANs can be used to accurately generate vessel map segmentations from flat-mount images. These vessel maps may be used to evaluate novel metrics of vascular tortuosity in OIR, such as CTI, and have the potential to accelerate research in treatments for ischemic retinopathies. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. |
| first_indexed | 2024-03-11T18:19:38Z |
| format | Article |
| id | doaj.art-f3adeaa2abf74ce192a0b29e10d3cb0c |
| institution | Directory Open Access Journal |
| issn | 2666-9145 |
| language | English |
| last_indexed | 2024-03-11T18:19:38Z |
| publishDate | 2024-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Ophthalmology Science |
| spelling | doaj.art-f3adeaa2abf74ce192a0b29e10d3cb0c2023-10-16T04:12:55ZengElsevierOphthalmology Science2666-91452024-01-0141100338Applications of Deep LearningJimmy S. Chen, MD0Kyle V. Marra, MD, PhD1Hailey K. Robles-Holmes, BS2Kristine B. Ly, BS3Joseph Miller, MS4Guoqin Wei, PhD5Edith Aguilar, MD6Felicitas Bucher, MD, PhD7Yoichi Ideguchi, BS8Aaron S. Coyner, PhD9Napoleone Ferrara, MD10J. Peter Campbell, MD, MPH11Martin Friedlander, MD, PhD12Eric Nudleman, MD, PhD13Shiley Eye Institute, Viterbi Family Department of Ophthalmology, University of California San Diego, San Diego, CaliforniaMolecular Medicine, the Scripps Research Institute, San Diego, California; School of Medicine, University of California San Diego, San Diego, CaliforniaShiley Eye Institute, Viterbi Family Department of Ophthalmology, University of California San Diego, San Diego, CaliforniaCollege of Optometry, Pacific University, Forest Grove, OregonShiley Eye Institute, Viterbi Family Department of Ophthalmology, University of California San Diego, San Diego, CaliforniaMolecular Medicine, the Scripps Research Institute, San Diego, CaliforniaMolecular Medicine, the Scripps Research Institute, San Diego, CaliforniaEye Center, Medical Center, Faculty of Medicine, University of Freiburg, Freiburg, GermanyMolecular Medicine, the Scripps Research Institute, San Diego, CaliforniaCasey Eye Institute, Department of Ophthalmology, Oregon Health & Science University, Portland, OregonShiley Eye Institute, Viterbi Family Department of Ophthalmology, University of California San Diego, San Diego, CaliforniaCasey Eye Institute, Department of Ophthalmology, Oregon Health & Science University, Portland, OregonMolecular Medicine, the Scripps Research Institute, San Diego, CaliforniaShiley Eye Institute, Viterbi Family Department of Ophthalmology, University of California San Diego, San Diego, California; Correspondence: Eric Nudleman, MD, PhD, 9415 Campus Point Dr, MC 0946, La Jolla, CA 92093-0946, USA.Objective: To develop a generative adversarial network (GAN) to segment major blood vessels from retinal flat-mount images from oxygen-induced retinopathy (OIR) and demonstrate the utility of these GAN-generated vessel segmentations in quantifying vascular tortuosity. Design: Development and validation of GAN. Subjects: Three datasets containing 1084, 50, and 20 flat-mount mice retina images with various stains used and ages at sacrifice acquired from previously published manuscripts. Methods: Four graders manually segmented major blood vessels from flat-mount images of retinas from OIR mice. Pix2Pix, a high-resolution GAN, was trained on 984 pairs of raw flat-mount images and manual vessel segmentations and then tested on 100 and 50 image pairs from a held-out and external test set, respectively. GAN-generated and manual vessel segmentations were then used as an input into a previously published algorithm (iROP-Assist) to generate a vascular cumulative tortuosity index (CTI) for 20 image pairs containing mouse eyes treated with aflibercept versus control. Main Outcome Measures: Mean dice coefficients were used to compare segmentation accuracy between the GAN-generated and manually annotated segmentation maps. For the image pairs treated with aflibercept versus control, mean CTIs were also calculated for both GAN-generated and manual vessel maps. Statistical significance was evaluated using Wilcoxon signed-rank tests (P ≤ 0.05 threshold for significance). Results: The dice coefficient for the GAN-generated versus manual vessel segmentations was 0.75 ± 0.27 and 0.77 ± 0.17 for the held-out test set and external test set, respectively. The mean CTI generated from the GAN-generated and manual vessel segmentations was 1.12 ± 0.07 versus 1.03 ± 0.02 (P = 0.003) and 1.06 ± 0.04 versus 1.01 ± 0.01 (P < 0.001), respectively, for eyes treated with aflibercept versus control, demonstrating that vascular tortuosity was rescued by aflibercept when quantified by GAN-generated and manual vessel segmentations. Conclusions: GANs can be used to accurately generate vessel map segmentations from flat-mount images. These vessel maps may be used to evaluate novel metrics of vascular tortuosity in OIR, such as CTI, and have the potential to accelerate research in treatments for ischemic retinopathies. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.http://www.sciencedirect.com/science/article/pii/S2666914523000702Artificial intelligenceData scienceOxygen-induced retinopathyVascular tortuosity |
| spellingShingle | Jimmy S. Chen, MD Kyle V. Marra, MD, PhD Hailey K. Robles-Holmes, BS Kristine B. Ly, BS Joseph Miller, MS Guoqin Wei, PhD Edith Aguilar, MD Felicitas Bucher, MD, PhD Yoichi Ideguchi, BS Aaron S. Coyner, PhD Napoleone Ferrara, MD J. Peter Campbell, MD, MPH Martin Friedlander, MD, PhD Eric Nudleman, MD, PhD Applications of Deep Learning Ophthalmology Science Artificial intelligence Data science Oxygen-induced retinopathy Vascular tortuosity |
| title | Applications of Deep Learning |
| title_full | Applications of Deep Learning |
| title_fullStr | Applications of Deep Learning |
| title_full_unstemmed | Applications of Deep Learning |
| title_short | Applications of Deep Learning |
| title_sort | applications of deep learning |
| topic | Artificial intelligence Data science Oxygen-induced retinopathy Vascular tortuosity |
| url | http://www.sciencedirect.com/science/article/pii/S2666914523000702 |
| work_keys_str_mv | AT jimmyschenmd applicationsofdeeplearning AT kylevmarramdphd applicationsofdeeplearning AT haileykroblesholmesbs applicationsofdeeplearning AT kristineblybs applicationsofdeeplearning AT josephmillerms applicationsofdeeplearning AT guoqinweiphd applicationsofdeeplearning AT edithaguilarmd applicationsofdeeplearning AT felicitasbuchermdphd applicationsofdeeplearning AT yoichiideguchibs applicationsofdeeplearning AT aaronscoynerphd applicationsofdeeplearning AT napoleoneferraramd applicationsofdeeplearning AT jpetercampbellmdmph applicationsofdeeplearning AT martinfriedlandermdphd applicationsofdeeplearning AT ericnudlemanmdphd applicationsofdeeplearning |