Computational Ophthalmology – Page 8 – big data. machine learning. data science.

April 10, 2019July 9, 2020

To Err is Both Human and AI

In a recent Outlook commentary in the journal Nature, Dr. Aaron Lee discusses the use of artificial intelligence (AI) in clinical medicine - both the significant potential for this technology as well as the various limitations and ethical issues involved with its implementation. He specifically addresses the case of AI screening systems for diabetic retinopathy - the first of these has already been approved for use by the FDA - but the issues with these systems apply to other areas of clinical medicine beyond ophthalmology. Dr. Lee makes the case that as this technology improves, and especially with the potential for deep learning algorithms to "learn" from their mistakes, the benefits from using AI as a tool for physicians may outweigh the risks.

April 5, 2019July 9, 2020

Ophthalmology-Based Neuropathology Risk Factors: Diabetic Retinopathy is Associated with Deep Microinfarcts in a Community-Based Autopsy Study.

Pubmed

Journal

Citation

The eye offers opportunities to capture critical information regarding the health of the brain, as the neurological anatomy and the microvasculature of the retina are very closely related to those of the brain and can be similarly affected by certain pathological processes. In this study published in the Journal of Alzheimer's Disease, Dr. Cecilia Lee and her co-authors compared brain neuropathology findings with known clinical information to determine if there were any associations between certain eye diseases and the presence of characteristic neuropathologic lesions of Alzheimer's disease. The authors analyzed the data from 676 patients from Adult Changes in Thought, a prospective longitudinal study in older patients who are followed over time and monitored for development of cognitive decline and dementia.

April 5, 2019April 15, 2020

Generating retinal flow maps from structural optical coherence tomography with artificial intelligence

Pubmed

Journal

Citation

Advances in both imaging technology and artificial intelligence are enabling many new developments in the field of ophthalmology. Optical coherence tomography angiography (OCTA) is one such development. OCTA measures blood flow in retinal microvasculature and can visualize the superficial and deep capillary plexus of the retinal vasculature without the need for dye. But the technology is also expensive and has a limited field of view. In this study, we sought to develop a deep learning model that could address this issue by first learning to infer between structure and retinal vascular function from structural OCT images and then generating an OCTA-like image. This would allow for accurate objectively-generated annotations, eliminating the need for expert-defined labels, and potentially enable the acquisition of new information about retinal blood circulation from preexisting OCT databases.

February 1, 2019June 15, 2020

Estimating Retinal Sensitivity Using Optical Coherence Tomography With Deep-Learning Algorithms in Macular Telangiectasia Type 2

Pubmed

Journal

Citation

Macular telangiectasia type 2 (MacTel) is a neurodegenerative disease affecting the parafoveal region of the retina, causing loss of central vision. Microperimetry testing is used to monitor progression of vision loss in patients with MacTel, during which multiple points on the retina are tested for sensitivity to light. Recent advances in imaging technology have correlated the focal light sensitivity defects that occur with MacTel to at least one structural defect, ellipsoid zone loss. We hypothesized that a deep learning model could be trained to estimate retinal sensitivity using structural details identified on optical coherence tomography (OCT) images and microperimetry data, providing a detailed high-resolution functional map of the retina that may be useful for research purposes or when microperimetry is not feasible.