In recent years, a shift in perceptions has occurred regarding the role of physicians, delivery systems, and payers from simply providing treatment for acute and chronic disease to the broader and more valued goal of enabling better health and quality of life for the public. With this shift, a need has emerged to better understand how to define and measure success more precisely in the context of shifting care paradigms enabled by continuing innovation. Innovations in ocular imaging, including ever-improving laser-based imaging methods such as OCT and wide-field scanning laser ophthalmoscopy, have changed our understanding and management of ophthalmic diseases and have transformed the delivery of eye care over the past decade. Not only have new imaging methods furthered our understanding of the pathophysiologic features of the eye, they also have helped to extend care into rural areas and places where shortages of eye care providers exist through, for instance, store-and-forward teleophthalmology. The rapid evolution and usefulness of ophthalmic imaging and algorithms designed to interpret it have created exciting opportunities, but also have created new challenges. At the same time, artificial intelligence (AI) applied to ophthalmic imaging is creating new eye care paradigms, as the United States Food and Drug Administration (FDA) approvals for autonomous AI in medicine for the detection of diabetic retinopathy have enabled, for the first time, the diagnosis of disease without the input of a clinician. As these advances in both hardware and software continue to spur evolution in how we deliver care to our patients, it is important to provide a forum for all stakeholders involved to have a voice in the crafting of these new paradigms.
These challenges impact many key stakeholders—including, patients, health care providers and systems, leaders in clinical and basic research and public health, governmental regulators, payers, and the pharmaceutical and medical device industry—and are best addressed by having those stakeholders work together to craft solutions. The Collaborative Community on Ophthalmic Imaging (CCOI) was formed in 2019 to clarify challenges, best practices, strategies, and standards while advancing innovation in the ophthalmic imaging space. The stakeholders involved were brought together to develop solutions and to help refine the diagnosis and management of patients with eye diseases, along with other medical conditions. In this editorial, we describe the origins of collaborative communities and specifically the CCOI: its mission and activities to date, summary of findings from its most recent public convening, and future directions.
What Is a Collaborative Community?
Collaborative communities are continuing forums where public and private sector members proactively work together to achieve common objectives and outcomes, to solve shared challenges, and to leverage collective opportunities in an environment of trust, respect, empathy, and openness. This novel collaborative group structure is facilitated by the internet1 and combines the knowledge of a global group of diverse experts primarily motivated by a collective mission, rather than solitary, autonomous efforts. The supportive but flexible organizational structure of collaborative communities is designed to foster not only innovation and agility, but also efficiency and scalability.2
In the case of modern collaborative communities formed in the not-for-profit world including mission-driven service organizations, governmental agencies, and non-governmental organizations, the preferred method is to bring together all relevant groups of stakeholders into a self-governing voluntary organization both to identify problems of shared interest and then to define best principles and practices to solve those problems. This type of collaborative community has, by design, a fairly flat structure relative to conventional hierarchical mission-driven industrial and governmental organizations because of its shared membership and governance. In addition to the qualitative goals of problem identification and prioritization, a need exists for quantitative precision in developing metrics that generally can be agreed on by the community as the arbiter of the finished work products, and then used to measure outcomes that help to define both success and failure. The potential advantage of this approach therefore lies in both its qualitative and quantitative features as well as its general applicability to all stakeholders and participants. The FDA’s Center for Devices and Radiological Health (CDRH) identified participating in collaborative communities as one of its strategic priorities.3 The FDA’s participation in collaborative communities with diverse stakeholders is key to ensuring that patients and providers have timely and continued access to safe, effective, and high-quality medical products. Although collaborative communities do not replace established regulatory mechanisms, they can create solutions that potentially can inform and streamline regulatory efforts. To date, the FDA participates in more than 10 collaborative communities,4 including the CCOI.
Genesis of the Collaborative Community on Ophthalmic Imaging
As part of a 2016 Research Collaboration Agreement between Stanford University’s Byers Eye Institute and the CDRH, several public cosponsored workshops were organized in collaboration with several professional organizations. These included the Ophthalmic Digital Health Workshop in 20175 and the Forum on Laser-Based Imaging in 2019.6 A broad variety of important topics addressed at these workshops included the ultra-high anatomic resolution and functional imaging capabilities of new forms of OCT and adaptive optics, remote imaging and functional assessment of ocular health with smart devices, the regulatory impacts of software as a medical device (SaMD), telehealth, nonclinical data sources, AI-enabled automated algorithmic image interpretation, and reimbursement. Software as a medical device is considered by the International Medical Device Regulators Forum (IMDRF) to be “software intended to be used for one or more medical purposes that perform these purposes without being part of a hardware medical device.”7 Artificial intelligence and machine learning technologies are considered to be a subset of SaMD. Discussions on these topics stimulated members of the meeting planning groups to consider ways to continue these efforts.
In response to the announcement by the FDA of collaborative communities as a strategic priority, leaders of the Byers Eye Institute of Stanford University’s Department of Ophthalmology and its Program in Ophthalmic Innovation spoke with a variety of experts and constituencies in the field for input. Organizations with global reach and representation that would be most knowledgeable and helpful in developing a strategic plan for a collaborative community in ophthalmic imaging were contacted. The leaders of those organizations selected members of their organizations to serve on the organizing committee. Selected members of the organizing committee agreed to serve on the executive committee on a voluntary, rotating basis and were tasked with helping to define the mission, vision, and substructure of the CCOI. This included identifying and recruiting domain experts to serve as members of specialized work-stream groups, as well as collaboratively identifying possible objectives and goals for each work effort.
The Vision and Mission of the Collaborative Community on Ophthalmic Imaging
The CCOI was established by its founding stakeholders out of the recognition that innovations in ophthalmic imaging have been advancing rapidly and that research and development of image-interpreting algorithms have the potential to drive major changes in eye care paradigms, including the capacity to diagnose disease and monitor progression. A charter was written reflecting these realities with the directive for public and private sector members to work together proactively to identify and clarify important challenges; to elucidate best practices, strategies, and standards; and to advance innovation within the general area of ophthalmic imaging. The path to doing so necessarily leads to a need for consensus among key stakeholders with experience and expertise in critical domains related to this field. The charter also outlined the scope of its work to “include but not be limited to related areas such as clinical endpoints, regulatory science, reimbursement, and artificial intelligence,” and the members committed to a “culture of collaboration, transparency, and trust fostered by shared governance by the stakeholders.” The members agreed to make their findings known through a variety of possible mechanisms including “where possible, a white paper summarizing the community’s views on a particular topic, workshops and forums in public venues, online publications and/or peer-reviewed journal articles or editorials, and through submission of findings to the Food & Drug Administration” where appropriate.8 The CCOI, with members from across the globe, extended an invitation for CDRH to participate as a formal member. Given the public health impact, this community was one of the first in which CDRH decided to participate.9
In addition to the FDA, the following organizations were invited and accepted membership in CCOI at the time of its formation: the American Academy of Ophthalmology, the American Society of Retinal Specialists, the American Academy of Cataract and Refractive Surgeons, the American Glaucoma Society, the European Society of Cataract and Refractive Surgery, the European Society of Retina Specialists (EURETINA), the Pan-American Academy of Ophthalmology, the Asia Pacific Academy of Ophthalmology, the Melanoma Research Foundation, the Glaucoma Research Foundation, the Lighthouse Guild, and the Byers Eye Institute at Stanford University.
Collaborative Community on Ophthalmic Imaging Organizational Structure
The executive committee of CCOI, with the Byers Eye Institute at Stanford serving as the convening center, agreed to plan and execute the first public meeting of the CCOI as well as to oversee the collaborative work products of the specialized workgroups. Currently, CCOI consists of six different work groups. The Foundational Principles of Ophthalmic Imaging and Algorithmic Interpretation group is horizontal in its purview and was organized to assist the disease-specific workgroups with expertise in the broader foundational principles of AI-enabled algorithmic image interpretation, including not only the technical aspects of coding, machine learning, and data security, but also related regulatory, reimbursement, and ethical considerations.10 Recently, a new, cross-cutting working group was formed centered on personally identifiable information entitled “Ocular Imaging—Personal Health Information (PHI) and Personally Identifiable Information (PII)” and is intended to be a major focus of the CCOI’s activities in upcoming public forums.
Currently 4 vertical work groups are focused on specific clinical indications: age-related macular degeneration (AMD), retinopathy of prematurity (ROP), glaucoma, and ocular oncology. The 4 diseases were chosen because they were deemed by the organizing committee to be (1) epidemiologically important areas of clinical need within the field of ophthalmology, (2) the indications where the most extensive and ongoing efforts are in progress to use AI to identify biomarkers of disease, and (3) as such, are arguably the ones where AI is closest to being used clinically in the near future. Leaders of each of these work groups are well-recognized authorities in their respective fields.
Diabetic retinopathy is also an important area of interest to the CCOI, but has not been designated at this time as a focus of a working group. The reasons for this are that (1) much attention in the research community already has been given to the topic and (2) to date, 2 technologies already have been cleared by the FDA for autonomous AI interpretation of images. The CCOI has chosen to focus on diseases where regulatory approval and clinical implementation of AI has not yet been achieved. It has identified AMD, ROP, ocular oncology, and glaucoma as the 4 emergent areas where a critical mass of research, knowledge, and expert opinion has been accumulated that are paving the way toward making algorithmic interpretation of images in these areas a clinical reality. The intent is to add additional disease-focused workgroups, including one focused on diabetic retinopathy among others, in the future.
Each workgroup leader has been charged with identifying other experts with global reach (including patients living with the disease) that could bring the necessary expertise and the patient perspective to identify unmet needs and potential solutions enabled by the use of advanced image acquisition and interpretation techniques to improve clinical outcomes (topics and the leaders of the individual workgroups, as well as members of the executive committee, are listed in Supplemental Materials, available at www.aaojournal.org). The organizational structure of the CCOI is relatively flat by design and comprises member organizations, to whom the CCOI organizing committee belong, and the CCOI executive committee, which includes both a board of directors and formal advisors to the board of directors. Under the purview of the executive committee are 2 horizontal workgroups–the Foundational Principles of Ophthalmic Imaging and Algorithmic Interpretation (FPOAI) group and the Ocular Imaging - PHI and PII? group–as well as 4 vertical work groups focused on AMD, ocular oncology, ROP, and glaucoma. An organizational chart summarizing the governance structure of the CCOI is shown in Figure S1 (available at www.aaojournal.org), with additional information about the participating organizations and the organizing committee members shown in Tables S1 and S2 (available at www.aaojournal.org). The full listing of participants in each workgroup are provided in Tables S3, S4, S5, S6, S7, and S8 (available at www.aaojournal.org).
Collaborative Community on Ophthalmic Imaging Work Products
Leveraging the work carried out by the IMDRF, each workgroup of the CCOI has made progress on defining the key challenges for ophthalmic image-based SaMD. The IMDRF agreed on the framework for risk categorization, the quality management system, and the clinical evaluation of SaMD. It describes clinical evaluation of SaMD as a process that consists of establishing valid clinical association, demonstrating adequate analytical validation, and performing clinical validation.11 Valid clinical association refers to the extent to which the SaMD’s output is accepted clinically and corresponds accurately in the real world to the SaMD’s targeted clinical condition.Analytical validation measures the ability of an SaMD to generate the intended technical output from the input data accurately, reliably, and precisely. Clinical validation, the final process of clinical evaluation, must show that SaMD has been tested in the target population for its intended use and that users can achieve clinically meaningful outcomes. The CCOI’s interest in both SaMD and AI center on these as tools to generate, enhance, and interpret ophthalmic images. General agreement exists within the CCOI that ophthalmic imaging, SaMD, and AI together can be key paradigm-shifting drivers in eye care delivery. The 4 CCOI vertical, disease-specific work groups have been discussing whether sufficient existing evidence is available currently or if new evidence needs to be generated for establishment of clinical associations in their respective disease areas.
Public conferences and resulting white papers and publications are the CCOI’s forums for identifying, presenting, and discussing unmet needs in the field of ophthalmic imaging. Through these venues, CCOI workgroups have put forth key considerations that need to be addressed to develop consensus on clinical meaningfulness for different functions of SaMD, such as diagnosing and predicting risk.
On September 3 and 4, 2020, CCOI held its first workshop entitled “The Future of Artificial Intelligence-Enabled Ophthalmic Image Interpretation: Accelerating Innovation and Implementation Pathways.”12 The workshop was hosted virtually by the convener, Stanford University, and garnered over 1100 registrants from more than 42 countries and 6 continents. The workshop covered key questions across a broad array of topics such as the role of AI in the diagnosis and management of intraocular tumors, a working diagnosis of glaucomatous optic neuropathy, machine parameters to identify possible or likely glaucoma, AI for population-based glaucoma screening, AI and machine learning in the prediction of AMD progression from fundus photographs, the use of AI and OCT in identifying AMD biomarkers, AI-driven approaches to classification and management of ROP in low-birthweight infants, telehealth for detection and management of ROP, and issues related to data acquisition, sharing, and ownership. The panel of speakers and discussants was a reflection of the stakeholder members of the CCOI, including members from a wide range of patient organizations. In the spirit of inclusion, the questions discussed during the meeting were posted for public comment so everyone had the opportunity to inform and influence the collective solutions being generated. To ensure the work being carried out by the CCOI is transparent, the workgroups have generated articles, in various stages of review and publication, reflecting the collaboratively defined challenges, some of which are being published alongside or following this editorial. As an example of the CCOI’s goal of identifying and discussing unmet needs in the field, the first 2 presentations of session 6 on day 2 of the September 2020 CCOI conference were entitled “Considerations in Detecting Glaucoma” and “The Challenge of the Definition and Grading of Glaucoma Using Advanced Imaging Techniques Including OCT and AI.” Drs. Garway-Heath and Schuman delineated the importance of developing consensus standards for glaucomatous optic neuropathy for diagnostic algorithms to become a clinical reality, with subsequent presentations given on potential solutions. Similar discourse on the unmet needs and potential solutions within AMD, ROP, and ocular oncology as they pertain to the clinical implementation of AI took place on day 1 of the conference.
Patient Perspectives: a Priority of the Collaborative Community on Ophthalmic Imaging
Patient representation is a priority of the CCOI. A unique aspect of the CCOI September 2020 convening was its integration and prioritization of patient and global health perspectives within each of the workgroup sessions. The CCOI conference included speakers and special panel discussions on both days of the conference from patient advocacy groups, and their interests and opinions continue to be heard across all working groups. Dr. Michelle Tarver, the director of patient science and engagement at the FDA; representatives from nonprofit organizations with disease-specific priorities such as the Cure Ocular Melanoma, Foundation Fighting Blindness, the Glaucoma Foundation, and the Glaucoma Research Foundation; as well as those with broader reach such as the Lighthouse Guild and Orbis International participated in live panels during the 2-day meeting.
An agile and responsive health care ecosystem is one that incorporates the perspectives of the impacted stakeholders and generates solutions that meet their collective needs. Typically, health care providers, payers, researchers, industry, and regulators are involved in this process, exchanging information and making decisions that impact patients and the public. However, one critical stakeholder often omitted from the solution creation process is the patient. Patients bring a rich perspective about what it is like to live daily with a medical condition and interact with the health care system. They are also increasingly activated to participate in not only their health care decisions, but also in the development and evaluation of technologies used in their care. Collaborative approaches that include patient input can have more impact when the patients are empowered to contribute. This might include providing educational materials in advance to help prepare them for the meetings, giving them an equal vote in the decisions, and including them in the dissemination process for those decisions. The CCOI has welcomed patient groups to participate on the steering committee as well as on relevant workstreams. During the public meeting held in September 2020, multiple panel sessions were held featuring the patient perspective. Those sessions revealed possible challenges as well as opportunities to advance the use of ophthalmic imaging for certain conditions. The inclusive approach the CCOI has taken is a call to action for other collaborative efforts to seek and include patient input. Representatives from the Melanoma Research Foundation, The Glaucoma Research Foundation, the Lighthouse Guild, Foundation Fighting Blindness, The Glaucoma Foundation, and Orbis International provided insights and engaged in live panel discussions on topics of advocacy and patient perceptions on the emergence of AI and home monitoring on care paradigms. A full list of panel questions discussed at the meeting can be found at: https://www.cc-oi.org/panel-questions.
The CCOI’s commitment to the perspectives and needs of patients is especially salient in light of the recent adoption by the United Nations General Assembly of a first-ever resolution on vision, which commits all 193 of its member states to ensure eye care for all by 2030. The resolution aims to address the clinical needs of 1.1 billion people around the world who experience visual impairment but have limited access to eye care.13 The CCOI’s mission aligns well with this resolution, with a specific focus on advancing innovations and the clinical implementation of ophthalmic imaging technologies as key drivers for increasing access to care worldwide.
Subsequent to the workshop, all of the recordings of the formal presentations and live panel discussions were made available publicly on the CCOI website (http://www.cc-oi.org). Each work group has reviewed the proceedings of the workshop and analyzed the comments received via the website during the conference and through the web portal. The articles included herein within this special issue as well as in additional journal publications represent a significant component of the September CCOI meeting and the follow-up considerations of each workgroup, and have provided the roadmap for ongoing discussions.
The Collaborative Community on Ophthalmic Imaging has been formed to accelerate innovation and the introduction of new ophthalmic imaging technologies into clinical practice to improve patient outcomes. These improved outcomes can include earlier and more accurate disease detection and classification as well as more effective treatment strategies. The CCOI can be instrumental in generating recommendations for appropriate clinical evaluation of SaMD based on ophthalmic imaging, expediting the establishment of adequate clinical evidence to help inform global regulators, clinicians, patients, and payers.
Supplementary Material
Disclosure(s):
All authors have completed and submitted the ICMJE disclosures form.
The author(s) have made the following disclosure(s): Emily Chew, an associate editor of this journal and Michael Repka, an editorial board member of this journal, were recused from the peer-review process of this article and had no access to information regarding its peer-review.
M.S.B.: – Verana Health; Patent - Smartphone-enabled vision testing. M.S.B. and D.M.: Support from the Byers Eye Institute and Ophthalmic Innovation Program at Stanford, as well as departmental core grants from the National Eye Institute (P30-EY026877) and Research to Prevent Blindness, Inc.
D.M.: Patent – Lens adapters for mobile anterior and posterior segment ophthalmoscopy.
The Food and Drug Administration participates as a member of the Collaborative Community on Ophthalmic Imaging. This article reflects the views of the authors and should not be construed to represent the Food and Drug Administration’s views or policies.
Collaborative Community on Ophthalmic Imaging Executive Committee Members
Mark S. Blumenkranz, MD, MMS – HJ Smead Professor Emeritus, Co-Director of the Ophthalmic Innovation Program, Byers Eye Institute at Stanford
Malvina B. Eydelman, MD – Director, Office of Health Technology 1, Ophthalmic, Anesthesia, Respiratory, ENT, & Dental Devices, Center for Devices and Radiological Health, Food and Drug Administration
Michael D. Abràmoff, MD, PhD – Robert C. Watzke Professor of Ophthalmology and Visual Sciences, University of Iowa. Founder and Executive Chairman, Digital Diagnostics Inc, Coralville, Iowa.
Emily Chew, MD – Director, Division of Epidemiology and Clinical Applications, National Eye Institute
Michael F. Chiang, MD – Director, National Eye Institute
Aaron Lee, MD – Associate Professor of Ophthalmology, University of Washington
David Myung, MD, PhD – Assistant Professor of Ophthalmology and, by courtesy, Chemical Engineering, Director - Ophthalmic Innovation Program, Byers Eye Institute at Stanford and VA Palo Alto Health Care System
Michael Repka, MD, MBA – Professor of Ophthalmology, Wilmer Eye Institute, Johns Hopkins University
Joel S. Schuman, MD – Elaine Langone Professor of Ophthalmology, Professor of Neuroscience and Physiology, Neural Science, Biomedical Engineering and Electrical and Computer Engineering, NYU Langone and NYU Tandon
Carol Shields, MD – Chief, Ocular Oncology, Wills Eye Hospital; Professor of Ophthalmology, Thomas Jefferson University
Michelle E. Tarver, MD, PhD – Director, Patient Science and Engagement Program, Office of Strategic Partnerships and Technology Innovation, CDRH, Food and Drug Administration
Footnotes
Supplemental material available at www.aaojournal.org.
Contributor Information
Mark S. Blumenkranz, Byers Eye Institute at Stanford.
Michelle E. Tarver, Patient Science and Engagement Program, Office of Strategic Partnerships and Technology Innovation, CDRH, Food and Drug Administration.
David Myung, Ophthalmic Innovation Program, Byers Eye Institute at Stanford and VA Palo Alto Health Care System.
Malvina B. Eydelman, Office of Health Technology 1, Ophthalmic, Anesthesia, Respiratory, ENT, & Dental Devices, Center for Devices and Radiological Health, Food and Drug Administration.
References
- 1.Bøllingtoft A, Donaldson L, Huber GP, et al. Collaborative Communities of Firms: Purpose, Process, and Design. Springer Science & Business Media, New York, NY; 2011. [Google Scholar]
- 2.Prusak L. Building a collaborative enterprise. Harv Bus Rev. 2011;89(7–8):94–101. [PubMed] [Google Scholar]
- 3.Shuren J. Charting our course for 2018–2020. January 17, 2018. Available at: https://www.fda.gov/news-events/fda-voices/charting-our-course-2018-2020. Accessed 30.09.21. [Google Scholar]
- 4.Food and Drug Administration. Collaborative communities: addressing health care challenges together. Available at: https://www.fda.gov/about-fda/cdrh-strategic-priorities-and-updates/collaborative-communities-addressing-health-care-challenges-together; 2020. Accessed 30.09.21. [Google Scholar]
- 5.Bodnar ZM, Schuchard R, Myung D, et al. Evaluating new ophthalmic digital devices for safety and effectiveness in the context of rapid technological development. JAMA Ophthalmol. 2019;137(8):939–944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Brodie F, Repka M, Burns SA, et al. Development, validation, and innovation in ophthalmic laser-based imaging: report from a US Food and Drug Administration-cosponsored forum. JAMA Ophthalmol. 2021;139(1):113–118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.International Medical Device Regulators Forum. Software as a medical device (SaMD): key definitions. Available at: https://www.imdrf.org/docs/imdrf/final/technical/imdrf-tech-131209-samd-key-definitions-140901.pdf; 2013. Accessed 30.09.21.
- 8.Collaborative Community on Ophthalmic Imaging. Homepage. Available at: https://www.cc-oi.org; 2020. Accessed 30.09.21.
- 9.Rosenfeld S. FDA announces participation in 2 collaborative communities to overcome med tech challenges. Available at: https://www.chiefhealthcareexecutive.com/view/fda-announces-participation-in-collaborative-communities; 2019. Accessed 30.09.21. [Google Scholar]
- 10.Abràmoff MD, Tobey D, Char DS. Lessons learned about autonomous AI: finding a safe, efficacious, and ethical path through the development process. Am J Ophthalmol. 2020;214:134–142. [DOI] [PubMed] [Google Scholar]
- 11.International Medical Device Regulators Forum. Software as a medical device (SaMD): clinical evaluation. Available at: http://www.imdrf.org/docs/imdrf/final/consultations/imdrf-cons-samd-ce.pdf; 2016. Accessed 30.09.21.
- 12.Collaborative Community on Ophthalmic Imaging. 2020 Conference program. Available at: https://www.cc-oi.org/conference-program; 2020. Accessed 30.09.21. [Google Scholar]
- 13.Lederer EM. UN adopts first resolution on vision, aims to help 1 billion. July 24, 2021. Available at: https://apnews.com/article/health-united-nations-united-nations-general-assembly-d2cd6376a10e89b9352fe4267c12da82. Accessed 30.09.21. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.