Forging Connections in Latin America to Advance AI in Radiology
The 1° Encontro Latino-Americano de IA em Saúde (1st Latin American Meeting on AI in Health) was held during the 2022 Jornada Paulista de Radiologia, the annual radiology meeting in the state of São Paulo. The event was created to foster discussion among Latin American countries about the complexity, challenges, and opportunities in developing and using artificial intelligence (AI) in those countries. Technological improvements in AI have created high expectations in health care. AI is recognized increasingly as a game changer in clinical radiology. To counter the fear that AI would “take over” radiology, the program included activities to educate radiologists. The development of AI in Latin America is in its early days, and although there are some pioneer cases, many regions still lack world-class technological infrastructure and resources. Legislation, regulation, and public policies in data privacy and protection, digital health, and AI are recent advances in many countries. The meeting program was developed with a broad scope, with expertise from different countries, backgrounds, and specialties, with the objective of encompassing all levels of complexity (from basic concepts to advanced techniques), perspectives (clinical, technical, ethical, and business), and specialties (both informatics and data science experts and the usual radiology clinical groups). It was an opportunity to connect with peers from other countries and share lessons learned about AI in health care in different countries and contexts.
Keywords: Informatics, Use of AI in Education, Impact of AI on Education, Social Implications
© RSNA, 2022
The 1° Encontro Latino-Americano de IA em Saúde (1st Latin American Meeting on AI in Health) offered an opportunity to connect with peers and share lessons learned about health care artificial intelligence in the Americas.
■ Technological improvements in artificial intelligence (AI) have raised expectations, enthusiasm, and concerns among radiologists.
■ To foster the scientific development of radiology, the São Paulo Radiology Society gathered radiology association representatives and world-class experts from the Americas to host the 1° Encontro Latino-Americano de IA em Saúde (1st Latin American Meeting on AI in Health).
■ The meeting program offered a broad scope and brought together expertise from varying countries, backgrounds, and specialties, from basic concepts to advanced techniques, and across clinical, technical, ethical, and managerial perspectives.
In recent years, substantial technological improvements in artificial intelligence (AI) have created high expectations among health care providers, policymakers, and radiologists (1). AI is recognized increasingly as a game changer in clinical radiology as machine learning (ML) techniques are being used to extract crucial information from radiologic images, radiology reports, and tabular data to improve workflow, formulate more precise diagnoses, increase efficiency, and reduce errors and costs (2,3).
In Brazil, the fear that AI would “take over” radiology has haunted radiologists for the past 6 years. Our better-informed colleagues are now more confident that AI will not replace us. However, many radiologists and physicians from other medical specialties recently have discovered AI and are spreading the opinion that AI will replace radiologists. To counter these concerns, the Sociedade Paulista de Radiologia e Diagnóstico por Imagem (SPR, São Paulo Society of Radiology and Diagnostic Imaging) envisioned activities to educate mainstream radiologists to be prepared for the future. We expect radiologists to know the basics of AI and its implications in clinical practice. That would also make our specialty more prepared to discuss the topic with physicians of other specialties.
The use of AI in Latin America faces unique challenges not found in the United States, Canada, Europe, Africa, Asia, or Australia. For instance, technological infrastructure, such as high-end computers and high-speed internet connections, is usually lacking in most of the region. As well, some diseases that are endemic in Latin America are unusual in other places. Legislation regarding data privacy and how AI should be developed and used are also at different levels of maturity in Latin America. The need to share experiences among Latin American countries about developing, validating, implementing, monitoring, and acquiring AI products led to the creation of the 1° Encontro Latino-Americano de IA em Saúde (1st Latin American Meeting on AI in Health) during the 52nd Jornada Paulista de Radiologia (JPR), organized by SPR and the Radiological Society of North America (RSNA).
The Role of SPR
SPR brings together more than 9000 members, covering all areas of diagnostic imaging (4). About 50% of its members are from the state of São Paulo, and others live throughout the other states of Brazil. SPR has corresponding members in North America, Europe, and other Latin American countries. SPR is affiliated with the Colégio Brasiliero de Radiologia e Diagnóstico por Imagem (Brazilian College of Radiology and Diagnostic Imaging) and has been dedicated to the scientific development and promotion of radiology with the highest principles of professionalism. SPR has a strong, long-standing partnership with the RSNA that has contributed greatly to the growth and development of radiology in Latin America.
SPR seeks international partnerships to promote scientific exchange, develop specific events, and enrich the contact between Brazilians and professionals from different regions. SPR organizes study groups for each radiology subspecialty, which are monthly events focused on case discussions. The Grupo de Estudos de Tecnologia e Informática em Radiologia (Information Technology Study Group) focuses on imaging informatics, including AI and data science. At these meetings, participants discuss commercial solutions, testing processes, and implementation of AI solutions, as well as issues related to information systems in radiology. The Grupo de Estudos de Imagem Quantitativa (Quantitative Imaging Study Group) deals with the image acquisition standardization profiles presented by the Quantitative Imaging Biomarker Alliance and with standardization procedures for extracting features, such as those proposed by the Image Biomarker Standardization Initiative to develop, evaluate, implement, and track results of AI algorithms that generate image biomarkers (5).
The Image Update Course (Prof Dr Feres Secaf) for trainees includes a module entirely dedicated to the study of AI concepts, explaining to professionals in training what their possible interfaces with algorithms might be and how they can get involved in projects on this topic. SPR has inserted AI into its educational activities, and it aims to keep pushing for a higher level of AI literacy in the radiology community by promoting didactic content in future events.
SPR annually hosts the JPR, the largest diagnostic imaging meeting in Latin America, which includes leading radiologists from RSNA every other year. In 2022, 110 vendors showcased their products. Thirty-four courses were presented by more than 700 lecturers, 60 of whom were from outside Brazil. JPR’s Radiology Informatics track has existed for more than 10 years; recently, it was reformulated to contemplate a broader scope in technology, including AI, innovation, and health technology startups. A dedicated AI track was created at JPR 2022, with theoretical sessions, hands-on sessions (Digital Imaging and Communications in Medicine and deep learning), and the 1° Encontro Latino-Americano de IA em Saúde, which invited radiology societies from other countries in Latin America.
The scientific committee assured that the workshop’s topics varied from clinical to more technical approaches, mixing entry-level concepts—to target general radiologists, residents, and students—with deep technological discussions. The program also covered topics such as ethics, regulation, implementation, business models, and the managerial issues of dealing with AI models, products, and companies. Finally, the informatics and AI committee worked alongside other subspecialty committees to include AI lectures, panels, and specialists in traditional clinical tracks, so that mainstream radiologists were exposed to AI content. Bringing AI into the already robust clinical communities SPR has built over the years is considered essential and strategic, as it will accelerate learning and technology adoption while creating legitimacy and trust.
1º Encontro Latino-Americano de IA em Saúde
The meeting had a multidisciplinary audience, including radiologists from various subspecialties, residents and fellows, biomedical scientists, technologists, researchers, information technology specialists, engineers, and undergraduate students from various fields of knowledge. The topics discussed in the meeting included insights into the development and adoption of AI in various Latin American countries. A panel discussed the return on investment in AI in health care. A didactic interactive session was offered for mainstream radiologists. Speakers from RSNA surveyed AI’s ethical and medicolegal aspects (6), presented practical aspects of AI adoption (7), summarized RSNA’s Imaging AI in Practice demonstration (8), and considered how AI could add value to radiology.
Brazil’s Lei Geral de Proteção de Dados (General Data Protection Law), which took effect in 2021, resembles the European Union’s General Data Protection Regulation in many aspects (9). The essential elements of the law are the following: (a) respect for privacy; (b) informative self-determination; (c) freedom of expression, information, communication, and opinion; (d) inviolability of intimacy, honor, and image; (e) economic and technological development and innovation; (f) free enterprise, free competition, and consumer protection; and (g) human rights, the free development of personality, dignity, and the exercise of citizenship by natural persons. The law represents a historic milestone in the regulation of the processing of personal data in Brazil, both in physical media and digital platforms. However, it does not guide how to use AI models.
Since 2019, four draft laws have been proposed to rule the principles and guidelines of the development and use of AI in Brazil (10–13). The topic is still being addressed by a commission of jurists on whether to compile all of the proposals into a single document and open it for public consultation. In March 2022, the Brazilian Health Regulatory Agency approved a resolution to regulate software as a medical device. Among many definitions, it allows the internal use of software developed in-house, provided some requirements are met, including being classified as class I or II, having the validation documented, and not interfering in regulated medical devices (14).
Chile recently approved its first National Artificial Intelligence Policy that brings together 70 priority actions and 185 initiatives that will impact the social and economic fields and nurture talent (15). The policy has contributed to creating an ecosystem that has favored the development of data science and AI with the consolidation of research groups, the creation of postgraduate degrees, and the acceleration of programs to encourage science and technology-based ventures. The adoption of these technologies by public initiatives has made it possible to promote the solution of problems on a large scale, as in the case of Digital Hospital, a telemedicine project that provides health care access to people throughout Chile, which has implemented the first diabetic retinopathy screening program based on a computer vision system (16). The Sociedad Chilena de Radiología (Chilean Society of Radiology) has promoted the education and participation of radiologists through initiatives such as the first Virtual Course on Artificial Intelligence in Radiology and the creation of the Artificial Intelligence Chapter, which aims to promote education in the basics of AI and its adoption in clinical practice.
To increase productivity and economic growth in the long term, Argentina has taken steps toward rapid and early adoption of AI. In July 2019, the National Plan on AI was presented, with the goal of “developing and implementing AI solutions centered around the augmentation of human capacities.” The document shows that education is critical in this transformation process (17). National and international AI experts from a variety of disciplines, including radiology, participated in the International Forum on AI in April 2022, which aspired to promote the development of AI in Argentina. The forum closed with creating the Multidisciplinary Center on AI, oriented toward talent development and public-private articulation (18). At present, there is no legislation or ethical principles regulating the AI industry in Argentina.
Argentina’s two national radiology societies—Sociedad Argentina de Radiología (SAR, Argentine Society of Radiology) and Federación Argentina de Asociaciones de Radiología, Diagnóstico por Imágenes y Terapia Radiante (Argentine Federation of Associations of Radiology, Imaging Diagnosis, and Radiation Therapy)—have made progress in the last couple of years. In 2021, the national meeting (Congreso Argentino de Diagnóstico por Imágenes, CADI) held annually by both organizations convened a session entirely devoted to AI for the first time. At CADI 2022, assigned time for AI has expanded widely, including two 2-hour sessions plus a workshop. Also, this year, SAR created a committee on AI to foster critical judgment about AI among radiologists and members-in-training. The committee discusses relevant papers in the field, shares knowledge through presentations, discusses research projects, and plans the inclusion of an AI module in the radiology resident’s curriculum. With these initiatives, the committee aims to solidify the workforce that will lead the AI transformation in the field of radiology in Argentina in the forthcoming years.
At the time this writing, Mexico has no specific law regarding AI. The Federación Mexicana de Radiología e Imagen (FMRI, Mexican Federation of Radiology and Imaging) has been supporting the knowledge and understanding of the role of AI as one of the main ways radiology has to impact intercolleague communication and improve patient care. FMRI has promoted many activities to support teaching and learning about AI, such as:
Active participation in national meetings and congresses: Almost every FMRI meeting and webinar include talks and roundtables with AI topics. Participation of radiologists from FMRI in meetings of other specialties varies in issues ranging from broad aspects of the role of AI in radiology to more practical applications in specific pathologic conditions.
Fostering the inclusion of an AI curriculum in the radiology residency programs: Every time FMRI has the opportunity to participate with professors in the different sites where there are radiology residency programs, FMRI emphasizes the importance that AI will have in the future to empower our specialty. For many years, FMRI has been promoting the idea of the “visible radiologist,” with all the implications and responsibilities that it represents. FMRI believes that AI will be a factor in empowering radiologists to have a more active role as doctors, consultants to other specialists, diagnosticians, and data scientists.
Inclusion of public, private, and academic health sectors: In all its academic activities, FMRI has sought the participation of private, public, and academic radiologists and clinicians to assure that all medical professionals are ready to collaborate for the benefit of our patients.
The Asociación Colombiana de Radiología (ACR, Colombian Association of Radiology) is a professional, scientific, and academic society with a 75-year history, representing more than 1400 physicians specializing in radiology and diagnostic imaging. In November 2019, the ACR established an Artificial Intelligence Committee, which comprises an interdisciplinary group of five members with expertise in ethics, bioethics, AI, and ML: two radiologists, one computer science engineer, one bioengineer, and one rehabilitation physician. The committee’s primary goal is to promote the ethical, safe, and effective use of AI in health care. In recent years, regulatory frameworks, ethics and bioethics, clinical evidence, and medical communication have served as the four pillars supporting the growth of activities of this committee. The committee has organized national educational initiatives, including a national congress on radiology AI and ML, the AI foundations for imaging technologists, as well as AI subspecialty theoretical and practical conferences. Moreover, the AI committee works closely with the International Symposium on Biomedical Imaging to organize the clinical day of the 2023 conference. The national dissemination of evidence-based Consolidated Standards of Reporting AI Trials (ie, CONSORT-AI) (19) and Standard Protocol Items: Recommendations for AI Interventional Trials (ie, SPIRIT-AI) (20) guidelines for clinical trials involving AI is also a component of additional international collaborations.
The AI committee of the ACR has participated in national initiatives that form the foundation of the Colombian Framework for applying new technologies in health care. These include contributions to the national white papers on the AI ethical framework in Colombia (21) and the ethical and responsible use of AI in health care (22). The committee works closely with the national AI expert mission to develop recommendations for implementing national AI policies. In July 2021, the committee formed an ethics and bioethics subcommittee; consequently, in October 2021, the ACR hosted its first national forum on the ethics and implementation of AI in health care. Particular initiatives of this subcommittee include the ongoing pilot program to implement informed consent for health care solutions based on AI. The AI committee also oversees regional and national projects such as the external validation of chest radiograph models for detecting misplaced tubes and lines, multi-institutional training of models to detect lung opacities, and developing segmentation networks to estimate prostate volume; all of these are in the revision process.
The expanding ACR initiatives coincide with the government’s digital transformation agenda, establishing Colombia as one of the regional leaders in regulatory technology (23). As part of this digital transformation policy, more than 30 initiatives have been developed since 2019. These initiatives include the adoption of an ethical framework for responsible and inclusive AI development, the exploitation of data as a cornerstone of AI development, and the establishment of a competitive AI market. The ACR endeavors to promote these components into its operational pillars.
Collaboration with Other Societies
The RSNA, which has had long relationships with Latin American radiology societies, was represented by several keynote speakers at the event. Radiologists from Latin American countries have contributed datasets and have been among the winners in past RSNA ML challenges (head CT hemorrhage dataset , chest CT pulmonary embolism dataset , the RSNA International COVID-19 Open Radiology Database ). Assuring that the Latin American population is represented in such datasets is essential so that the best models created in those competitions will be generalizable to the Latin American population. Latin American radiologists are also involved in the RSNA’s Machine Learning Steering Committee and the Radiology Informatics Subcommittee of the Scientific Program Committee of the RSNA Annual Meeting.
The Society for Imaging Informatics in Medicine (SIIM), an organization of leading experts in imaging informatics and defining standards, was represented by the chair of the board of directors, who served as a keynote speaker. SIIM’s Global Track South America was created to highlight the contribution and uniqueness of South America in imaging informatics; it is a yearly virtual meeting that took place for the second time in 2022. There has been an increase in the number of volunteers from Latin America at SIIM, including in the Education Committee, the Hackathon Committee, and the Machine Learning Subcommittee.
The 1° Encontro Latino-Americano de IA em Saúde surpassed the organizers’ expectations. The program was developed with a broad scope strategy, mixing experts and expertise from different countries, backgrounds, and specialties, with the objective of encompassing all levels of complexity (from basic concepts to advanced techniques), perspectives (clinical, technical, ethical, and business management), and specialties (informatics, data science, and radiology). It was an opportunity to connect with peers from other countries and share lessons learned about AI in health care in different countries and contexts.Disclosures of conflicts of interest: F.C.K. Consultant for MD.ai and GE Healthcare; co-chair of the Machine Learning Education Subcommittee at the Society for Imaging Informatics in Medicine; member of Radiological Society of North America (RSNA) Machine Learning Steering Committee; former member of Radiology In Training trainee editorial board. F.B.P.d.N. No relevant relationships. G.E.R. Leadership or fiduciary role in the RSNA Committee on International Radiology Education, Educación Committee at Federación Mexicana de Radiología e Imagen, and Evaluación Committee at Consejo Mexicano de Radiología e Imagen. H.C. Consulting fees from Entelai. H.H.L. No relevant relationships. E.S.M. No relevant relationships. T.J. Informatics director of Sociedade Paulista de Radiologia. A.J.d.R. No relevant relationships. C.H.N. No relevant relationships.
We thank the staff of SPR and JPR for organizing the events. We thank all the societies and their organizers and representatives, as well as all the professors and speakers who made the event possible.
Author contributions: Guarantors of integrity of entire study, F.C.K., F.B.P.d.N., E.S.M., T.J., A.J.d.R., C.H.N.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, F.C.K., F.B.P.d.N., G.E.R., H.C., H.H.L., E.S.M., T.J., C.H.N.; and manuscript editing, all authors
Authors declared no funding for this work.
- 1. Gartner Hype Cycle. Gartner. https://www.gartner.com/en/research/methodologies/gartner-hype-cycle. Accessed June 13, 2022. Google Scholar
- 2. . Implementation of artificial intelligence (AI) applications in radiology: hindering and facilitating factors. Eur Radiol 2020;30(10):5525–5532. Crossref, Medline, Google Scholar
- 3. . Overview of noninterpretive artificial intelligence models for safety, quality, workflow, and education applications in radiology practice. Radiol Artif Intell 2022;4(2):e210114. Link, Google Scholar
- 4. SPR - Sociedade Paulista de Radiologia e Diagnóstico por Imagem. https://www.spr.org.br/. Accessed June 13, 2022. Google Scholar
- 5. . The Image Biomarker Standardization Initiative: Standardized quantitative radiomics for high-throughput image-based phenotyping. Radiology 2020;295(2):328–338. Link, Google Scholar
- 6. . Ethics of using and sharing clinical imaging data for artificial intelligence: a proposed framework. Radiology 2020;295(3):675–682. Link, Google Scholar
- 7. . The state of radiology AI: considerations for purchase decisions and current market offerings. Radiol Artif Intell 2020;2(6):e200004. Link, Google Scholar
- 8. . Imaging AI in practice: a demonstration of future workflow using integration standards. Radiol Artif Intell 2021;3(6):e210152. Link, Google Scholar
- 9. Lei Geral de Proteção de Dados Pessoais (LGPD) No 13.709. http://www.planalto.gov.br/ccivil_03/_ato2015-2018/2018/lei/L13709compilado.htm. Accessed June 15, 2022. Google Scholar
- 10. Projeto de Lei N° 5691, de 2019. https://legis.senado.leg.br/sdleg-getter/documento?dm=8031122&ts=1649876906724&disposition=inline. Accessed June 15, 2022. Google Scholar
- 11. Projeto de Lei N° 5051, de 2019. https://legis.senado.leg.br/sdleg-getter/documento?dm=8009064&ts=1646420511147&disposition=inline. Accessed June 15, 2022. Google Scholar
- 12. Projeto de Lei No21, de 2020. https://www.camara.leg.br/proposicoesWeb/prop_mostrarintegra?codteor=1853928&filename=PL+21/2020. Accessed June 15, 2022. Google Scholar
- 13. Projeto de Lei N° 872, de 2021. https://legis.senado.leg.br/sdleg-getter/documento?dm=8940096&ts=1651259203176&disposition=inline. Accessed June 15, 2022. Google Scholar
Imprensa Nacional. RDC No 657. https://www.in.gov.br/en/web/dou/-/resolucao-de-diretoria-colegiada-rdc-n-657-de-24-de-marco-de-2022-389603457. Accessed June 16, 2022. Google Scholar
- 15. Política Nacional de Inteligencia Artificial. Project. http://www.minciencia.gob.cl/areas-de-trabajo/inteligencia-artificial/politica-nacional-de-inteligencia-artificial/. Accessed June 15, 2022. Google Scholar
- 16. . Clinical validation of an artificial intelligence-based diabetic retinopathy screening tool for a national health system. Eye (Lond) 2022;36(1):78–85.[Published correction appears in Eye (Lond) 2021;35(10):2910.] Crossref, Medline, Google Scholar
- 17. Desconferencia sobre Inteligencia Artificial. Argentina.gob.ar. https://www.argentina.gob.ar/ciencia/desconferencia-sobre-inteligencia-artificial. Published 2019. Accessed June 15, 2022. Google Scholar
- 18. Un foro para fomentar el desarrollo de la Inteligencia Artificial en la Argentina. Argentina.gob.ar. https://www.argentina.gob.ar/noticias/un-foro-para-fomentar-el-desarrollo-de-la-inteligencia-artificial-en-la-argentina. Published 2022. Accessed June 15, 2022. Google Scholar
- 19. . Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. Lancet Digit Health 2020;2(10):e537–e548. Crossref, Medline, Google Scholar
- 20. . Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. Nat Med 2020;26(9):1351–1363. Crossref, Medline, Google Scholar
- 21. Ethical Framework for Artificial Intelligence in Colombia. OECD AI Policy Observatory Portal. https://oecd.ai/en/dashboards/policy-initiatives/http:%2F%2Faipo.oecd.org%2F2021-data-policyInitiatives-26737. Accessed June 15, 2022. Google Scholar
- 22. Uso ético y responsable de IA en salud. C4IR.CO. Centro para la Cuarta Revolución Industrial de Colombia - C4IR.CO. https://c4ir.co/uso-etico-y-responsable-de-ia-en-salud/. Published 2021. Accessed June 15, 2022. Google Scholar
- 23. Regulatory Technology for the 21st Century. Foro Económico Mundial. https://es.weforum.org/whitepapers/regulatory-technology-for-the-21st-century/. Accessed June 15, 2022. Google Scholar
- 24. . Construction of a machine learning dataset through collaboration: The RSNA 2019 Brain CT Hemorrhage Challenge. Radiol Artif Intell 2020;2(3):e190211[Published correction appears in Radiol Artif Intell 2020;2(4):e209002.]. Link, Google Scholar
- 25. . The RSNA Pulmonary Embolism CT Dataset. Radiol Artif Intell 2021;3(2):e200254. Link, Google Scholar
- 26. . The RSNA International COVID-19 Open Radiology Database (RICORD). Radiology 2021;299(1):E204–E213. Link, Google Scholar
Article HistoryReceived: June 26 2022
Revision requested: July 18 2022
Revision received: Aug 6 2022
Accepted: Aug 12 2022
Published online: Aug 31 2022