Rafael Scherer
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Rafael Scherer has a total of 43 co-authors. Here are the top co-authors by number of shared publications:
- Douglas Rodrigues da Costa
Bascom Palmer Eye Institute, University of Miami, Miami, FL, USA
6 shared publications - Felipe A. Medeiros
Bascom Palmer Eye Institute, University of Miami, Miami, USA
5 shared publications - Newton Kara-Junior
Department of Ophthalmology, Universidade de São Paulo, São Paulo, SP, Brazil
4 shared publications - Camila Ribeiro Koch
Universidade de São Paulo, São Paulo, SP, Brazil
3 shared publications Newton Kara-Junior
Univ Sao Paulo, Ophthalmol Dept, Sao Paulo, SP, Brazil
3 shared publications- Alessandro Adad Jammal
Department of Ophthalmology, University of Miami Miller School of Medicine, Bascom Palmer Eye Institute, Palm Beach Gardens, FL
2 shared publications
Latest Publications
- Open Access
- 6 January 2026
- Published by JMIR Publications Inc. in JMIR AI
- Vol. 5 (1) , e72101
Performance of a Small Language Model Versus a Large Language Model in Answering Glaucoma Frequently Asked Patient Questions: Development and Usability Study
Background: Large language models (LLMs) have been shown to answer patient questions in ophthalmology similar to human experts. However, concerns remain regarding their use, particularly related to patient privacy and potential inaccuracies that could compromise patient safety. Objective: This study aimed to compare the performance of an LLM in answering frequently asked patient questions about glaucoma with that of a small language model (SLM) trained locally on ophthalmology-specific literature. Methods: We compiled 35 frequently asked questions on glaucoma, categorized into 6 domains, including pathogenesis, risk factors, clinical manifestations, diagnosis, treatment and prevention, and prognosis. Each question was posed to both a SLM using a retrieval-augmented generation framework, trained on ophthalmology-specific literature, and to a LLM (ChatGPT 4.0, OpenAI). Three glaucoma specialists from a single institution independently assessed the answers using a 3-tier accuracy rating scale: poor (score=1), borderline (score=2), and good (score=3). Each answer received a quality score ranging from 3 to 9 points based on the sum of ratings from the 3 graders. Readability grade level was assessed using 4 formulas, such as the Flesch-Kincaid Level, the Gunning Fog Index, the Coleman-Liau Index, and the Simple Measure of Gobbledygook Index. Results: The answers from the SLM demonstrated comparable quality with ChatGPT 4.0, scoring mean 7.9 (SD 1.2) and mean 7.4 (SD 1.5), respectively, out of a total of 9 points ( P =.13). The accuracy rating was consistent overall and across all 6 glaucoma care domains. Both models provided answers considered unsuitable for health care–related information, as they were difficult for the average layperson to read. Conclusions: Both models generated accurate content, but the answers were considered challenging for the average layperson to understand, making them unsuitable for health care–related information. Given the specialized SLM’s comparable performance to the LLM, its high customization potential, lower cost, and ability to operate locally, it presents a viable option for deploying natural language processing in real-world ophthalmology clinical settings.
- 6 January 2026
- Published by JMIR Publications Inc. in JMIR AI
- Vol. 5 (1) , e72101
Open AccessPerformance of a Small Language Model Versus a Large Language Model in Answering Glaucoma Frequently Asked Patient Questions: Development and Usability Study
Background: Large language models (LLMs) have been shown to answer patient questions in ophthalmology similar to human experts. However, concerns remain regarding their use, particularly related to patient privacy and potential inaccuracies that could compromise patient safety. Objective: This study aimed to compare the performance of an LLM in answering frequently asked patient questions about glaucoma with that of a small language model (SLM) trained locally on ophthalmology-specific literature. Methods: We compiled 35 frequently asked questions on glaucoma, categorized into 6 domains, including pathogenesis, risk factors, clinical manifestations, diagnosis, treatment and prevention, and prognosis. Each question was posed to both a SLM using a retrieval-augmented generation framework, trained on ophthalmology-specific literature, and to a LLM (ChatGPT 4.0, OpenAI). Three glaucoma specialists from a single institution independently assessed the answers using a 3-tier accuracy rating scale: poor (score=1), borderline (score=2), and good (score=3). Each answer received a quality score ranging from 3 to 9 points based on the sum of ratings from the 3 graders. Readability grade level was assessed using 4 formulas, such as the Flesch-Kincaid Level, the Gunning Fog Index, the Coleman-Liau Index, and the Simple Measure of Gobbledygook Index. Results: The answers from the SLM demonstrated comparable quality with ChatGPT 4.0, scoring mean 7.9 (SD 1.2) and mean 7.4 (SD 1.5), respectively, out of a total of 9 points ( P =.13). The accuracy rating was consistent overall and across all 6 glaucoma care domains. Both models provided answers considered unsuitable for health care–related information, as they were difficult for the average layperson to read. Conclusions: Both models generated accurate content, but the answers were considered challenging for the average layperson to understand, making them unsuitable for health care–related information. Given the specialized SLM’s comparable performance to the LLM, its high customization potential, lower cost, and ability to operate locally, it presents a viable option for deploying natural language processing in real-world ophthalmology clinical settings.
- Open Access
- 2 January 2026
- Published by Elsevier in American Journal of Ophthalmology
- Vol. 284, 88-100
Artificial Intelligence-Guided Endpoint Selection for Neuroprotection Trials in Glaucoma
- 2 January 2026
- Published by Elsevier in American Journal of Ophthalmology
- Vol. 284, 88-100
Open AccessArtificial Intelligence-Guided Endpoint Selection for Neuroprotection Trials in Glaucoma
- Open Access
- 5 October 2025
- Published by Elsevier in American Journal of Ophthalmology
- Vol. 281, 395-403
RNFL Thickness in a Population-Based Cohort: The Canadian Longitudinal Study on Aging M2M (Machine-to-Machine) Study
- 5 October 2025
- Published by Elsevier in American Journal of Ophthalmology
- Vol. 281, 395-403
Open AccessRNFL Thickness in a Population-Based Cohort: The Canadian Longitudinal Study on Aging M2M (Machine-to-Machine) Study
- Open Access
AI-GUIDED ENDPOINT SELECTION FOR NEUROPROTECTION TRIALS IN GLAUCOMA
A bstract: Standard Automated Perimetry (SAP) is the mainstay for monitoring glaucoma progression and has been accepted by the U.S. Food and Drug Administration (FDA) as a trial endpoint, but only under stringent criteria of ≥7 dB loss in five pre-specified test locations. Identifying such locations a priori has remained a major barrier for neuroprotection trials. We developed an attention-based graph neural network (GNN) to predict the visual field points most likely to deteriorate (High-5) using baseline SAP data. The model was trained in the Bascom Palmer Ophthalmic Registry (BPOR; 6,996 eyes, 5,405 patients, 40,914 tests) and externally validated in the Duke Glaucoma Registry (DGR; 5,211 eyes, 3,933 patients, 31,225 tests) and the University of Washington Humphrey Visual Field dataset (UWHVF; 2,030 eyes, 1,195 patients, 10,310 tests). In internal validation, the mean slope at High-5 points among progressors was −2.16±0.80 dB/year, compared to −0.55±0.44 dB/year for Low-5 and −1.02±0.40 dB/year for mean deviation (MD). Similar results were observed in DGR (–2.05 vs −0.45 vs −0.93 dB/year) and UWHVF (−2.32 vs −0.66 vs −1.14 dB/year). High-5 showed superior discrimination of progressors from non-progressors with areas under the ROC curve of 0.883, 0.898, and 0.937 across the three cohorts, consistently outperforming MD (0.871–0.911) and Low-5 (0.668–0.731). Nearly all progressing eyes exhibited a repeatable ≥ 7 dB loss in average High-5 sensitivity during follow-up, compared to fewer than 30% when using MD. In sample size projections, High-5 increased the absolute effect size and lowered the σ 2 /Δ 2 ratio, translating into an estimated 42% reduction in required trial size compared to MD. In conclusion, this GNN-based framework enables data-driven identification of high-risk SAP locations, aligning with regulatory definitions of progression while substantially improving trial efficiency and sensitivity to detect meaningful visual field change. Funding: This work was supported in part by NIH R01 (EY036593) and by the Glaucoma Research Foundation (grant Endpoints2025MedeF).
Open AccessAI-GUIDED ENDPOINT SELECTION FOR NEUROPROTECTION TRIALS IN GLAUCOMA
A bstract: Standard Automated Perimetry (SAP) is the mainstay for monitoring glaucoma progression and has been accepted by the U.S. Food and Drug Administration (FDA) as a trial endpoint, but only under stringent criteria of ≥7 dB loss in five pre-specified test locations. Identifying such locations a priori has remained a major barrier for neuroprotection trials. We developed an attention-based graph neural network (GNN) to predict the visual field points most likely to deteriorate (High-5) using baseline SAP data. The model was trained in the Bascom Palmer Ophthalmic Registry (BPOR; 6,996 eyes, 5,405 patients, 40,914 tests) and externally validated in the Duke Glaucoma Registry (DGR; 5,211 eyes, 3,933 patients, 31,225 tests) and the University of Washington Humphrey Visual Field dataset (UWHVF; 2,030 eyes, 1,195 patients, 10,310 tests). In internal validation, the mean slope at High-5 points among progressors was −2.16±0.80 dB/year, compared to −0.55±0.44 dB/year for Low-5 and −1.02±0.40 dB/year for mean deviation (MD). Similar results were observed in DGR (–2.05 vs −0.45 vs −0.93 dB/year) and UWHVF (−2.32 vs −0.66 vs −1.14 dB/year). High-5 showed superior discrimination of progressors from non-progressors with areas under the ROC curve of 0.883, 0.898, and 0.937 across the three cohorts, consistently outperforming MD (0.871–0.911) and Low-5 (0.668–0.731). Nearly all progressing eyes exhibited a repeatable ≥ 7 dB loss in average High-5 sensitivity during follow-up, compared to fewer than 30% when using MD. In sample size projections, High-5 increased the absolute effect size and lowered the σ 2 /Δ 2 ratio, translating into an estimated 42% reduction in required trial size compared to MD. In conclusion, this GNN-based framework enables data-driven identification of high-risk SAP locations, aligning with regulatory definitions of progression while substantially improving trial efficiency and sensitivity to detect meaningful visual field change. Funding: This work was supported in part by NIH R01 (EY036593) and by the Glaucoma Research Foundation (grant Endpoints2025MedeF).
- 10 January 2025
- Published by Elsevier in Ophthalmology Glaucoma
- Vol. 8 (3) , 293-301
Effect of Blood Pressure on Rates of Progression in Focal Ischemic vs. Generalized Cup Enlargement Glaucoma Phenotypes
- 10 January 2025
- Published by Elsevier in Ophthalmology Glaucoma
- Vol. 8 (3) , 293-301
Effect of Blood Pressure on Rates of Progression in Focal Ischemic vs. Generalized Cup Enlargement Glaucoma Phenotypes
- Open Access
- 1 January 2025
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 88 (5) , 1-6
Artificial intelligence-based model for the interpretation and reporting of standard automated perimetry
Standard automated perimetry has been the standard method for measuring visual field changes for several years. It can measure an individual's ability to detect a light stimulus from a uniformly illuminated background. In the management of glaucoma, the primary objective of perimetry is the identification and quantification of visual field abnormalities. It also serves as a longitudinal evaluation for the detection of disease progression. The development of artificial intelligence--based models capable of interpreting tests could combine technological development with improved access to healthcare. In this observational, cross-sectional, descriptive study, we used an artificial intelligence-based model [Inception V3] to interpret gray-scale crops from standard automated perimetry that were performed in an ophthalmology clinic in the Brazilian Amazon rainforest between January 2018 and December 2022. The study included 1,519 standard automated perimetry test results that were performed using Humphrey HFA-II-i-750 (Zeiss Meditech). The Subsequently, 70%, 10%, and 20% of the dataset were used for training, validation, and testing, respectively. The model achieved 80% (68.23%-88.9%) sensitivity and 94.64% (88.8%-98%) specificity for detecting altered perimetry results. Furthermore, the area under the receiver operating characteristic curve was 0.93. The integration of artificial intelligence in the diagnosis, screening, and monitoring of pathologies represents a paradigm shift in ophthalmology, enabling significant improvements in safety, efficiency, availability, and accessibility of treatment.
- 1 January 2025
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 88 (5) , 1-6
Open AccessArtificial intelligence-based model for the interpretation and reporting of standard automated perimetry
Standard automated perimetry has been the standard method for measuring visual field changes for several years. It can measure an individual's ability to detect a light stimulus from a uniformly illuminated background. In the management of glaucoma, the primary objective of perimetry is the identification and quantification of visual field abnormalities. It also serves as a longitudinal evaluation for the detection of disease progression. The development of artificial intelligence--based models capable of interpreting tests could combine technological development with improved access to healthcare. In this observational, cross-sectional, descriptive study, we used an artificial intelligence-based model [Inception V3] to interpret gray-scale crops from standard automated perimetry that were performed in an ophthalmology clinic in the Brazilian Amazon rainforest between January 2018 and December 2022. The study included 1,519 standard automated perimetry test results that were performed using Humphrey HFA-II-i-750 (Zeiss Meditech). The Subsequently, 70%, 10%, and 20% of the dataset were used for training, validation, and testing, respectively. The model achieved 80% (68.23%-88.9%) sensitivity and 94.64% (88.8%-98%) specificity for detecting altered perimetry results. Furthermore, the area under the receiver operating characteristic curve was 0.93. The integration of artificial intelligence in the diagnosis, screening, and monitoring of pathologies represents a paradigm shift in ophthalmology, enabling significant improvements in safety, efficiency, availability, and accessibility of treatment.
- Open Access
- 1 January 2025
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 88 (4) , 1-6
Evaluation of the effectiveness of a portable device for diagnosing cataract
Purpose: Access to cataract treatment and diagnostic tools continues to be hindered by financial and logistical barriers. Thus, photography-based cataract analysis via portable devices offers a promising solution for the detection of cataracts in remote regions. In this study, the accuracy of a portable device that is based on the Lens Opacities Classification III System for diagnosing cataracts was analyzed. Methods: Photographs of the anterior segment of the eye were taken in a low-light environment, and the pupillary region markings were automatically delineated using infrared photography. The captured images were automatically analyzed using a convolutional neural network. The study group included patients with cataracts, and the control group included patients without cataracts. Results: A total of 270 eyes were analyzed, which included 143 eyes with cataracts and 127 control eyes. A total of 599 photos were analyzed. The isolated nuclear cataract was the most frequently detected subtype (37.5%), followed by a nuclear cataract associated with a cortical cataract (30.3%). The device’s accuracy was 88.5% (Confidence intervals (CI), 83.19%-94.69%), specificity was 84.62% (CI 71.79%-97.30%), positive predictive value was 91.78% (CI 74.36%-97.30%), and negative predictive value was 82.50% (CI 74.36%-97.30%). Conclusion: The portable device is a simplified user-friendly cataract screening technique that can interpret results in remote regions. This innovation could mitigate the occurrence of cataract-induced blindness and prevent premature surgical interventions in early-stage cataracts.
- 1 January 2025
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 88 (4) , 1-6
Open AccessEvaluation of the effectiveness of a portable device for diagnosing cataract
Purpose: Access to cataract treatment and diagnostic tools continues to be hindered by financial and logistical barriers. Thus, photography-based cataract analysis via portable devices offers a promising solution for the detection of cataracts in remote regions. In this study, the accuracy of a portable device that is based on the Lens Opacities Classification III System for diagnosing cataracts was analyzed. Methods: Photographs of the anterior segment of the eye were taken in a low-light environment, and the pupillary region markings were automatically delineated using infrared photography. The captured images were automatically analyzed using a convolutional neural network. The study group included patients with cataracts, and the control group included patients without cataracts. Results: A total of 270 eyes were analyzed, which included 143 eyes with cataracts and 127 control eyes. A total of 599 photos were analyzed. The isolated nuclear cataract was the most frequently detected subtype (37.5%), followed by a nuclear cataract associated with a cortical cataract (30.3%). The device’s accuracy was 88.5% (Confidence intervals (CI), 83.19%-94.69%), specificity was 84.62% (CI 71.79%-97.30%), positive predictive value was 91.78% (CI 74.36%-97.30%), and negative predictive value was 82.50% (CI 74.36%-97.30%). Conclusion: The portable device is a simplified user-friendly cataract screening technique that can interpret results in remote regions. This innovation could mitigate the occurrence of cataract-induced blindness and prevent premature surgical interventions in early-stage cataracts.
- Open Access
- 1 January 2024
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 87 (5) , 01-05
Analysis of a mobile learning app for ophthalmology in Brazil
Purpose: To determine and analyze the usability metrics of a free mobile learning app for ophthalmology in Brazil. Methods: Metric data from the management dashboard of the CBOQUIZ app were used. All users registered on the platform between March 2019 and June 30, 2021 were included. The number of questions answered, number of correct answers, number of questions answered and correct answers by subject area, and user performance by geographic region were analyzed. Results: There were 458 active users during the research period and 107,245 questions answered (average, 234.16 questions per user). Of the questions answered, 81,600 (75.5%) were correct and 2,645 were incorrect. The states in Brazil with the best performance were Espírito Santo, Paraiba, and Paraná. The subject area with the lowest hit rate was basic sciences (69.1%), within which embryology demonstrated the lowest hit rate (58.28%). The posterior segment had the highest number of questions answered, followed by miscellaneous topics and the anterior segment. Questions on strabismus were the least answered. Conclusion: The app was used consistently throughout the period studied, and participants adhered to this teaching modality. Performance asymmetry was observed across the Brazil states. The CBOQUIZ app can be used to homogenize ophthalmology teaching in the country.
- 1 January 2024
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 87 (5) , 01-05
Open AccessAnalysis of a mobile learning app for ophthalmology in Brazil
Purpose: To determine and analyze the usability metrics of a free mobile learning app for ophthalmology in Brazil. Methods: Metric data from the management dashboard of the CBOQUIZ app were used. All users registered on the platform between March 2019 and June 30, 2021 were included. The number of questions answered, number of correct answers, number of questions answered and correct answers by subject area, and user performance by geographic region were analyzed. Results: There were 458 active users during the research period and 107,245 questions answered (average, 234.16 questions per user). Of the questions answered, 81,600 (75.5%) were correct and 2,645 were incorrect. The states in Brazil with the best performance were Espírito Santo, Paraiba, and Paraná. The subject area with the lowest hit rate was basic sciences (69.1%), within which embryology demonstrated the lowest hit rate (58.28%). The posterior segment had the highest number of questions answered, followed by miscellaneous topics and the anterior segment. Questions on strabismus were the least answered. Conclusion: The app was used consistently throughout the period studied, and participants adhered to this teaching modality. Performance asymmetry was observed across the Brazil states. The CBOQUIZ app can be used to homogenize ophthalmology teaching in the country.
- Open Access
- 1 January 2024
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 87 (6) , 1-5
Situation of ophthalmology education in Brazil: supply versus demand
PURPOSE: The number of medical schools in Brazil has increased in recent years; however, vacancies for specialization in ophthalmology probably have not kept up with the growing demand. This study wants to estimate the increase in medical schools, the demand for ophthalmology specialization, and evaluate learning opportunities in ophthalmology. METHODS: This was a retrospective study with research from the Ministry of Education and Brazilian Council of Ophthalmology database from 2002 to 2021. These data were checked through 120 notices published by the institutions in 2021. RESULTS: The number of medical school vacancies increased by 370%, whereas the number of certified ophthalmology vacancies increased by 64%. There was an 11.4% misalignment between the Brazilian Council of Ophthalmology data in the Ministry of Education. CONCLUSION: The proportion of medical graduates has increased much more than opportunities for ophthalmology specialization. The effect on the search for unaccredited specialization positions is unknown, and policies for monitoring the specialization of ophthalmology vacancies should be established.
- 1 January 2024
- Published by GN1 Sistemas e Publicacoes Ltd. in Arquivos Brasileiros de Oftalmologia
- Vol. 87 (6) , 1-5
Open AccessSituation of ophthalmology education in Brazil: supply versus demand
PURPOSE: The number of medical schools in Brazil has increased in recent years; however, vacancies for specialization in ophthalmology probably have not kept up with the growing demand. This study wants to estimate the increase in medical schools, the demand for ophthalmology specialization, and evaluate learning opportunities in ophthalmology. METHODS: This was a retrospective study with research from the Ministry of Education and Brazilian Council of Ophthalmology database from 2002 to 2021. These data were checked through 120 notices published by the institutions in 2021. RESULTS: The number of medical school vacancies increased by 370%, whereas the number of certified ophthalmology vacancies increased by 64%. There was an 11.4% misalignment between the Brazilian Council of Ophthalmology data in the Ministry of Education. CONCLUSION: The proportion of medical graduates has increased much more than opportunities for ophthalmology specialization. The effect on the search for unaccredited specialization positions is unknown, and policies for monitoring the specialization of ophthalmology vacancies should be established.