About

Ruohao Chen is a Professional Program Development Researcher at Teachers College, Columbia University. He is affiliated with the Organization & Leadership department, specifically within the Adult Learning and Leadership program. His office is located at 210A Zankel Hall, and he can be contacted via phone at 212-678-3703. The biography indicates his role involves professional program development, contributing to the academic and leadership initiatives at Teachers College.

Research topics

• Medicine
• Internal medicine
• Virology
• Family medicine
• Demography

Selected publications

• RETINAL DETACHMENT AFTER ENDOGENOUS ENDOPHTHALMITIS

  Retina · 2026-02-18

  article

  PURPOSE: To assess factors associated with retinal detachment (RD) in eyes presenting with endogenous endophthalmitis (EE). METHODS: Multicenter retrospective study of individuals with an EE diagnosis from seven tertiary care centers. Demographics, medical history, ocular examination findings, microbial cultures (blood or intraocular), and management strategies were assessed relative to subsequent RD occurring within 6 months of EE diagnosis. Of note, this study did not distinguish between intraocular cultures without growth and eyes that were not sampled for an intraocular organism. RESULTS: A total of 201 eyes from 171 patients presented with EE. Of these, 22.4% (45/201) had a subsequent RD. Eyes that experienced a subsequent RD had worse presenting visual acuity ( P = 0.006) and lower intraocular pressure at presentation (RD: 13.7 ± 5.4 mmHg vs. No RD: 15.7 ± 6.5 mmHg, P = 0.04). Compared with other forms of bacteremia, S. aureus bacteremia was associated with increased odds of subsequent RD (OR: 6.10, CI: 1.82-20.41, P = 0.003). Eyes with a subsequent RD did not have significantly greater rates of therapeutic pars plana vitrectomy (RD: 11/48, 22.9% vs. No RD: 36/153, 23.5%). CONCLUSION: Eyes presenting with visual acuity ≤ light perception or S. aureus bacteremia had a higher risk of subsequent RD and may warrant closer surveillance. Although no therapeutic strategies were associated with reduced odds of subsequent RD, a pars plana vitrectomy was not associated with an increased risk of RD, suggesting that it may aid in eyes with lackluster responses to systemic and intravitreal pharmacotherapy.

  PublisherDOI

• Reply

  Ophthalmology Science · 2026-03-17

  articleOpen access

  We thank Dr Bair for the thoughtful and constructive response to our paper and for extending the discussion to the critical issue of post-deployment performance monitoring for ophthalmic artificial intelligence (AI) systems.As highlighted in the letter, there is growing recognition across medicine that clinical AI models may exhibit performance drift over time due to changes in patient populations, imaging devices, acquisition protocols, and clinical workflows.Several groups have documented these challenges during clinical deployment. 1 This growing body of work highlights the importance of extending AI evaluation beyond initial deployment to include ongoing surveillance and quality assurance.We agree that although regulatory frameworks and post-market surveillance concepts are evolving, they are becoming increasingly concrete.A notable example is the Food and Drug Administration's addition of Predetermined Change Control Plans (PCCP), which provide a formal pathway for manufacturers to prespecify and implement model updates, such as retraining or fine-tuning, without requiring new marketing submissions. 2However, these frameworks can still feel abstract from the perspective of the practicing ophthalmologist.This reality underscores the importance of empowering clinicians with practical tools, such as critical evaluation of datasets and a clear understanding of performance metrics, core principles emphasized in our original framework that remain relevant across all phases of model development, validation, and clinical use. 3 In practice, responsibility for continuous monitoring is likely to fall on institutions and clinical departments in addition to centralized regulatory bodies.Artificial intelligence performance is highly sensitive to local factors, including device vendor, acquisition protocols, patient demographics, and site-specific fine-tuning, many of which can be structured under formal PCCPs.Technical approaches, including distribution shift detection, domain adaptation, and retraining, may help identify and address performance drift, but these methods will also require careful local validation and oversight. 4e therefore support the development of institutional AI governance structures that parallel, but extend beyond, traditional institutional review board oversight for AI systems deployed within a given health system or clinical environment, where performance may depend on local patient populations, imaging devices, and workflows.Similar to institutional review board continuing reviews, hospital-level AI governance or algorithm review boards could require that deployed AI systems be logged, periodically reviewed, and accompanied by updated documentation such as "living model cards" or internal performance audits. 5Such committees could provide a formalized mechanism for annual review, monitoring for domain shift, and triggering recalibration or suspension when performance degrades.These local efforts would complement federal PCCP protocols by ensuring that preauthorized model updates perform as expected within specific clinical environments.Importantly, such oversight mechanisms should be streamlined and pragmatic to minimize administrative burden.Ultimately, clinicians bear the responsibility for safe integration of these technologies into

  PublisherOA PDFDOI

• Conditional Diffusion Model Optimization for Real-Time Point-of-Care Image-Quality Enhancement on Nvidia Jetson Orin Nano

  2026-04-08

  article

  High-resolution optical coherence tomography (OCT) is often inaccessible in low-resource settings. While diffusion models can enhance low-cost OCT scans, their computational cost prohibits embedded deployment. We solve this problem by presenting the first real-time deployment of a diffusion-based OCT enhancement model on an Nvidia Jetson Orin Nano. Our systematic pipeline achieves a 33 ms inference time by first pruning the model from 224 M (million) to 42.9 M parameters (80.8% reduction) with minimal fidelity loss (no statistically significant difference in quality (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p&gt;0.3$</tex>) ). We then apply a novel hybrid INT8 quantization approach, strategically offloading unstable Softmax operations to a lightweight custom CUDA kernel. Expert evaluation shows that downstream diagnostic utility of the resulting images is also not impacted. This work provides a concrete blueprint for translating large generative models into practical, affordable medical diagnostic tools.

  PublisherDOI

• Delayed-onset Staphylococcus epidermidis endophthalmitis masquerading as chronic anterior uveitis 15 years after cataract surgery – case report

  Case Reports in Ophthalmology · 2026-04-18

  articleOpen access

  Purpose: To report a case of chronic postoperative endophthalmitis that presented 15 years after cataract surgery. Case report: A 73-year-old man presented with ocular inflammation and blurred vision. He was diagnosed with bilateral chronic uveitis by his previous providers. The patient’s ocular inflammation persisted despite multiple immunosuppressive modalities, both topical and systemic. An infectious etiology was suspected, and an intraocular lens/Soemmering’s ring explanation surgery was performed. Specimens were sent to pathology and microbiology. Results: Pathology and microbiology reports came back positive for staphylococcus epidermidis from both the IOL and the capsular bag. Discussion/Conclusion: To the best of our knowledge this is a rare and unique case of culturally proven Staphylococcus epidermidis postoperative endophthalmitis diagnosed 20 years after the inoculating surgery. This case demonstrates the challenges of managing a severe case of ocular inflammation and the resulting increased intraocular pressure not responding to multiple modalities of topical and systemic immunosuppressive medications.

  PublisherDOI

• A Practical Guide to Evaluating Artificial Intelligence Imaging Models in Scientific Literature

  Ophthalmology Science · 2025-06-09 · 6 citations

  articleOpen access

  Objective: Recent advances in artificial intelligence (AI) are revolutionizing ophthalmology by enhancing diagnostic accuracy, treatment planning, and patient management. However, a significant gap remains in practical guidance for ophthalmologists who lack AI expertise to effectively analyze these technologies and assess their readiness for integration into clinical practice. This paper aims to bridge this gap by demystifying AI model design and providing practical recommendations for evaluating AI imaging models in research publications. Design: Educational review: synthesizing key considerations for evaluating AI papers in ophthalmology. Participants: This paper draws on insights from an interdisciplinary team of ophthalmologists and AI experts with experience in developing and evaluating AI models for clinical applications. Methods: A structured framework was developed based on expert discussions and a review of key methodological considerations in AI research. Main Outcome Measures: A stepwise approach to evaluating AI models in ophthalmology, providing clinicians with practical strategies for assessing AI research. Results: This guide offers broad recommendations applicable across ophthalmology and medicine. Conclusions: As the landscape of health care continues to evolve, proactive engagement with AI will empower clinicians to lead the way in innovation while concurrently prioritizing patient safety and quality of care. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

  PublisherOA PDFDOI

• Using artificial intelligence to screen for retinal tears in patients with acute, symptomatic posterior vitreous detachments

  2025-03-20

  article
  PublisherDOI

• Optic disc pit: A trigger for secondary multiple evanescent white dot syndrome

  Retinal Cases & Brief Reports · 2025-04-23 · 3 citations

  article

  PURPOSE: We describe a case of secondary multiple evanescent white dot syndrome (MEWDS) with optic disc pit as the underlying triggering pathology. METHODS: Observational case report. RESULTS: A 41-year-old well man with a background of right optic disc pit presented with right eye loss of vision and photopsia. Visual acuity at presentation was 20/100. He was found to have clinical and multimodal imaging features consistent with right unilateral MEWDS, and the spatial distribution of lesions made it likely that the optic disc pit was the trigger. Fundus autofluorescence revealed hyperautofluorescent 'spots' that gradually faded over serial imaging, with the complete absence of hyperautofluorescent 'dots'. Over three months of follow-up visual acuity recovered to 20/25. CONCLUSION: Optic disc pit may act as a trigger for secondary MEWDS, which is an increasingly recognized epiphenomenon. Secondary MEWDS may present as 'spots without dots', and we hypothesize that this feature may differentiate it from primary MEWDS.

  PublisherDOI

• Anterior Uveitis From Retained Perfluorocarbon Liquid in the Anterior Chamber, A Case Report

  International Medical Case Reports Journal · 2025-03-01

  articleOpen accessSenior author

  Purpose: Perfluorocarbon liquids (PFCL) are routinely used in vitreoretinal surgery to flatten the retina and push subretinal fluid. Perfluorocarbon liquid retention is a common complication of vitreoretinal surgery. While retention in the posterior chamber can lead to an inflammatory response, PFCL in the anterior chamber has been shown to be generally well tolerated. Patients and Methods: This report summarizes the findings and treatment of a patient with pathologic myopia and a history of retinal detachment in the left eye repaired with pars plana vitrectomy presenting with one week of blurry vision in the left eye. He was found to have an initial intraocular pressure of 54 mmHg in the left eye along with 1+ conjunctival injection, fine and medium-sized keratic precipitates on the cornea, and 30 cells per high power field. Additionally, there was a retained perfluorocarbon liquid bubble in the anterior chamber from his original retinal detachment repair surgery. Results: After being started on topical steroids and pressure lowering medications, the patient had persistent inflammation and elevated intraocular pressures. The patient's anterior uveitis and elevated intraocular pressure resolved with removal of the perfluorocarbon liquid bubble by anterior chamber tap. Conclusion: Retained PFCL in the anterior chamber can cause anterior uveitis and can be promptly treated by removal.

  PublisherOA PDFDOI

• Visual Outcomes in Cases of Endogenous Endophthalmitis

  Ophthalmology Retina · 2025-07-22 · 3 citations

  article
  PublisherDOI

• Acute Bilateral Scleritis After Recent Streptococcal Infection

  Annals of Internal Medicine Clinical Cases · 2025-02-01

  articleOpen accessSenior author

  A 67-year-old woman with no prior ocular history and culture-confirmed streptococcal pharyngitis presented with bilateral eye redness, pain with eye movement, swelling, and blurry vision. Her condition worsened with initial conservative treatment. Repeat imaging indicated bilateral scleritis, and laboratory test results showed elevated antistreptolysin-O titers. After steroid treatment, her vision returned to baseline and symptoms resolved. Acute scleritis is important to consider in a patient with recent streptococcal infection despite penicillin treatment. Besides ocular findings, patients may not present with dermatologic or rheumatologic symptoms. Internists should consider repeat imaging and laboratory work-up, as early steroid treatment can rapidly improve the scleritis and recover vision.

  PublisherDOI

Recent grants

• P30 - Core Grant for Vision Research

  NIH · $13.2M · 1997–2030

Frequent coauthors

• Mark P. Breazzano

  University of Rochester Medical Center

  46 shared

• Aliaa H. Abdelhakim

  27 shared

• Lora R. Dagi Glass

  Columbia University

  23 shared

• Omar Moussa

  Columbia University

  19 shared

• Bryan J. Winn

  University of California, San Francisco

  19 shared

• George A. Cioffi

  Columbia University

  16 shared

• Jason Horowitz

  Columbia University Irving Medical Center

  14 shared

• Jeffrey M. Liebmann

  13 shared