Tak Yee Tania Tai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Tak Yee Tania Tai has a total of 60 co-authors. Here are the top co-authors by number of shared publications:
- Richard B. Rosen
Department of Ophthalmology, New York Eye and Ear Infirmary of Mount Sinai, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA
2 shared publications - Alexander Pinhas
Department of Ophthalmology, Icahn School of Medicine at Mount Sinai
1 shared publication - Alice Verticchio Vercellin
Department of Ophthalmology, Icahn School of Medicine at Mount Sinai, 1468 Madison Avenue, Annenberg 22-86, New York, NY 10029, USA
1 shared publication - Ang Li
Cleveland Clinic Cole Eye Institute, USA
1 shared publication - Anthony J. Aldave
Stein Eye Institute, David Geffen School of Medicine at UCLA, Los Angeles, CA
1 shared publication - Avnish Deobhakta
Ophthalmology, Icahn School of Medicine at Mount Sinai and New York Eye and Ear Infirmary of Mount Sinai, New York, New York, USA
1 shared publication
Latest Publications
- Open Access
- 16 November 2026
- Published by MDPI AG in Preprints.org
The Evolution of Canaloplasty: Two Decades of Restoring and Reinforcing the Conventional Outflow Pathway
Surgery of Schlemm’s canal has evolved over two decades from bleb-dependent, ab-externo viscocanalostomy into a family of tissue-sparing, ab-interno, and scaffold-reinforced procedures. We trace five generations of canal surgery and evaluate whether durable structural reinforcement of the restored canal, with preservation of the trabecular meshwork, determines long-term outcome. Imaging and histology show that the glaucomatous canal is narrowed and partially collapsed, with canal area inversely related to intraocular pressure. Viscoelastic injected into the canal is cleared within 24 hours, whereas a tensioning suture sustains canal expansion beyond 20 months and outperforms viscocanalostomy, particularly at stringent pressure targets. Ab-interno canaloplasty sustains pressure and medication reduction for 3 to 6 years but leaves no structural element in the canal. A nitinol intracanalicular scaffold has 5-year randomized evidence of benefit but provokes persistent angle reactivity, with cumulative peripheral anterior synechiae of 14.6% versus 3.7% after cataract surgery alone. Acellular scleral allograft is safe and non-migrating as a uveoscleral cleft scaffold but has not been studied inside Schlemm’s canal. Biologic intracanalicular reinforcement is a coherent but untested next step that requires randomized, morphometric evaluation.
- 16 November 2026
- Published by MDPI AG in Preprints.org
Open AccessThe Evolution of Canaloplasty: Two Decades of Restoring and Reinforcing the Conventional Outflow Pathway
Surgery of Schlemm’s canal has evolved over two decades from bleb-dependent, ab-externo viscocanalostomy into a family of tissue-sparing, ab-interno, and scaffold-reinforced procedures. We trace five generations of canal surgery and evaluate whether durable structural reinforcement of the restored canal, with preservation of the trabecular meshwork, determines long-term outcome. Imaging and histology show that the glaucomatous canal is narrowed and partially collapsed, with canal area inversely related to intraocular pressure. Viscoelastic injected into the canal is cleared within 24 hours, whereas a tensioning suture sustains canal expansion beyond 20 months and outperforms viscocanalostomy, particularly at stringent pressure targets. Ab-interno canaloplasty sustains pressure and medication reduction for 3 to 6 years but leaves no structural element in the canal. A nitinol intracanalicular scaffold has 5-year randomized evidence of benefit but provokes persistent angle reactivity, with cumulative peripheral anterior synechiae of 14.6% versus 3.7% after cataract surgery alone. Acellular scleral allograft is safe and non-migrating as a uveoscleral cleft scaffold but has not been studied inside Schlemm’s canal. Biologic intracanalicular reinforcement is a coherent but untested next step that requires randomized, morphometric evaluation.
- Open Access
- 9 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Do Calibrated-Probability Claims Hold Up in Clinical Diagnosis? A Pre-registered Evaluation of a Decision Model Against Frontier Language Models (Preprint)
BACKGROUND A new class of AI "decision models" returns a probability for every option in a single pass and is marketed as calibrated, which invites use in clinical triage. Whether those probabilities hold up on clinical diagnosis has not been independently tested. OBJECTIVE To test the diagnostic accuracy, calibration, and probabilistic coherence of a decision model (Jev 1.13, TypeSafe AI) against frontier language models, with and without reasoning. METHODS In a pre-registered study (OSF, registered 26 September 2026), we compared Jev with Claude Opus 5.5 (adaptive reasoning, always on) and Claude Opus 5 (reasoning disabled) on three case sets: the 45 Semigran symptom-checker vignettes, 1,000 stratified synthetic patients from DDXPlus with reference differentials, and 80 simulated ophthalmic complaints (40 diagnoses at two levels of detail). Each model returned a probability for every diagnosis on a fixed list. Outcomes were top-1 and top-3 accuracy, expected calibration error (ECE), Brier score, the rate of exactly zero probability on the true diagnosis, sensitivity to option order, behavior on uninformative inputs, and behavior when the true diagnosis was absent. RESULTS On DDXPlus, top-1 accuracy was 72.0% for Jev, 79.5% for Opus 5 without reasoning, and 87.4% for Opus 5.5 (both comparisons with Jev P<.001). Jev assigned exactly zero probability to the true diagnosis in 62 of 1,000 cases (6.2%); neither comparator did so in any case. Calibration error did not differ between Jev and the registered comparator (ECE difference 0.007, 95% CI -0.026 to 0.038). Offered an explicit "none of these" option, Jev declined to diagnose all 30 uninformative inputs; without that option, it assigned a mean 58% to a single diagnosis. Order sensitivity was small in all models. Given the option list with no patient information, Jev placed up to 96% on a single diagnosis, while the comparators spread probability almost evenly. On the ophthalmic complaints, accuracy did not differ (Jev 88.8%, Opus 5 91.2%, Opus 5.5 88.8%), and Jev was better calibrated than the registered comparator (ECE difference -0.081, 95% CI -0.124 to -0.018). CONCLUSIONS The decision model matched frontier models on short, distinctive ophthalmic complaints but fell well behind on multi-symptom synthetic patients, including behind a comparator without reasoning, and there it excluded the correct diagnosis outright in about 1 in 16 cases. Its probabilities signaled uncertainty only when the question offered a way to abstain. Calibration claims for clinical decision models should be tested on clinical tasks, with explicit abstention options and zero-probability rates reported. CLINICALTRIAL Open Science Framework (OSF) Registries osf.io/fpmg3; https://osf.io/fpmg3
- 9 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessDo Calibrated-Probability Claims Hold Up in Clinical Diagnosis? A Pre-registered Evaluation of a Decision Model Against Frontier Language Models (Preprint)
BACKGROUND A new class of AI "decision models" returns a probability for every option in a single pass and is marketed as calibrated, which invites use in clinical triage. Whether those probabilities hold up on clinical diagnosis has not been independently tested. OBJECTIVE To test the diagnostic accuracy, calibration, and probabilistic coherence of a decision model (Jev 1.13, TypeSafe AI) against frontier language models, with and without reasoning. METHODS In a pre-registered study (OSF, registered 26 September 2026), we compared Jev with Claude Opus 5.5 (adaptive reasoning, always on) and Claude Opus 5 (reasoning disabled) on three case sets: the 45 Semigran symptom-checker vignettes, 1,000 stratified synthetic patients from DDXPlus with reference differentials, and 80 simulated ophthalmic complaints (40 diagnoses at two levels of detail). Each model returned a probability for every diagnosis on a fixed list. Outcomes were top-1 and top-3 accuracy, expected calibration error (ECE), Brier score, the rate of exactly zero probability on the true diagnosis, sensitivity to option order, behavior on uninformative inputs, and behavior when the true diagnosis was absent. RESULTS On DDXPlus, top-1 accuracy was 72.0% for Jev, 79.5% for Opus 5 without reasoning, and 87.4% for Opus 5.5 (both comparisons with Jev P<.001). Jev assigned exactly zero probability to the true diagnosis in 62 of 1,000 cases (6.2%); neither comparator did so in any case. Calibration error did not differ between Jev and the registered comparator (ECE difference 0.007, 95% CI -0.026 to 0.038). Offered an explicit "none of these" option, Jev declined to diagnose all 30 uninformative inputs; without that option, it assigned a mean 58% to a single diagnosis. Order sensitivity was small in all models. Given the option list with no patient information, Jev placed up to 96% on a single diagnosis, while the comparators spread probability almost evenly. On the ophthalmic complaints, accuracy did not differ (Jev 88.8%, Opus 5 91.2%, Opus 5.5 88.8%), and Jev was better calibrated than the registered comparator (ECE difference -0.081, 95% CI -0.124 to -0.018). CONCLUSIONS The decision model matched frontier models on short, distinctive ophthalmic complaints but fell well behind on multi-symptom synthetic patients, including behind a comparator without reasoning, and there it excluded the correct diagnosis outright in about 1 in 16 cases. Its probabilities signaled uncertainty only when the question offered a way to abstain. Calibration claims for clinical decision models should be tested on clinical tasks, with explicit abstention options and zero-probability rates reported. CLINICALTRIAL Open Science Framework (OSF) Registries osf.io/fpmg3; https://osf.io/fpmg3
- Open Access
- 6 September 2025
- Published by MDPI AG in Diagnostics
- Vol. 15  (17) , 2256
The Effects of Age on the Human Tear Film Assessed with a Novel Imaging Device
Purpose: We aimed to analyze the effects of age on human tear film (TF) using a novel nanometer resolution TF imaging device (Tear Film Imager, TFI, AdOM, Israel). Methods: 44 healthy adult subjects (≥18 years of age) without ocular or systemic diseases or prior eye treatments with ages spanning seven decades were enrolled in this prospective cross-sectional study. Subjects underwent a comprehensive ophthalmic examination and completed the Ocular Surface Disease Index questionnaire (OSDI). All study participants underwent TF imaging using the TFI, including assessment of muco-aqueous layer thickness (MALT), lipid-layer thickness (LLT), inter-blink interval, and lipid map uniformity. Associations between TFI parameters and age were tested using linear regression (accounting for multiple eyes). Results: A total of 80 eyes (44 subjects) were imaged: 19 eyes from 10 subjects in the 3rd decade of life (aged 20–29); 10 eyes from 5 subjects in the 4th decade of life (aged 30–39); 5 eyes from 3 subjects in the 5th decade of life (40–49); 12 eyes from 7 subjects in the 6th decade of life (50–59), 19 eyes from 11 subjects in the 7th decade of life (60–69); 11 eyes from 6 subjects in the 8th decade of life (70–79); and 4 eyes from 2 subjects in the 9th decade of life (80–89). With increasing age, MALT significantly decreased (p = 0.024), and LLT significantly increased (p = 0.001). No statistically significant linear age effects were found for the other TFI parameters (p > 0.05) or the OSDI scores of study participants of different ages (p = 0.786). Conclusions: Quantitative TF biomarkers varied significantly with advancing age in healthy individuals, highlighting the importance of accounting for age in TF assessments.
- 6 September 2025
- Published by MDPI AG in Diagnostics
- Vol. 15  (17) , 2256
Open AccessThe Effects of Age on the Human Tear Film Assessed with a Novel Imaging Device
Purpose: We aimed to analyze the effects of age on human tear film (TF) using a novel nanometer resolution TF imaging device (Tear Film Imager, TFI, AdOM, Israel). Methods: 44 healthy adult subjects (≥18 years of age) without ocular or systemic diseases or prior eye treatments with ages spanning seven decades were enrolled in this prospective cross-sectional study. Subjects underwent a comprehensive ophthalmic examination and completed the Ocular Surface Disease Index questionnaire (OSDI). All study participants underwent TF imaging using the TFI, including assessment of muco-aqueous layer thickness (MALT), lipid-layer thickness (LLT), inter-blink interval, and lipid map uniformity. Associations between TFI parameters and age were tested using linear regression (accounting for multiple eyes). Results: A total of 80 eyes (44 subjects) were imaged: 19 eyes from 10 subjects in the 3rd decade of life (aged 20–29); 10 eyes from 5 subjects in the 4th decade of life (aged 30–39); 5 eyes from 3 subjects in the 5th decade of life (40–49); 12 eyes from 7 subjects in the 6th decade of life (50–59), 19 eyes from 11 subjects in the 7th decade of life (60–69); 11 eyes from 6 subjects in the 8th decade of life (70–79); and 4 eyes from 2 subjects in the 9th decade of life (80–89). With increasing age, MALT significantly decreased (p = 0.024), and LLT significantly increased (p = 0.001). No statistically significant linear age effects were found for the other TFI parameters (p > 0.05) or the OSDI scores of study participants of different ages (p = 0.786). Conclusions: Quantitative TF biomarkers varied significantly with advancing age in healthy individuals, highlighting the importance of accounting for age in TF assessments.
- Open Access
The Role of Prompt Engineering for Multimodal LLM Glaucoma Diagnosis
Background and Aim: This study evaluates the diagnostic performance of multimodal large language models (LLMs), GPT-4o and Claude Sonnet 3.5, in detecting glaucoma from fundus images. We specifically assess the impact of prompt engineering and the use of reference images on model performance. Methods: We utilized the ACRIMA public dataset, comprising 705 labeled fundus images, and designed four prompt types, ranging from simple instructions to more refined prompts with reference images. The two model were tested across 5640 API runs, with accuracy, sensitivity, specificity, PPV, and NPV assessed through non-parametric statistical tests. Results: Claude Sonnet 3.5 achieved a highest sensitivity of 94.92%, a specificity of 73.46%, and F1 score of 0.726. GPT-4o reached a highest sensitivity of 81.47%, a specificity of 50.49%, and F1 score of 0.645. The incorporation of prompt engineering and reference images improved GPT-4o’s accuracy by 39.8% and Claude Sonnet 3.5’s by 64.2%, significantly enhancing both models’ performance. Conclusion: Multimodal LLMs demonstrated potential in diagnosing glaucoma, with Claude Sonnet 3.5 achieving a sensitivity of 94.92%, far exceeding the 22% sensitivity reported for primary care physicians in the literature. Prompt engineering, especially with reference images, significantly improved diagnostic performance. As LLMs become more integrated into medical practice, efficient prompt design may be key, and training doctors to use these tools effectively could enhance clinical outcomes.
Open AccessThe Role of Prompt Engineering for Multimodal LLM Glaucoma Diagnosis
Background and Aim: This study evaluates the diagnostic performance of multimodal large language models (LLMs), GPT-4o and Claude Sonnet 3.5, in detecting glaucoma from fundus images. We specifically assess the impact of prompt engineering and the use of reference images on model performance. Methods: We utilized the ACRIMA public dataset, comprising 705 labeled fundus images, and designed four prompt types, ranging from simple instructions to more refined prompts with reference images. The two model were tested across 5640 API runs, with accuracy, sensitivity, specificity, PPV, and NPV assessed through non-parametric statistical tests. Results: Claude Sonnet 3.5 achieved a highest sensitivity of 94.92%, a specificity of 73.46%, and F1 score of 0.726. GPT-4o reached a highest sensitivity of 81.47%, a specificity of 50.49%, and F1 score of 0.645. The incorporation of prompt engineering and reference images improved GPT-4o’s accuracy by 39.8% and Claude Sonnet 3.5’s by 64.2%, significantly enhancing both models’ performance. Conclusion: Multimodal LLMs demonstrated potential in diagnosing glaucoma, with Claude Sonnet 3.5 achieving a sensitivity of 94.92%, far exceeding the 22% sensitivity reported for primary care physicians in the literature. Prompt engineering, especially with reference images, significantly improved diagnostic performance. As LLMs become more integrated into medical practice, efficient prompt design may be key, and training doctors to use these tools effectively could enhance clinical outcomes.
- 1 May 2024
- Published by Medknow in The Pan-American Journal of Ophthalmology
- Vol. 6  (3)
Elevated intraocular pressure due to spontaneous hemorrhage into the endplate capsule of a glaucoma drainage implant
Tube occlusion with elevation of the intraocular pressure (IOP) is a well-described complication of glaucoma drainage implant (GDI) surgery. Most commonly, obstruction occurs at the proximal tube tip from iris, vitreous, or blood. Further intervention is often required to remove the obstruction and reestablish aqueous outflow. Herein, we describe a patient in whom obstruction of the distal tube ostium by clotted blood from a spontaneous hemorrhage into the endplate capsule of an Ahmed Glaucoma Valve (AGV) resulted in severe IOP elevation. This is the first report in the literature of distal tube occlusion by blood from a spontaneous hemorrhage into the endplate capsule of a GDI. Prompt recognition of this entity is important for appropriate treatment planning and patient counselling.
- 1 May 2024
- Published by Medknow in The Pan-American Journal of Ophthalmology
- Vol. 6  (3)
Elevated intraocular pressure due to spontaneous hemorrhage into the endplate capsule of a glaucoma drainage implant
Tube occlusion with elevation of the intraocular pressure (IOP) is a well-described complication of glaucoma drainage implant (GDI) surgery. Most commonly, obstruction occurs at the proximal tube tip from iris, vitreous, or blood. Further intervention is often required to remove the obstruction and reestablish aqueous outflow. Herein, we describe a patient in whom obstruction of the distal tube ostium by clotted blood from a spontaneous hemorrhage into the endplate capsule of an Ahmed Glaucoma Valve (AGV) resulted in severe IOP elevation. This is the first report in the literature of distal tube occlusion by blood from a spontaneous hemorrhage into the endplate capsule of a GDI. Prompt recognition of this entity is important for appropriate treatment planning and patient counselling.
- 22 April 2022
- Published by Elsevier in Ophthalmology Glaucoma
- Vol. 5  (6) , 663-671
Head and Neck Region Dermatological Ultraviolet-Related Cancers are Associated with Exfoliation Syndrome in a Clinic-Based Population
- 22 April 2022
- Published by Elsevier in Ophthalmology Glaucoma
- Vol. 5  (6) , 663-671
Head and Neck Region Dermatological Ultraviolet-Related Cancers are Associated with Exfoliation Syndrome in a Clinic-Based Population
- 1 July 2021
- Published by Elsevier in Ophthalmology Glaucoma
- Vol. 4  (4) , 421-426
A Prospective Randomized Trial of Ahmed Glaucoma Drainage Device Implantation with or without Ologen Collagen Matrix
- 1 July 2021
- Published by Elsevier in Ophthalmology Glaucoma
- Vol. 4  (4) , 421-426
A Prospective Randomized Trial of Ahmed Glaucoma Drainage Device Implantation with or without Ologen Collagen Matrix
- Open Access
- 1 January 2018
- Published by Asia Pacific Academy of Ophthalmology in Asia-Pacific Journal of Ophthalmology
- Vol. 7  (5) , 352-355
Visual Evoked Potentials and Glaucoma
The Asia-Pacific Journal of Ophthalmology (APJO) is the official publication of the Asia-Pacific Academy of Ophthalmology. The APJO positions itself as a first-class bimonthly (6 issues per year) peer-viewed e-journal representative of ophthalmic and visual science developments in the Asia-Pacific region. The APJO welcomes the submission of original, previously unpublished manuscripts which describe clinical investigations, clinical observations and clinically relevant laboratory investigations.
- 1 January 2018
- Published by Asia Pacific Academy of Ophthalmology in Asia-Pacific Journal of Ophthalmology
- Vol. 7  (5) , 352-355
Open AccessVisual Evoked Potentials and Glaucoma
The Asia-Pacific Journal of Ophthalmology (APJO) is the official publication of the Asia-Pacific Academy of Ophthalmology. The APJO positions itself as a first-class bimonthly (6 issues per year) peer-viewed e-journal representative of ophthalmic and visual science developments in the Asia-Pacific region. The APJO welcomes the submission of original, previously unpublished manuscripts which describe clinical investigations, clinical observations and clinically relevant laboratory investigations.
- 1 August 2017
- Published by Elsevier in Advances in Ophthalmology and Optometry
- Vol. 2  (1) , 301-319
Secondary Angle Closure
- 1 August 2017
- Published by Elsevier in Advances in Ophthalmology and Optometry
- Vol. 2  (1) , 301-319
Secondary Angle Closure
- Open Access
- 1 June 2009
- Published by Wolters Kluwer Health in Cornea
- Vol. 28  (5) , 589-593
Keratoconus Associated With Corneal Stromal Amyloid Deposition Containing TGFBIp
Purpose: To report the identification and characterization of stromal amyloid deposits in patients with keratoconus. Methods: The excised corneal buttons from 2 patients diagnosed clinically with keratoconus underwent histochemical analysis with Masson trichrome, Congo red, Alcian blue, and periodic acid-Schiff stains, and immunohistochemical analysis for the transforming growth factor beta-induced gene (TGFBI) protein product (TGFBIp), prealbumin, lysozyme, and kappa and lambda light chain expression. After the collection of DNA from both patients, exons 4, 11, 12, 13 and 14 of TGFBI were amplified and sequenced to search for mutations previously associated with dystrophic corneal stromal amyloid deposition. Results: Light microscopic examination of the corneal buttons revealed stromal thinning, epithelial basement membrane abnormalities, and focal disruption of Bowman layer. Multiple stromal deposits were identified that stained red with Masson trichrome, pink with periodic acid-Schiff, and red with Congo red; the Congo red-stained deposits demonstrated birefringence and dichroism with crossed polarizing lenses. Immunohistochemical staining demonstrated reactivity of the stromal deposits with antibodies to TGFBIp but no reactivity with antibodies against prealbumin, lysozyme, or kappa and lambda light chains. Screening of TGFBI exons 4, 11, 12, 13, and 14 revealed 2 previously identified single nucleotide polymorphisms present in the heterozygous state in both individuals but no other coding region variants. Conclusions: Two cases of keratoconus with clinically unsuspected, presumed secondary stromal amyloid deposition are described. Although TGFBIp is identified in the stromal deposits, no previously reported amyloidogenic mutations are identified in TGFBI in either affected individual, indicating a previously undescribed mechanism of stromal amyloid deposition.
- 1 June 2009
- Published by Wolters Kluwer Health in Cornea
- Vol. 28  (5) , 589-593
Open AccessKeratoconus Associated With Corneal Stromal Amyloid Deposition Containing TGFBIp
Purpose: To report the identification and characterization of stromal amyloid deposits in patients with keratoconus. Methods: The excised corneal buttons from 2 patients diagnosed clinically with keratoconus underwent histochemical analysis with Masson trichrome, Congo red, Alcian blue, and periodic acid-Schiff stains, and immunohistochemical analysis for the transforming growth factor beta-induced gene (TGFBI) protein product (TGFBIp), prealbumin, lysozyme, and kappa and lambda light chain expression. After the collection of DNA from both patients, exons 4, 11, 12, 13 and 14 of TGFBI were amplified and sequenced to search for mutations previously associated with dystrophic corneal stromal amyloid deposition. Results: Light microscopic examination of the corneal buttons revealed stromal thinning, epithelial basement membrane abnormalities, and focal disruption of Bowman layer. Multiple stromal deposits were identified that stained red with Masson trichrome, pink with periodic acid-Schiff, and red with Congo red; the Congo red-stained deposits demonstrated birefringence and dichroism with crossed polarizing lenses. Immunohistochemical staining demonstrated reactivity of the stromal deposits with antibodies to TGFBIp but no reactivity with antibodies against prealbumin, lysozyme, or kappa and lambda light chains. Screening of TGFBI exons 4, 11, 12, 13, and 14 revealed 2 previously identified single nucleotide polymorphisms present in the heterozygous state in both individuals but no other coding region variants. Conclusions: Two cases of keratoconus with clinically unsuspected, presumed secondary stromal amyloid deposition are described. Although TGFBIp is identified in the stromal deposits, no previously reported amyloidogenic mutations are identified in TGFBI in either affected individual, indicating a previously undescribed mechanism of stromal amyloid deposition.