About

Igor Kozak, MD, PhD, is a professor in the Department of Ophthalmology and Vision Science at the University of Arizona College of Medicine in Tucson. He is affiliated with the Retina Department within the Ophthalmology Department. His professional address is 655 N. Alvernon Way, Suite 204, Tucson, AZ 85711. Dr. Kozak's role involves teaching, research, and patient care in the field of ophthalmology, with a focus on retinal health and vision science. His work contributes to advancing understanding and treatment in ophthalmic medicine, supporting the university's mission in medical education and research.

Research topics

• Ophthalmology
• Medicine
• Endocrinology
• Optometry
• Optics
• Physics
• Materials science

Selected publications

• Impact of baseline visual acuity, time-in-range and early treatment on functional outcomes in DMO: insights from the IRISS outcomes

  Eye · 2025-12-03 · 1 citations

  articleOpen access1st author

  BACKGROUND/OBJECTIVES: The initial visual acuity (VA) prior to treatment initiation can significantly influence long-term visual outcomes. The current analysis aimed to examine change in VA by baseline vision categories and their effects on time spent within visual change categories in patients with diabetic macular oedema (DMO) who underwent treatment with the intravitreal fluocinolone acetonide (FAc) implant. SUBJECTS/METHODS: This was a post-hoc analysis of the IRISS-Registry Data. Time-in-range (TIR) was calculated based on three VA letter-score-thresholds: ≥70, ≥65, and ≥60 ETDRS letters after treatment initiation. TIR was stratified by baseline VA in three groups: 0-33, 34-68, and 69-100 letters. The primary outcome was the mean TIR for the ≥70 letters threshold (equivalent to 6/12 in Snellen). RESULTS: A total of 671 eyes from 542 patients were included. VA improved significantly in all VA swimlane groups, with 84.8%, 71.7%, and 60.0% of eyes in the 0-33, 34-68, and 69-100 baseline VA categories, respectively, showing maintained or improved VA at 36 months (p = 0.0367). The mean TIR for the ≥70 letter threshold was significantly longer in the 69-100 letters subgroup (892.7 ± 413.4 days) compared to the 34-68 (648.4 ± 366.4 days) and 0-33 (251.3 ± 175.9 days) subgroups (p < 0.0001). No significant differences in TIR were observed based on the duration of DMO or the number of previous anti-angiogenic injections. CONCLUSIONS: Eyes with better initial VA maintained functionally better visual acuity for longer following FAc implant treatment. TIR emerged as a potentially clinically relevant endpoint for evaluating long-term treatment outcomes in DMO, offering a broader perspective than traditional VA measures.

  PublisherOA PDFDOI

• Feasibility of teleguided training of retinal laser photocoagulation to optimize education of ophthalmology residents

  Graefe s Archive for Clinical and Experimental Ophthalmology · 2025-08-06

  articleSenior authorCorresponding
  PublisherDOI

• Application of Deep Learning for Advanced Detection and Quantification of Drusen in Nonexudative AMD From Retinal Multispectral Imaging

  Translational Vision Science & Technology · 2025-11-21

  articleOpen access1st authorCorresponding

  Purpose: To propose a novel deep learning-based methodology for drusen detection and quantification in early age-related macular degeneration (AMD) using retinal multispectral images. The retinal multispectral images highlight features in several nonoverlapping spectral bands that the deep learning models leverage for automatic drusen detection and quantification in dry AMD. Methods: The proposed novel methodology comprises quality assessment of retinal images, region of interest extraction, drusen segmentation, and drusen quantification stages. Different deep learning models (such as UNet++ convolutional neural network with EfficientNetV2 encoder) have been implemented for these stages. A total of 170 drusen and 150 nondrusen retinal images (single eye) were split into four training and validation data sets to analyze the performance of a deep learning model for drusen segmentation. Results: The proposed methodology achieved an average score, recall, and precision of 0.691, 0.668, and 0.776, respectively, across all four validation sets. This work also analyzed the performance of the proposed deep learning model for discriminating drusen and drusen-like lesions, achieving a pixel-wise segmentation accuracy of 99.998%. The number and the diameter of the detected drusen were also computed. A Dice score distribution for drusen segmentation with different numbers and sizes of drusen per eye is also shown. Conclusions: This work demonstrates that deep learning models applied to retinal multispectral images can provide accurate and clinically significant drusen segmentation and quantification, thereby facilitating early detection, longitudinal monitoring, and reduction of the risk of vision loss from AMD. Translational Relevance: Deep learning-assisted detection of drusen from multispectral retinal images will refine and improve clinical diagnosis of early nonexudative age-related macular degeneration.

  PublisherOA PDFDOI

• Simultaneous GA and CNV/MNV: incidence, characteristics, and treatments

  Graefe s Archive for Clinical and Experimental Ophthalmology · 2025-04-14 · 2 citations

  reviewOpen accessSenior author

  PURPOSE: Understanding the clinical characteristics and underlying mechanisms of simultaneous geographic atrophy (GA) and choroidal neovascularization (CNV)/macular neovascularization (MNV) is necessary for the long-term management of late age-related macular degeneration (AMD) in clinical practice. METHODS: The authors reviewed the literature on the incidence, risk factors, and clinical characteristics of simultaneous GA and CNV/MNV and developed consensus recommendations for the diagnosis, assessment, and management of simultaneous GA and CNV/MNV in clinical practice. RESULTS: The incidence rate of CNV/MNV in eyes with GA is reported as 7.4% per patient-year or 13.8% in 4.1 years, while that of macular atrophy (MA) subsequent to CNV/MNV is reported as 24.4% to 37% in 24 months. Recent studies using optical coherence tomography angiography (OCT-A) revealed the presence of subclinical CNV/MNV in 11% to 16% of eyes with GA. Fundus autofluorescence is used to detect MA; optical coherence tomography (OCT) and OCT-A are useful for detecting MA, especially around the fovea, with OCT-A offering high sensitivity and specificity in the detection of both MA and CNV/MNV. GA and CNV/MNV share the genetic risk factors of HTRA1, complement factor H, complement factors 3 and 2, and ARMS2, and clinical risk factors of large drusen, cuticular drusen, intraretinal hyperreflective foci, and subretinal drusenoid deposits, suggesting that simultaneous GA and CNV/MNV represents a continuum of AMD. Anti-vascular endothelial growth factor therapy for CNV/MNV is reported to have no impact on the speed or magnitude of MA development or enlargement. An association has been observed between CNV subtype and MA progression, with the latter being slower in the presence of type 1 CNV/MNV. CONCLUSIONS: These findings suggest there is a high probability of coexistence of GA and CNV/MNV and that they should not be considered separately. Future clinical studies should assess the two conditions simultaneously using OCT and OCT-A. KEY MESSAGES: What is known Owing to differences in their clinical characteristics, geographic atrophy (GA) and choroidal neovascularization (CNV)/macular neovascularization (MNV) have historically been regarded as two separate entities; however, several cases of coexistent GA and CNV/MNV have been reported recently in the published literature. What is new The findings of this review confirm that GA and CNV/MNV share common genetic risk factors and clinical characteristics, and suggest that these two entities are part of a continuum of late-stage age-related macular degeneration (AMD). The potential for GA and CNV/MNV to coexist should be considered in any discussion of the long-term management of late AMD; moreover, clinicians should assess for CNV/MNV in patients with GA, and for GA in those with CNV/MNV, using multimodal imaging.

  PublisherOA PDFDOI

• Laser Prophylaxis for Dry Age-Related Macular Degeneration: Current Evidence

  Journal of Clinical Medicine · 2025-06-11

  reviewOpen accessSenior author

  PURPOSE: To review the role of prophylactic panmacular laser treatment for age-related macular degeneration (AMD). METHOD: A narrative review of studies employing laser treatment for non-exudative ("dry") AMD listed in the PubMed database. RESULTS: In multiple published studies, macular laser treatment that causes laser-induced retinal damage (LIRD) has shown either no overall benefit or has accelerated disease progression and vision loss in eyes with dry AMD, particularly in high-risk eyes with more severe pre-treatment disease. Conversely, other studies, including randomized and matched propensity-scored real-world-data (RWD) clinical studies, indicate that avoidance of LIRD may allow laser to be used safely and effectively to prevent vision loss from AMD by slowing disease progression and preventing neovascular conversion without adverse treatment effects, benefiting the highest-risk eyes most. CONCLUSIONS: AMD is the leading cause of irreversible vision loss worldwide. Current evidence suggests that laser prophylaxis for dry AMD may hold great promise in preventing age-related vision loss. More studies, especially prospective clinical trials, from more investigators, are needed. If current evidence is confirmed, laser prophylaxis would rise to a preeminent position as the safest, most effective, and most economical and accessible approach to reducing vision loss from AMD.

  PublisherOA PDFDOI

• Development and standardisation of ‘time-in-range’ measurement for clinical endpoints in retinal diseases

  Eye · 2025-12-16

  reviewOpen accessSenior author

  Clear and updated endpoints are required to measure outcomes of a disease course and/or a therapeutic intervention. The aim of this review is to identify a reliable 'time-in-range' endpoint of clinical outcomes in ocular conditions, with a particular focus on exudative diseases involving the posterior pole of the eye, and to explore possible applications of this endpoint. A PubMed search was carried out pertaining to: 'time-in-range', 'clinical-outcome', 'clinical-endpoint', 'clinical trial', 'metrics', 'retina', 'retinopathy', 'macular-oedema', 'maculopathy', 'ophthalmology', 'visual-function', 'visual acuity end-point' and 'OCT'. The results showed that both functional and morphological endpoints have been used in the evaluation of retinal diseases. At present, the most widely accepted and clinically meaningful marker of ocular disease is 'mean change' in best corrected visual acuity (BCVA). While comparisons to baseline at various timepoints are commonly recommended to evaluate statistical and clinically relevant differences, few metrics capture the disease course continuously over time. In other medical fields, 'time-in-range' has been introduced to provide more complete information on the fluctuations characterising the course of a disease. The application of 'time-in-range' on BCVA in exudative diseases involving the posterior pole seems feasible, reliable and applicable in clinical practice. BCVA 'time-in-range' offers a useful and practical endpoint in retinal diseases, evaluating both visual function at the end of an observation/treatment and fluctuations in disease over time. It may also be applied to other clinical and morphological endpoints in ocular diseases, including macular thickness. This review presents a hypothesis-generating framework proposing 'time-in-range' as a supplementary metric, pending prospective validation.

  PublisherOA PDFDOI

• Time-in-Range Analysis of Responses after Intravitreal Dexamethasone Therapy in Eyes with Diabetic Macular Edema

  Ophthalmology Science · 2025-05-26 · 3 citations

  articleOpen access1st authorCorresponding

  Purpose: Time in range (TIR) is a novel end point that assesses the time during which an outcome remains within a predetermined range. Because the range includes normal parameters, it is indicative of clinically meaningful benefit. The MEAD trial comprised two 3-year randomized, multicenter, sham-controlled, phase III clinical studies that evaluated the efficacy and safety of dexamethasone intravitreal implant (DEX-I) in patients with diabetic macular edema. Dexamethasone intravitreal implant significantly improved best-corrected visual acuity (BCVA) and central retinal thickness (CRT) compared with sham treatment. We present a post hoc analysis of the MEAD trial to investigate TIR benefit across various thresholds of BCVA and CRT with DEX-I versus sham. Design: Post hoc analysis of results from the randomized, multicenter, sham-controlled, phase III MEAD trial (NCT00168337 and NCT00168389). Participants: Adults with type 1 or 2 diabetes mellitus and fovea-involved macular edema associated with diabetic retinopathy. Intervention: Intravitreal injection of DEX-I 0.7 mg or sham procedure. Main Outcome Measures: Time in range during year 1 was evaluated using BCVA thresholds of ≥69, ≥51, ≥59, and ≥64 letters, and CRT thresholds of <300, <353, <446, and <551 μm (the latter cutoffs being quartile [Q] 1, Q2, and Q3 of pooled baseline BCVA and CRT, respectively). Results: 0.001). Conclusions: Patients receiving intravitreal DEX-I 0.7 mg had a longer time with BCVA above the driving threshold and below the normal limit of CRT during year 1 of the MEAD trial versus those who received sham. These results suggest that patients treated with dexamethasone experience a longer time with clinically meaningful outcomes than with sham, such as being able to drive or regaining normal structural retinal features. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

  PublisherDOI

• Long-term follow-up demonstrates change in conformation shape of the focal choroidal excavation lesions

  BMC Ophthalmology · 2024-04-02

  articleOpen accessSenior author

  PURPOSE: This study aims to present long-term observation of 5 eyes with focal choroidal excavation (FCE), focusing on morphological changes in conformity of the lesion. METHODS: A retrospective case series was conducted, including 5 eyes of 5 patients with FCE. The study utilized multimodal imaging including color fundus photography, optical coherence tomography (OCT), enhanced depth imaging OCT (EDI-OCT), fundus fluorescein angiography (FFA), fundus autofluorescence (FAF), red free imaging, and OCT angiography. RESULTS: The mean age at diagnosis was 51 ± 10.65 years, with a mean follow-up period 37 ± 13.59 months. All cases were unilateral, with 1 presenting FCE as an isolated lesion, and one patient exhibiting 2 FCEs in one eye. The mean choroidal thickness measured by EDI-OCT was 268.2 ± 63.39 μm in the affected eye. One patient displayed choroidal thickening and pachyvessels. Of the 5 eyes, one had conforming and 4 non-conforming FCE. We observed a conversion in conformity in all patients, with 4 cases transitioning from non-conforming FCE to conforming type (3 spontaneously, 1 treatment-induced). In conforming FCE, a hyporeflective space appeared twice between neuroretina and retinal pigment epithelium with spontaneous regression. CONCLUSION: We observed change in shape from the conforming to non-conforming FCE and vice versa in all patients. We consider this small change in the hyporeflective space as non-pathologic and clinically insignificant.

  PublisherOA PDFDOI

• Variability of Thermal Subthreshold Retinal Laser Treatment Plans

  Research Square · 2024-06-07

  preprintOpen access

  <title>Abstract</title> Purpose To investigate the variability in subthreshold laser treatment plans for patients with diabetic macular edema or central serous chorioretinopathy. Methods Diagnostic images from 20 patients were utilized, and 25 retina specialists generated subthreshold treatment plans along with a self-rated experience level. Evaluation comprised of i)Area Variability(AV): quantifies the consensus regarding the covered area and is the difference between the areas planned by 75% and 25% of the participants ii)mean Centroid Distance(CD): level of agreement on the localization of a treatment. Subgroup analysis investigated the impact of participants' experience levels, utilizing the Mann-Whitney-Wilcoxon test. Results The predominant plan style is a targeted treatment approach(92%) and avoidance of subfoveal region(89%). Mean CD is 71.0 ± 37.5 pixels (≈half disc diameter) and mean AV is 9.8%±8.9%. A slight difference is observed between the 50%-75% areas, but a notable distinction exists between the 25%-50% areas. Subgroup analysis revealed CD and AV value of 75.9 pixels and 24% in the lower experience level group as opposed to 55.9 pixels and 8.6% in the higher experience level group. Conclusions There is significant variability in treatment planning which reduces with increased experience of retina specialists. While consensus is observed around focal points, differences in the surrounding extents persist.

  PublisherOA PDFDOI

• Novel ocular observations in a child with Joubert syndrome type 6 due to pathogenic variant in TMEM67 gene

  American Journal of Ophthalmology Case Reports · 2024-06-15 · 1 citations

  articleOpen accessSenior authorCorresponding

  To describe unique ocular features in a child with Joubert syndrome type 6. A 4-year-old male patient presented with right microphthalmia and non-dilating pupil and left primary position nystagmus. Brain MRI revealed a “molar tooth sign” of the midbrain and a “batwing sign” of the fourth ventricle along with large retroorbital cysts bilaterally. The diagnosis of autosomal recessive Joubert syndrome type 6 due to homozygous pathogenic variant c.725A > G p. (Asn242Ser) in TMEM67 gene was confirmed by whole exome sequencing. Left eye had nystagmus and the left optic nerve and retina showed epipapillary and subretinal fibrosis, respectively. Scleral buckle was performed for left non-rhegmatogenous retinal detachment which then improved and has been stable. We present a rare case of JS with some unique ophthalmic features which expand clinical knowledge on this complex systemic and ocular entity.

  PublisherDOI

Frequent coauthors

• J. Fernando Arévalo

  271 shared

• Jay Chhablani

  267 shared

• Marcos Pereira de Ávila

  257 shared

• Guillermo Salcedo-Villanueva

  Asociacion Para Evitar la Ceguera en México Hospital Dr. Luis Sánchez Bulnes

  256 shared

• José A. Roca

  256 shared

• Maurício Maia

  256 shared

• Josep Badal

  Althaia

  256 shared

• Roberto Gallego‐Pinazo

  ACT Government

  256 shared

Labs

• McKay LabPI

Education

• Master's in Clinical Research, School of Medicine

  University of California San Diego

  2007

• PhD

  University of Veterinary Medicine in Košice

  2003

• Medical Doctor

  University of Pavol Jozef Šafárik

  1995