Denizcan Özizmirliler
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Denizcan Özizmirliler has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Ceren Durmaz Engin
İzmir Democracy University Buca Seyfi Demirsoy Education and Research Hospital
5 shared publications - Canan Asli Utine
Department of Ophthalmology, Dokuz Eylul University Faculty of Medicine, Izmir, Turkey
2 shared publications - Meltem Söylev Bajin
Department of Ophtalmology, Dokuz Eylul University School of Medicine, Izmir, Turkey
2 shared publications - Mustafa Alper Selver
Department of Electrical and Electronics Engineering, Dokuz Eylul University, Izmir, Turkey
2 shared publications - Ufuk Besenk
Dokuz Eylül University Faculty of Engineering, Department of Electrical and Electronics Engineering, İzmir, Türkiye
2 shared publications - Ziya Ayhan
Department of Ophthalmology, Dokuz Eylül University, Izmir, Turkey
2 shared publications
Latest Publications
- Open Access
MCG-Net: Multiscale Residual Context Learning for Retinal OCT Classification with External Model Comparison
Background and Objectives: Retinal optical coherence tomography (OCT) classification requires recognition of subtle lesions within their anatomical surroundings, yet expanding contextual coverage can obscure local structural information. MCG-Net integrates local morphology and multiscale context within a gated convolutional hierarchy. Methods: A retained depthwise local pathway is augmented by two compressed separable context branches with bounded channel-wise residual scales. This study distinguishes choroidal neovascularization, diabetic macular edema, drusen, and normal retina. OCT-C8 provides development images; independent CAVRI-H5 and OCTDL datasets provide 384 external observations. Both external datasets remain excluded from parameter fitting, development validation, and checkpoint selection, but their ablation results inform selection of the Full 11/23 architecture. Paired comparisons, source-specific assessment, and visual error analysis characterize this external model comparison. Results: MCG-Net achieved accuracy, macro-F1, and Matthews correlation coefficient. Combined context improved macro-F1 over the local baseline by 2.18 percentage points, with a paired 98% interval of 0.50–4.04 and adjusted . Exploratory comparisons identified differences against six alternatives, while advantages over the strongest comparators remained uncertain. Source-specific accuracy was 90.86% on CAVRI-H5 and 95.49% on OCTDL; drusen remained the most difficult category. The model contains 26.82 million parameters and has 4.28 ms median inference latency on an RTX 5090. Conclusions: Compressed residual context strengthens gated convolutional OCT classification without replacing local processing. Reuse of the external cohort for architecture selection limits confirmatory interpretation. Untouched external validation, complete-development patient metadata, and representative clinical sampling are needed before clinical use.Open AccessMCG-Net: Multiscale Residual Context Learning for Retinal OCT Classification with External Model Comparison
Background and Objectives: Retinal optical coherence tomography (OCT) classification requires recognition of subtle lesions within their anatomical surroundings, yet expanding contextual coverage can obscure local structural information. MCG-Net integrates local morphology and multiscale context within a gated convolutional hierarchy. Methods: A retained depthwise local pathway is augmented by two compressed separable context branches with bounded channel-wise residual scales. This study distinguishes choroidal neovascularization, diabetic macular edema, drusen, and normal retina. OCT-C8 provides development images; independent CAVRI-H5 and OCTDL datasets provide 384 external observations. Both external datasets remain excluded from parameter fitting, development validation, and checkpoint selection, but their ablation results inform selection of the Full 11/23 architecture. Paired comparisons, source-specific assessment, and visual error analysis characterize this external model comparison. Results: MCG-Net achieved accuracy, macro-F1, and Matthews correlation coefficient. Combined context improved macro-F1 over the local baseline by 2.18 percentage points, with a paired 98% interval of 0.50–4.04 and adjusted . Exploratory comparisons identified differences against six alternatives, while advantages over the strongest comparators remained uncertain. Source-specific accuracy was 90.86% on CAVRI-H5 and 95.49% on OCTDL; drusen remained the most difficult category. The model contains 26.82 million parameters and has 4.28 ms median inference latency on an RTX 5090. Conclusions: Compressed residual context strengthens gated convolutional OCT classification without replacing local processing. Reuse of the external cohort for architecture selection limits confirmatory interpretation. Untouched external validation, complete-development patient metadata, and representative clinical sampling are needed before clinical use. - Open Access
- 23 September 2026
- Published by Springer Nature in BMC Ophthalmology
- Vol. 26 (1) , 593
Age-related changes in corneal parameters in the Turkish population: a Pentacam-based study
Purpose: To evaluate anterior and posterior corneal surfaces and anterior segment parameters in the Turkish population using Pentacam, and to analyze age-related changes. Methods: This retrospective study included healthy participants aged 3–86 years who underwent Pentacam imaging at Dokuz Eylül University Faculty of Medicine. Participant-level descriptive summaries were based on right-eye measurements, whereas inferential analyses incorporated both eyes using linear mixed-effects models with a participant-level random intercept. Age was evaluated both categorically (3–19, 20–39, 40–59, and ≥ 60 years) and continuously using centered linear and quadratic age terms. Results: A total of 1202 eyes of 601 participants (49.1% male, 50.9% female; mean age 35.25 ± 23.02 years) were included. Age-group effects were supported for anterior K2, Kmean, Rflat, Rsteep, Rmean, anterior astigmatism magnitude and asphericity, posterior K1, posterior astigmatism magnitude and asphericity, thinnest corneal point, corneal volume, anterior chamber volume, and anterior chamber depth (nominal p < 0.05). Age effects for anterior K1, posterior K2, posterior Rsteep, and anterior and posterior steep astigmatism axes were laterality-dependent. Continuous-age models demonstrated quadratic trajectories for anterior K1, K2, Kmean, Rflat, Rsteep, Rmean, and anterior asphericity, with approximate turning points between 32 and 46 years. Thinnest corneal point, corneal volume, anterior chamber volume, and anterior chamber depth decreased with increasing age; the age slopes for anterior chamber volume and depth differed by sex. Conclusions: Pentacam-derived corneal and anterior segment parameters show non-uniform age-related variation. The most consistent changes involved anterior corneal curvature and shape, corneal and anterior chamber dimensions, whereas several posterior corneal and astigmatic-axis measures were laterality-dependent. These findings support age-aware interpretation of corneal tomography rather than a single uniform aging pattern.
- 23 September 2026
- Published by Springer Nature in BMC Ophthalmology
- Vol. 26 (1) , 593
Open AccessAge-related changes in corneal parameters in the Turkish population: a Pentacam-based study
Purpose: To evaluate anterior and posterior corneal surfaces and anterior segment parameters in the Turkish population using Pentacam, and to analyze age-related changes. Methods: This retrospective study included healthy participants aged 3–86 years who underwent Pentacam imaging at Dokuz Eylül University Faculty of Medicine. Participant-level descriptive summaries were based on right-eye measurements, whereas inferential analyses incorporated both eyes using linear mixed-effects models with a participant-level random intercept. Age was evaluated both categorically (3–19, 20–39, 40–59, and ≥ 60 years) and continuously using centered linear and quadratic age terms. Results: A total of 1202 eyes of 601 participants (49.1% male, 50.9% female; mean age 35.25 ± 23.02 years) were included. Age-group effects were supported for anterior K2, Kmean, Rflat, Rsteep, Rmean, anterior astigmatism magnitude and asphericity, posterior K1, posterior astigmatism magnitude and asphericity, thinnest corneal point, corneal volume, anterior chamber volume, and anterior chamber depth (nominal p < 0.05). Age effects for anterior K1, posterior K2, posterior Rsteep, and anterior and posterior steep astigmatism axes were laterality-dependent. Continuous-age models demonstrated quadratic trajectories for anterior K1, K2, Kmean, Rflat, Rsteep, Rmean, and anterior asphericity, with approximate turning points between 32 and 46 years. Thinnest corneal point, corneal volume, anterior chamber volume, and anterior chamber depth decreased with increasing age; the age slopes for anterior chamber volume and depth differed by sex. Conclusions: Pentacam-derived corneal and anterior segment parameters show non-uniform age-related variation. The most consistent changes involved anterior corneal curvature and shape, corneal and anterior chamber dimensions, whereas several posterior corneal and astigmatic-axis measures were laterality-dependent. These findings support age-aware interpretation of corneal tomography rather than a single uniform aging pattern.
- Open Access
- 6 August 2026
- Published by Springer Nature in BMC Ophthalmology
- Vol. 26 (1)
Role of surgeon seniority in predicting surgically induced astigmatism after phacoemulsification surgery: a machine learning study
Purpose: To evaluate whether surgeon seniority influences postoperative astigmatic change after phacoemulsification cataract surgery and to assess the performance of machine-learning models in predicting postoperative astigmatic outcomes using preoperative clinical data. Methods: This study included 169 eyes of 169 patients who underwent phacoemulsification cataract surgery. Surgeons were categorized as residents ( n = 101) or specialists ( n = 68). Astigmatism was quantified using both classical magnitude-based analysis and vector decomposition. Potential determinants of postoperative astigmatic change were evaluated using multivariable regression analyses. Machine-learning models, including Gradient Boosting, K-Nearest Neighbors, Decision Tree, and Support Vector Machine models, were used to predict classical and vector astigmatism outcomes. Model performance was assessed using mean absolute error, root mean square error, and the coefficient of determination. Results: Of the 169 eyes, 101 were operated on by residents and 68 by specialists. Postoperative astigmatic change did not differ significantly between resident- and specialist-performed procedures in either classical or vector analyses. Within the resident group, increasing surgical seniority was not associated with postoperative astigmatic change. In multivariable analysis, preoperative cylindrical error was the only significant predictor in the vector model, whereas no significant predictor was identified in the classical model. Overall, machine-learning models demonstrated limited predictive performance, with only modest utility for classical astigmatism and poor explanatory performance for vector outcomes. Conclusion: Surgeon seniority was not a significant determinant of surgically induced astigmatism after phacoemulsification cataract surgery. Machine-learning models based on preoperative clinical data provided only limited predictive value, particularly for vector astigmatism outcomes. Incorporating intraoperative variables may improve future predictive performance.
- 6 August 2026
- Published by Springer Nature in BMC Ophthalmology
- Vol. 26 (1)
Open AccessRole of surgeon seniority in predicting surgically induced astigmatism after phacoemulsification surgery: a machine learning study
Purpose: To evaluate whether surgeon seniority influences postoperative astigmatic change after phacoemulsification cataract surgery and to assess the performance of machine-learning models in predicting postoperative astigmatic outcomes using preoperative clinical data. Methods: This study included 169 eyes of 169 patients who underwent phacoemulsification cataract surgery. Surgeons were categorized as residents ( n = 101) or specialists ( n = 68). Astigmatism was quantified using both classical magnitude-based analysis and vector decomposition. Potential determinants of postoperative astigmatic change were evaluated using multivariable regression analyses. Machine-learning models, including Gradient Boosting, K-Nearest Neighbors, Decision Tree, and Support Vector Machine models, were used to predict classical and vector astigmatism outcomes. Model performance was assessed using mean absolute error, root mean square error, and the coefficient of determination. Results: Of the 169 eyes, 101 were operated on by residents and 68 by specialists. Postoperative astigmatic change did not differ significantly between resident- and specialist-performed procedures in either classical or vector analyses. Within the resident group, increasing surgical seniority was not associated with postoperative astigmatic change. In multivariable analysis, preoperative cylindrical error was the only significant predictor in the vector model, whereas no significant predictor was identified in the classical model. Overall, machine-learning models demonstrated limited predictive performance, with only modest utility for classical astigmatism and poor explanatory performance for vector outcomes. Conclusion: Surgeon seniority was not a significant determinant of surgically induced astigmatism after phacoemulsification cataract surgery. Machine-learning models based on preoperative clinical data provided only limited predictive value, particularly for vector astigmatism outcomes. Incorporating intraoperative variables may improve future predictive performance.
- Open Access
- 21 March 2026
- Published by Elsevier in AJO International
- Vol. 3 (2)
Evaluation of automated machine learning platforms against an expert-designed deep learning model for optic disc abnormality detection
- 21 March 2026
- Published by Elsevier in AJO International
- Vol. 3 (2)
Open AccessEvaluation of automated machine learning platforms against an expert-designed deep learning model for optic disc abnormality detection
- Open Access
- 1 July 2025
- Published by Springer Nature in BMC Ophthalmology
- Vol. 25 (1) , 1-8
Exploring computerized tomography findings and ocular trauma score in open globe injuries: assessing imaging’s predictive value
To investigate the relationship between computerized tomography (CT) findings and the Ocular Trauma Score (OTS) in patients with open globe injury (OGI). This study included 123 individuals with OGI, examining injury characteristics, clinical symptoms, and CT scans for scleral irregularity, reduced globe volume, dislocated crystalline lens, aberrant vitreous density, chorioretinal thickening, and foreign body/air presence. The prognostic value of CT imaging in OGI was evaluated by assessing its relationship with OTS findings. The most common OTS category was stage 3, found in 71 eyes (57.7%). The prevalent CT findings included severe scleral irregularity or globe collapse (34.1%), dislocated crystalline lens (33.3%), and abnormal vitreous density (26.8%). Multivariate analysis revealed that severe scleral irregularity or globe collapse (Odds Ratio [OR] = 1.36, p = 0.006) and abnormal vitreous density (OR = 1.7, p < 0.001) were significantly associated with more advanced OTS stages (I or II), indicating more severe ocular trauma. The presence of three CT findings was associated with a higher likelihood of falling into OTS stages I or II (OR = 6.82, p = 0.001). Conversely, the absence of CT findings was associated with a decreased probability of having more severe OTS stages (OR = 0.05, p < 0.001). The study concludes that imaging findings are valuable for correlating with the severity of open globe injuries, highlighting the importance of CT in assessing ocular trauma.
- 1 July 2025
- Published by Springer Nature in BMC Ophthalmology
- Vol. 25 (1) , 1-8
Open AccessExploring computerized tomography findings and ocular trauma score in open globe injuries: assessing imaging’s predictive value
To investigate the relationship between computerized tomography (CT) findings and the Ocular Trauma Score (OTS) in patients with open globe injury (OGI). This study included 123 individuals with OGI, examining injury characteristics, clinical symptoms, and CT scans for scleral irregularity, reduced globe volume, dislocated crystalline lens, aberrant vitreous density, chorioretinal thickening, and foreign body/air presence. The prognostic value of CT imaging in OGI was evaluated by assessing its relationship with OTS findings. The most common OTS category was stage 3, found in 71 eyes (57.7%). The prevalent CT findings included severe scleral irregularity or globe collapse (34.1%), dislocated crystalline lens (33.3%), and abnormal vitreous density (26.8%). Multivariate analysis revealed that severe scleral irregularity or globe collapse (Odds Ratio [OR] = 1.36, p = 0.006) and abnormal vitreous density (OR = 1.7, p < 0.001) were significantly associated with more advanced OTS stages (I or II), indicating more severe ocular trauma. The presence of three CT findings was associated with a higher likelihood of falling into OTS stages I or II (OR = 6.82, p = 0.001). Conversely, the absence of CT findings was associated with a decreased probability of having more severe OTS stages (OR = 0.05, p < 0.001). The study concludes that imaging findings are valuable for correlating with the severity of open globe injuries, highlighting the importance of CT in assessing ocular trauma.
- Open Access
- 25 June 2025
- Published by Galenos Yayinevi in Turkish Journal of Ophthalmology
- Vol. 55 (3) , 120-126
Comparative Analysis of Automated vs. Expert-Designed Machine Learning Models in Age-Related Macular Degeneration Detection and Classification
Comparative Analysis of Automated vs. Expert-Designed Machine Learning Models in Age-Related Macular Degeneration Detection and Classification - Turkish Journal of Ophthalmology
- 25 June 2025
- Published by Galenos Yayinevi in Turkish Journal of Ophthalmology
- Vol. 55 (3) , 120-126
Open AccessComparative Analysis of Automated vs. Expert-Designed Machine Learning Models in Age-Related Macular Degeneration Detection and Classification
Comparative Analysis of Automated vs. Expert-Designed Machine Learning Models in Age-Related Macular Degeneration Detection and Classification - Turkish Journal of Ophthalmology
- 29 January 2025
- Published by Springer Nature in International Ophthalmology
- Vol. 45 (1) , 1-11
Orbital inflammatory disease: a joint clinical experience in rheumatology and ophthalmology
Purpose This retrospective study aimed to characterize the clinical features, histopathological findings, and treatment outcomes of patients diagnosed with orbital inflammatory disease (OID) co-managed by the rheumatology and ophthalmology departments in a tertiary hospital. Methods Medical records of 14 patients with OID were analyzed. Data on demographics, clinical presentation, laboratory investigations, radiological imaging, histopathological results, treatment regimens, and disease outcomes were collected and reviewed. Results The mean age of the patients was 34.5 years, with a female (71.4%) predominance. Periorbital pain (62.5%), periorbital edema (50%) and diplopia (50%) were the most common presenting symptoms. Histopathological evaluation revealed diverse subtypes, including non-specific orbital inflammation (NSOI) (42.8%), orbital myositis (28.5%) and IgG4-related disease (14.3%). Initial treatment with systemic corticosteroids achieved remission in 14.3% of patients, while 85.7% required additional immunosuppressive therapy. Radiotherapy was effective in one patient unresponsive to steroid and azathioprine treatment. Conclusion OID presents with diverse clinical manifestations and histopathological subtypes, often requiring multidisciplinary management. While corticosteroids remain the first-line therapy, adjunctive immunosuppressive agents or radiotherapy may be necessary in refractory cases.
- 29 January 2025
- Published by Springer Nature in International Ophthalmology
- Vol. 45 (1) , 1-11
Orbital inflammatory disease: a joint clinical experience in rheumatology and ophthalmology
Purpose This retrospective study aimed to characterize the clinical features, histopathological findings, and treatment outcomes of patients diagnosed with orbital inflammatory disease (OID) co-managed by the rheumatology and ophthalmology departments in a tertiary hospital. Methods Medical records of 14 patients with OID were analyzed. Data on demographics, clinical presentation, laboratory investigations, radiological imaging, histopathological results, treatment regimens, and disease outcomes were collected and reviewed. Results The mean age of the patients was 34.5 years, with a female (71.4%) predominance. Periorbital pain (62.5%), periorbital edema (50%) and diplopia (50%) were the most common presenting symptoms. Histopathological evaluation revealed diverse subtypes, including non-specific orbital inflammation (NSOI) (42.8%), orbital myositis (28.5%) and IgG4-related disease (14.3%). Initial treatment with systemic corticosteroids achieved remission in 14.3% of patients, while 85.7% required additional immunosuppressive therapy. Radiotherapy was effective in one patient unresponsive to steroid and azathioprine treatment. Conclusion OID presents with diverse clinical manifestations and histopathological subtypes, often requiring multidisciplinary management. While corticosteroids remain the first-line therapy, adjunctive immunosuppressive agents or radiotherapy may be necessary in refractory cases.
- Open Access
- 30 March 2021
- Published by Springer Nature in Eye and Vision
- Vol. 8 (1) , 1-9
The number of intracorneal ring segments in asymmetric and central cones
Background: To compare the results of single versus double intracorneal ring segment (ICRS) (KeraRing) implantation in keratoconus with respect to different cone locations.Methods: Twenty-two eyes of 18 patients with totally asymmetric cones (20–80% or 0–100% distribution along steep axis) were implanted with single ICRS (Group 1), 38 eyes of 32 patients with central or partially asymmetric cones (50–50% or 40–60% distribution along steep axis) were implanted with double ICRS (Group 2), at a depth of 80% of the site of implantation, in channels created with femtosecond laser device. All patients had uncorrected and corrected distance visual acuities (UDVA and CDVA, respectively) of ≤ 0.3 Snellen lines.Results: In both groups, patients had median UDVA and CDVA gain of 3 Snellen lines (P > 0.05). Postoperative improvement in indices of vertical asymmetry and height decentration in Group 1; simulated keratometry, corneal astigmatism and anterior corneal asphericity values in Group 2 were greater (P < 0.05). A total of 10 eyes (45.5%) in Group 1 were recommended double ring implantation by the manufacturer’s nomogram, but underwent single ICRS implantation and achieved visual, refractive, tomographic outcomes comparable to that in Group 2, although corneal cylindrical correction was less and final topographic astigmatism was greater.Conclusion: Double ICRS implantation seems to be superior in terms of keratometry, corneal astigmatism and anterior corneal asphericity improvement. Single ICRS implantation in totally asymmetric cones seems to provide satisfactory visual, refractive and tomographic results, similar to double ICRS implantation in central and partially asymmetric cones, by inducing central shift of the cone.
- 30 March 2021
- Published by Springer Nature in Eye and Vision
- Vol. 8 (1) , 1-9
Open AccessThe number of intracorneal ring segments in asymmetric and central cones
Background: To compare the results of single versus double intracorneal ring segment (ICRS) (KeraRing) implantation in keratoconus with respect to different cone locations.Methods: Twenty-two eyes of 18 patients with totally asymmetric cones (20–80% or 0–100% distribution along steep axis) were implanted with single ICRS (Group 1), 38 eyes of 32 patients with central or partially asymmetric cones (50–50% or 40–60% distribution along steep axis) were implanted with double ICRS (Group 2), at a depth of 80% of the site of implantation, in channels created with femtosecond laser device. All patients had uncorrected and corrected distance visual acuities (UDVA and CDVA, respectively) of ≤ 0.3 Snellen lines.Results: In both groups, patients had median UDVA and CDVA gain of 3 Snellen lines (P > 0.05). Postoperative improvement in indices of vertical asymmetry and height decentration in Group 1; simulated keratometry, corneal astigmatism and anterior corneal asphericity values in Group 2 were greater (P < 0.05). A total of 10 eyes (45.5%) in Group 1 were recommended double ring implantation by the manufacturer’s nomogram, but underwent single ICRS implantation and achieved visual, refractive, tomographic outcomes comparable to that in Group 2, although corneal cylindrical correction was less and final topographic astigmatism was greater.Conclusion: Double ICRS implantation seems to be superior in terms of keratometry, corneal astigmatism and anterior corneal asphericity improvement. Single ICRS implantation in totally asymmetric cones seems to provide satisfactory visual, refractive and tomographic results, similar to double ICRS implantation in central and partially asymmetric cones, by inducing central shift of the cone.
- Open Access
- 1 January 2021
- Published by Kare Publishing in European Eye Research
- Vol. 1 (2) , 107-112
Use of amniotic membrane at outpatient conditions for acute ocular surface involvement of Stevens Johnson Syndrome
- 1 January 2021
- Published by Kare Publishing in European Eye Research
- Vol. 1 (2) , 107-112
Open AccessUse of amniotic membrane at outpatient conditions for acute ocular surface involvement of Stevens Johnson Syndrome