Objective: This study determined the incidence and density of Demodex species on the eyelashes of subjects with normal eyelids, anterior blepharitis (AB), meibomian-gland dysfunction (MGD), and mixed blepharitis (MB). Methods: Consecutive patients diagnosed with AB, MGD, and MB were recruited, along with 50 normal individuals who served as controls. All underwent a standard eye exam. Data on demographics and ocular symptomatology were gathered. Digital photographs of the lid margins were taken. Lash sampling was done using the modified Coston procedure for epilating lashes. The collected lashes were checked for Demodex based on morphology and motility, and the total number of Demodex identified were tabulated for each eye. Patients with AB and MB had eyelid bacterial culture. Data were subjected to statistical analyses. Results: A total of 167 patients with mean age of 53.7 years were included: 40% with MGD, 12% with AB, 19% with MB, and 30% normal. The overall incidence of Demodex infestation was 73%; 85% for MGD, 95% for AB, 97% for MB, and 34% for control group. The mean density count (MDC) was 13.63 for MB, 8.95 for AB, 4.08 for MGD, and 0.98 for the normal group. The incidence and density of Demodex infestation increased with age. Symptoms of irritation correlated with high MDC (>9 mites) (p <0.04). Asymptomatic patients had low MDC (0 to 2 mites) (p < 0.01). Sixty-eight percent of patients with cylindrical dandruff (p <0.001) and 36.1% with plugged meibomian glands (p < 0.001) had high MDC, while 82.7% with no eyelash findings had low MDC (p < 0.001). Conclusion The incidence and density of Demodex infestation was highest among patients with AB and MGD. Lid irritation and presence of cylindrical dandruff were indicative of high density count. These should alert the clinician to treat concomitant Demodex infestation of the eyelashes.
Meibomian Gland Dysfunction

Objective: This study compared the effectiveness of prophylactic administration of topical ketorolac tromethamine 0.5% and nepafenac 0.1% on maintaining mydriasis during phacomulsification. Methods: This is a prospective, randomized, double-masked comparative study involving adult cataract patients given topical NSAIDs (ketorolac or nepafenac) or balanced salt solution (control) prior to phacoemulsification and capsular bag intraocular-lens (IOL) implantation at a tertiary hospital. Horizontal and vertical diameters of pupil were measured at different stages of cataract surgery and the mean values were compared across the three groups. Results: A total of 47 eyes of 44 cataract surgery patients, 13 males and 34 females, with a mean age of 66.04 ± 8.87 years, were included in the study. The mean horizontal and vertical diameters of the three groups were similar at the start of surgery. Significant differences were seen after IOL implantation, with the nepafenac group having the largest mean diameters in both horizontal (p = 0.012) and vertical (p = 0.012) pupil measurements. Conclusion Topical nepafenac has been shown to be a more effective inhibitor of miosis during phacoemulsification and provides a more stable mydriatic effect throughout the surgical procedure compared to topical ketorolac and placebo.
Cataract ; Phacoemulsification ; Mydriasis ; Ketorolac

OBJECTIVE

This study determined the diagnostic accuracies of Growth and Retinopathy of Prematurity (GROP) criteria and a novel modified G-ROP criteria on identifying retinopathy of prematurity (ROP) in infants referred for screening at a tertiary hospital.

METHODS

This was a single-center, cross-sectional, retrospective study. Medical records of infants referred for ROP screening from January 2012 to December 2021 were reviewed. Infants were labelled as “requiring ROP examination” if they met the 2020 Philippine Academy of Ophthalmology – ROP Working Group (PAO-ROPWG) screening consensus, G-ROP, or modified G-ROP criteria. We compared the accuracy of each criterion in predicting prethreshold ROP, evaluating sensitivity, specificity, and predictive values, as well as percentage of low-risk infants. Statistical analysis used Chi-square tests and one-way ANOVA with post hoc testing.

RESULTS

Of the 873 infants, 162 infants (18.6%) were noted to have ROP of any stage. Type 1 ROP developed in 15.4%, and type 2 ROP in 16.7%. The 2020 PAO-ROPWG consensus had 100% sensitivity (95% CI: 86.3%- 100%) in detecting type 1 and 2 ROP while 323 infants (37%) were low-risk. G-ROP criteria had 100% (95% CI: 86.3%-100%) sensitivity and 79.2% (95% CI: 76.4%-81.9%) specificity in predicting type 1 ROP, and 88.89% (95% CI: 70.84%-97.65%) sensitivity and 79.1% (95% CI: 76.2%-81.8%) specificity in predicting type 2 ROP, while 672 infants (77%) were classified as low-risk. Modified G-ROP criteria had a 100% (95% CI: 86.3%-100%) sensitivity in predicting type 1 and 2 ROP, 54.9% (95% CI: 51.5%-58.3%) and 55.1% (95% CI: 51.7%-58.5%) specificity in predicting type 1 and type 2 ROP, respectively, while 472 infants (54%) were classified as low-risk.

CONCLUSION

G-ROP and modified G-ROP criteria showed high sensitivity and better specificity compared to the 2020 PAO-ROPWG consensus. Their stricter criteria for gestational age and birth weight likely enhanced specificity. Further research is needed to confirm these findings in a broader population.

Screening ; Mass Screening ; Retinopathy Of Prematurity

OBJECTIVE

To identify the current practices of ophthalmology subspecialists involved in the treatment of retinopathy of prematurity (ROP) in the Philippines.

METHODS

A survey was conducted among members of the Vitreo-Retina Society of the Philippines (VRSP) and the Philippine Society of Pediatric Ophthalmology and Strabismus (PSPOS) who treat ROP. The electronic questionnaire covered treatment preferences and factors influencing these preferences regarding the use of anti-vascular endothelial growth factor (anti-VEGF), laser indirect ophthalmoscopy (LIO), cryotherapy, and surgical interventions. Questions about referral system and ROP recurrences encountered by the respondents were included.

RESULTS

A total of 73 out of 86 (85.00% response rate) possible respondents were included in the study, majority (80.82%) of whom were retina subspecialists. The initial treatment preference for type 1 ROP was laser indirect ophthalmoscopy (43.84%), while anti-VEGF injection (67.12%) was preferred for aggressive posterior ROP (APROP). Among the available anti-VEGF agents, bevacizumab was the most favored (68.11%). Most ROP consultations (59.79%) occurred through referrals, primarily from pediatricians or neonatologists (91.78%). Subspecialists reported encountering more recurrences with anti-VEGF injections compared to LIO. LIO was the most preferred treatment after failed initial anti-VEGF therapy, with most recurrences occurring in patients older than 50 weeks post-conceptual age.

CONCLUSION

This study underscored the diverse treatment practices for ROP in the Philippines, influenced by factors such as variable access to equipment or drugs, clinician experience with specific treatments, and ocular factors such as faster regression of neovascularization. LIO was preferred as the initial therapy for Type 1 ROP and for managing recurrences, while anti-VEGF injections were favored for APROP and as an alternative treatment for Type 1 ROP. Given the absence of an established standard for treating recurrent ROP cases, long-term monitoring is necessary to facilitate timely interventions and prevent blindness.

Retinopathy Of Prematurity ; Bevacizumab

OBJECTIVE

This study compared the demographics, clinical profile, treatment, and outcomes of retinoblastoma patients seen at medical institutions in the Philippines between two time periods: 2010 to 2015 and 2016 to 2020.

METHODS

This was a multicenter, analytical, cohort study using review of medical charts and databases of retinoblastoma patients seen in 11 medical institutions from 2010 to 2020.

RESULTS

There were 636 patients (821 eyes) included in this study: 330 patients were seen in 2010 to 2015 while 306 in 2016 to 2020. More cases per annum were seen in the latter timeline. The number of patients with unilateral disease was not significantly different between the two time periods (p=0.51). Age at onset of symptom, age at initial consultation, and delay in consult were also not significantly different between the two time periods (p > 0.05). Patients had significantly different distributions of intraocular grades (p < 0.0001) and systemic staging (p < 0.0001) between the two time periods. Enucleation was the most common surgical treatment performed in both timelines. There was significant difference in the status of patients based on the need for systemic chemotherapy (p < 0.01). There was significant difference in outcome between the two time periods, including the proportions of living and deceased patients.

CONCLUSION

This study compared the most comprehensive data on retinoblastoma patients in the country. There was no improvement in early health seeking behavior based on similar age at initial consult and delay in consult. Enucleation remained the most common treatment mode as opposed to chemotherapy due to similar percentage of patients with unilateral disease, an indication for enucleation rather than chemotherapy.

Human ; Retinoblastoma ; Philippines ; Epidemiology ; Treatment ; Therapeutics

OBJECTIVE

This study determined which of the anti-vascular endothelial growth factors (anti-VEGF) agents is the most cost-effective in treating patients with diabetic macular edema (DME).

METHODS

This study was a cost-effectiveness analysis. A decision-analytic Markov cohort model of the natural history and treatment of DME was developed. Data was obtained from a meta-analysis by Virgili et al. on anti-VEGFs for DME in which intravitreal injections of bevacizumab given monthly, 6-weekly, and 12-weekly; ranibizumab given monthly, bimonthly, and as necessary; aflibercept given monthly, bimonthly, and as necessary; and macular laser therapy were evaluated for efficacy and safety in 4,413 eyes. Costs were obtained from local standard retail price at a tertiary government institution and assumed an out-of-pocket expenditure. The study measured and compared gains in quality-adjusted life years (QALYs) and incremental costeffectiveness ratios (ICERs) for each treatment regimen.

RESULTS

Quarterly bevacizumab, monthly ranibizumab (3.82 QALY), and bimonthly ranibizumab injections were the three most beneficial dosing schedules in terms of clinical effectiveness at 3.81, 3.82, and 3.89 QALY, respectively. However, in terms of cost, bevacizumab was substantially most affordable. Quarterly dosing of bevacizumab provided the best value for money, with an ICER of PhP 9,661.70 per QALY gained.

CONCLUSION

Quarterly intravitreal injections of bevacizumab were identified as the most cost-effective treatment regimen for DME. To be considered cost-effective alternatives, ranibizumab requires an 85% price reduction, while aflibercept needs a price reduction exceeding 95%. We recommend quarterly bevacizumab injections be included in the national insurance coverage package, given their cost-effectiveness and clinical efficacy in the treatment of DME.

Cost-benefit Analysis ; Intravitreal Injections

OBJECTIVE

To determine if dilated smartphone photography with a 3D-printed adaptor for a fundus lens can be an alternative screening tool for diabetic retinopathy.

METHODS

This was a single-center, prospective, comparative study of 102 eyes of patients with diabetes mellitus. DR screening was performed using binocular indirect ophthalmoscopy (BIO) as the reference standard, alongside two imaging modalities: a traditional fundus camera and a smartphone equipped with a 3D-printed adaptor. Coded images were evaluated by three masked retina specialists. Sensitivity, specificity, predictive values, and likelihood ratios were calculated for both imaging modalities.

RESULTS

DR gradings of both imaging modalities were compared to BIO and showed very good agreement (κ 0.94-0.96 95% CI 0.93-0.99). Both have high levels of interobserver reliability (kappa value 0.88-0.92 95% CI 0.84-0.96) and intraobserver reliability (kappa value 0.85-1.00 95% CI 0.64-1.00). Smartphone and Visucam 500 images can detect presence of DR with a sensitivity of 93.9% (95% CI 83.1-98.7) and 91.8% (95% CI 80.4- 97.7); and a specificity of 90.6% (95% CI 79.3-96.9) and 92.5% (95% CI 81.8-97.9), respectively, as well as the presence of vision threatening DR with a sensitivity of 100% (95% CI 91.2-100) and specificity of 85.5% (95% CI 74.2-93.1). Image quality was similar between the two imaging modalities (P=1.00).

CONCLUSION

Smartphone-based imaging with an attached 3D-printed adaptor offers high sensitivity and reliability comparable to a traditional fundus camera, and can be used an alternative for DR screening.

Diabetic Retinopathy

OBJECTIVES

This study evaluated the accuracy of an artificial intelligence (AI) model in identifying retinal lesions, validated its performance on a Filipino population dataset, and evaluated the impact of dataset diversity on AI analysis accuracy.

METHODS

This cross-sectional, analytical, institutional study analyzed standardized macula-centered fundus photos taken with the Zeiss Visucam®. The AI model’s output was compared with manual readings by trained retina specialists.

RESULTS

A total of 215 eyes from 109 patients were included in the study. Human graders identified 109 eyes (50.7%) with retinal abnormalities. The AI model demonstrated an overall accuracy of 73.0% (95% CI 66.6% – 78.8%) in detecting abnormal retinas, with a sensitivity of 54.1% (95% CI 44.3% – 63.7%) and specificity of 92.5% (95% CI 85.7% – 96.7%).

CONCLUSION

The availability and sources of AI training datasets can introduce biases into AI algorithms. In our dataset, racial differences in retinal morphology, such as differences in retinal pigmentation, affected the accuracy of AI image-based analysis. More diverse datasets and external validation on different populations are needed to mitigate these biases.

Human ; Artificial Intelligence ; Deep Learning

OBJECTIVE

To develop and evaluate the diagnostic performance of an automated machine learning (AutoML) model for the detection of referable diabetic retinopathy (refDR) in ultrawide field (UWF) retinal images from local Philippine retinal image datasets.

METHODS

A Google AutoML Vision model was trained using 2000 UWF images with a 50/50 ratio of refDR/non-refDR. Images were labeled according to the Early Treatment Diabetic Retinopathy Study (ETDRS) severity grading. RefDR was defined as moderate nonproliferative DR or worse. The dataset was split with 80% for training, 10% for validation, and 10% for testing. Two sets of published UWF image sets were used for external validation. Sensitivity and specificity were calculated in accordance with United States Food and Drug Administration (US FDA) performance requirements of 0.85 and 0.825, respectively.

RESULTS

The area under the precision-recall curve was 0.998. External validation against two datasets showed a sensitivity/specificity of 0.88/0.83 (95% CI 0.80-0.94/0.74-0.89) and 0.83/0.80 (95% CI 0.74-0.89/0.72-0.86), respectively. Positive and negative predictive values were 0.81/0.89 (95% CI 0.73-0.89/0.82-0.94) and 0.75/0.86 (95% CI 0.66-0.83/0.79-0.91), respectively.

CONCLUSION

The pilot performance of the custom AutoML model constructed using local Philippine data approaches US FDA requirements for the diagnosis of referable DR. The ease of use and intuitiveness of the platform, combined with its performance, support the potential of no-code AI in the detection of refDR.

Artificial Intelligence ; Machine Learning

OBJECTIVE

This study compared the choroidal vascularity indices (CVI) among eyes with neovascular, nonneovascular age-related macular degeneration (AMD) and healthy controls.

METHODS

Spectral-domain optical coherence tomography (OCT) with enhanced depth imaging (EDI) from 52 eyes of 33 subjects were analyzed and designated into 3 groups: control, non-neovascular AMD, and neovascular AMD. Using Image J software, a 1.5 mm subfoveal choroidal area was segmented and binarized to measure total and luminal choroidal areas. The CVI was calculated as the ratio of luminal to total choroidal area. Correlation studies were done to assess relationship of CVI with best-corrected visual acuity (BCVA) and disease severity.

RESULTS

The overall median CVI was 0.66 (IQR = 0.62 – 0.69), with overall median total choroidal area of 696,707.60 (IQR = 530,776.80 – 806,348.00), overall median luminal choroidal area of 442,884.60 (IQR = 351,612.80 – 549,540.30), and an overall median choroidal thickness of 237.10 (IQR = 178.43 – 270.25). The overall median LogMAR BCVA was 0.30 (IQR = 0.10 – 0.54). Statistical comparisons showed no significant differences in the median CVI, median total choroidal area, median luminal choroidal area, and median choroidal thickness among the three groups (p > 0.05). However, median BCVA was significantly different among the groups (χ2=35.98, p=0.001). Specifically, those with non-neovascular AMD and neovascular AMD had significantly worse visual acuity compared to the control group (p < 0.05).

CONCLUSION

The study found that CVI, as measured by binarization of EDI-OCT images, was not significantly different among AMD and control groups. BCVA, however, was significantly affected by AMD. These findings suggest that while CVI may not vary with AMD severity, BCVA remains a crucial diagnostic factor. Further research is needed to explore the relationship between choroidal vascularity and AMD using advanced imaging techniques.

Tomography, Optical Coherence