Moderator
TBD
Panelists
TBD
Viewing Papers
Expand a paper title to the right to view the paper abstract and authors. Use the video link to jump to that poster in the session.
Presenting Author
Raheem Remtulla, MD
Co-Authors
Siddharth Nath PhD, MD, Robert Koenekoop PhD, FRCSC, MD
Purpose
Surgery for congenital cataracts is inherently challenging given natural growth of pediatric eyes and fluctuating refractive errors; leading to sub-optimal outcomes. Implanting intraocular lenses (IOL) is difficult due to anatomical differences and measurement challenges. Machine learning may offer data-driven guidance in IOL implantation
Methods
Retrospective data from 2001 to 2007 on patients with congenital cataracts who had secondary intraocular lens implantation were analyzed using a single-layer neural network with 10 nodes in MATLAB. Training used scaled conjugate gradient and cross-entropy performance tracking. Input factors included sex, ages at cataract extraction and IOL implantation, age at last follow-up, laterality, follow-up time, opacity type, amblyopia therapy, visual axis obscuration, and post-op complications. The target was visual acuity of at least 20/40. Data were split into training (75%), validation (10%), and testing (15%) sets. Model performance was evaluated using accuracy, sensitivity, and specificity.
Results
A total of 110 patients were collected, of whom 54 patients (49%) had a final visual acuity of at least 20/40. Neural network training was completed in 14 epochs, reaching the maximum cross-entropy value. The sensitivity in the training set was 85%, 80.0% in the validation set, and 88.9% in the test set. The specificity was 85.7% in the training set, 83.3% in the validation set, and 87.5% in the test set. The total accuracy was 85.4% in the training set, 81.8% in the validation set, and 88.2% in the test set. The overall network sensitivity, specificity, and accuracy were 85.2%, 85.7%, and 85.5%, respectively.
Conclusion
A neural network approach demonstrated strong sensitivity and specificity in predicting visual acuity post congenital cataract surgery. Importantly, the network did not demonstrate significant overfitting. A neural network effectively accounted for the age of IOL implantation and could guide clinicians on optimal IOL insertion timing in the future
Presenting Author
Parul Chawla Gupta, MS
Co-Authors
Abhilasha Kumari MBBS, Jaspreet Sukhija MS, Savleen Kaur MS, Surinder Pandav MS, Srishti Raj MS
Purpose
To compare the complications and outcomes of pediatric cataract surgery using early (2 weeks) removal of 10-0 Nylon sutures in one eye versus late (6 weeks) removal of 10-0 Nylon sutures in another eye of patients having bilateral congenital or developmental cataract in a contralateral eye study
Methods
This was a Prospective, Interventional, Assessor Blinded, Randomized Controlled Trial including children (up to 8 years of age) having bilateral congenital or developmental cataracts. All the patients underwent cataract surgery in both eyes, with early (2 weeks) removal of 10-0 Nylon sutures in one eye included in Group A and later (6 weeks) removal of 10-0 Nylon sutures in the other eye included in Group B. Primary outcome was the rate of suture-related complications like corneal vascularization, mucus infiltration, loose sutures, and suture breakage between the 2 groups.Need for sedation for suture-related complications, and the corneal incision architecture were also compared using ASOCT.
Results
In this study of 82 eyes of 41 patients, there was a statistically significant difference in the rate of suture-related complications (24.4% in Group A and 100% in Group B)such as loosening of sutures, mucus infiltration at the suture site, and corneal neovascularization at the suture site (p = <0.001) between="" both="" groups.="" however,="" the="" difference="" in="" the="" rate="" of="" suture="" breakage="" was="" not="" statistically="" significant="" between="" both="" groups="" (p="0.241)." no="" significant="" corneal="" incision="" architectural="" changes="" were="" seen="" asoct="" between="" both="" the="" groups.="" there="" was="" also="" no="" case="" of="" sedation="" required="" for="" the="" management="" of="" suture-related="" complications="" in="" either="">
Conclusion
10-0 Nylon sutures should be used for surgical incision closure after pediatric cataract surgery and should be removed early i.e. 2-3 weeks after surgery due to significantly lower risk of suture-related complications and satisfactory healing of corneal surgical incisions in children.
Presenting Author
Insa M Mannstadt, BSc, BA
Co-Authors
Lisa Park MD, Annette Goulak BA, Timothy Daniel BA
Purpose
Myopia, or nearsightedness, is an increasing global health concern. While extensive research has been conducted in developed nations, data from East Africa, particularly regarding pediatric myopia, is limited. This study aims to examine the prevalence of pediatric myopia and its association with demographic and geographic factors in East Africa.
Methods
This retrospective study analyzed data from annual primary school vision screenings conducted by trained staff in East Africa between 2019 and 2024. A total of 34 primary schools participated, comprising 29 schools in Addis Ababa, Ethiopia, and 5 schools in Pwani, Tanzania. Demographic information was collected for students and schools. The schools were characterized by their level of urbanization, including size, population, and location. Parental consent was obtained, and basic eye examinations were performed on students.
Results
Among the 4204 screened students, the overall prevalence of myopia was 3%, with significant variation across geographic regions. Urban areas exhibited a higher prevalence (5%) compared to rural areas (3%) (p=0.05). There was an upward temporal trend; annual screenings showed an overall increase of 9% across all regions, with a more pronounced rise in urban (11%) compared to rural regions (7%). The 11-13 age group had the highest prevalence, peaking at 15%. Myopia prevalence was significantly associated with urban-rural settings, with children in urban areas having higher odds (aOR: 1.24, 95% CI: 1.15-1.42) after adjusting for age, sex, and country.
Conclusion
This study sheds light on a significant variation in the prevalence of pediatric myopia across East Africa, with a notable urban-rural divide. The upward trend in myopia cases, particularly in urban settings, emphasizes the need for targeted interventions and preventative measures.
Presenting Author
Purendra Bhasin, MS
Purpose
To avoid visual axis opacification, UGH syndrome, and IOL decentration. We report our experience of phaco-aspiration/emulsification (PA/PE) and foldable IOL implantation with primary posterior capsulotomy/capsulorhexis and IOL optic posterior capture with or without automated anterior vitrectomy in paediatric traumatic cataracts.
Methods
This prospective interventional case series involved 82 pediatric traumatic cataract patients, with surgeries performed under general anesthesia by a single surgeon. Corneal/corneoscleral tears were repaired first, and after stabilizing the eye, phaco-aspiration/emulsification (PA/PE) and IOL implantation were done at least 2 weeks post-trauma. A multi-piece hydrophobic acrylic foldable IOL was implanted with posterior optic capture in the primary posterior capsulotomy/capsulorhexis. Patients with retinal detachment, endophthalmitis, no light perception, inadequate capsule support, uncontrolled glaucoma, or less than 3 months of follow-up were excluded.
Results
The mean age of patients was 7.8 years (range 2-15) yrs. The mean follow-up was 14.7 months (range 6-42) yrs. Preoperative BCVA (logMAR) was 2.04 Mean BCVA at the last follow-up was 0.253 (logMAR). Visual axis remained clear in all the patients, with no incidence of posterior capsule opacification and no IOL decentration till the last follow-up.
Conclusion
Posterior optic capture of foldable PCIOL in pediatric traumatic cataract cases, after primary posterior capsulotomy with/without AV, prevents Visual Axis Opacification and ensures well-centered IOL placement. This technique reduces the risk of low-grade uveitis and amblyopia while retaining the benefits of small incision surgery for the child.
Presenting Author
Bryan Sim, MBBS, FRCOphth
Purpose
To assess the efficacy of combination myopia control spectacle lenses (HALT) with atropine eye drops in slowing down myopia in children still progressing whilst on low dose atropine (LDA).
Methods
Children with myopia aged 6-11 years old still progressing on atropine (by >-0.5D in at least 1 eye in past 6 months) were started on Highly Aspherical Lenslet Target (HALT) (Essilor��Stellest�) spectacle lenses. Change in spherical�equivalent (SE), cycloplegic autorefraction (cAR) and axial length (AL), were calculated prior and after treatment.
Results
50 multi-ethnic children (age 8.9�1.1 years) were recruited Gp A:Atropine 0.01% (A0.01) daily (n20) Gp B:A0.01 twice daily (n5); or A0.025 daily (n25). Prior to study, myopia progression was -0.60D/0.24mm over 6m. After addition of HALT, change in SE/AL decreased to -0.06D/0.06mm over first 6m. At 12 months, change in SE/AL was -0.15D/0.14mm. Benefit present in all age groups. At 12m, a hyperopic shift in AL was seen in 11 children (24%). However, progression of >0.5D was noted in 20%, with 18% and 40% progressing by >0.3mm and >0.15mm respectively. SE change was greater in those on higher atropine dose, while change in AL was greater in younger and smaller baseline pupil size.
Conclusion
Children progressing on LDA benefited from addition of HALT lenses. Significant reductions in SE and AL was noted in children as young as 7YO. Study limitations include small sample size, no control group and short study duration. Benefit in first 6m was better than the second 6m and a longer study is required to sustainability of myopia control.
Presenting Author
Michael X. Lin, BA
Co-Authors
Janet Alexander MD
Purpose
The purpose of this study was to quantify the retinal blood flow of pediatric patients with cataracts or amblyopia compared to blood flow of pediatric patients without conditions affecting the retina.
Methods
Patients 18 years old or below with post-operative cataracts or amblyopia were included in this prospective observational study. Patients 18 years old or below without conditions affecting the retina were included as controls. Patients� eyes were imaged using laser-speckle contrast imaging to measure peak blood flow velocity index, dip blood flow velocity index, and the mean blood flow velocity of the optic nerve. Mean blood flow velocities were compared relative to control groups or the contralateral eye of the patient using unpaired student�s t-tests. Age-matched comparisons were further performed.
Results
23 control eyes, 8 cataract eyes, and 8 amblyopia eyes were included. Before age-matching, mean blood flow velocity in cataract patients was decreased compared to controls (11.1 v. 12.4, p=0.325). After age-matching, mean blood flow velocity in cataract patients was decreased compared to age-matched controls (11.1 v. 12.0, p=0.621). Mean blood flow in unilateral cataract patients had slower blood flow in the cataract eye compared to contralateral non-cataract eye (8.0 v. 10.9, p=0.353). Mean blood flow velocity of amblyopia patients was increased compared to age-matched controls (12.8 v. 11.8, p=0.450). Data collection is still ongoing.
Conclusion
A non-significant trend of decreased blood flow velocity among pediatric cataract patients was observed compared to age-matched controls. Additionally, a non-significant trend of increased blood flow velocity among pediatric amblyopia patients patients was observed compared to age-matched controls.
Presenting Author
Larissa Gouvea, MD
Co-Authors
M. Edward Wilson MD, Rupal Trivedi MD, Antonio Lottelli MD, PhD
Purpose
To compare the accuracy of three mathematical models in predicting postoperative axial length (AL) in children with bilateral cataract surgery with primary IOL implantation.
Methods
Eyes were included only if AL data were available before surgery and at least 6 months after surgery. Three mathematical models were created, one logarithmic model without age correction (Log Model 1), one logarithmic model with age correction at 0.6 years (Log Model 2), and a general estimation equation model (GEE). R2 values, Lin agreement coefficient and Bland-Altman analysis were established, with measured and predicted values.
Results
One hundred one eyes were included. Forty-five eyes of 23 children were female, and 56 eyes of 28 children were male. Between one and eight postoperative AL measurements were performed in a total of 401 measurements. Predicting postoperative AL using Log 1 Model = baseline AL + 2.87 x log10 (age at follow-up/baseline age); Log 2 model = Baseline AL + 3.56 x log10 (age at follow-up +0.6/baseline age +0.6); GEE model = -1.89 + 1.09 x baseline AL + 2.65 x Log 10 Age at follow-up -3.17 x log10 baseline age + 0.52 x (log10 initial age x log10 age at follow-up) (Respective values: R2 0.86, 0.74 and 0.87; Lin's coefficient of agreement 0.915, 0.58, 0.93)
Conclusion
The logarithmic model 1 and the GEE model are good predictors of future AL from baseline AL, age at surgery and age at follow-up.
Presenting Author
Sathya T. Ravilla, MS
Co-Authors
Rengaraj Venkatesh MD
Purpose
To evaluate the feasibility, patient satisfaction, and initial outcomes of implementing shared medical appointments (SMA) for pediatric eye conditions that require long term compliance to ensure visual recovery. We aim to determine if this group-based approach can improve patient engagement and knowledge among their guardians.
Methods
A prospective pilot study done on children with anisometropic amblyopia on patching therapy and their families. Questionnaires regarding their knowledge and compliance to patching therapy were filled by the parents� pre and post session after obtaining consent. During the shared appointment, each child was examined individually in the same room; but queries were answered via group counselling and facilitation of discussion among parents. The primary objective was to increase the knowledge about the eye condition among the families, solve their queries, alleviate their fears and most importantly support peer to peer learning.
Results
Of the 26 families with anisometropic amblyopia contacted, 11 attended the shared medical appointments. The attendees were categorized into two age groups: 5 patients aged 7�11 years and 6 patients aged 12�15 years. Feedback collected from all participants indicated high levels of satisfaction with the SMA format. Participants appreciated the group setting for its supportive environment and the opportunity to share experiences and discuss treatment challenges collectively.
Conclusion
Our results support use of SMA as a potential to increase the knowledge and attitude towards anisometric amblyopia. This can improve the compliance to patching, follow up rates and thereby improving success rate of the treatment. SMA can also be tried for other conditions like Strabismic amblyopia and Aphakia.
Presenting Author
Philip Dockery, MD, MPH
Co-Authors
Ayesh Ali BEng, Abdelrahman Elhusseiny MD
Purpose
To review pediatric cataracts and identify primary suspect drugs with possible causal relationships using the FDA Adverse Events Reporting System (FAERS).
Methods
The FAERS reports of congenital cataract, subcapsular cataract, nuclear cataract, and cortical cataract from 2004 to 2024 were observed in patient 18 years of age or younger. Descriptive statistics and a signal detection model screened for positive pharmacovigilance signals, and five data mining algorithms were used to indicated greater-than-expected drug-adverse event reporting frequencies.
Results
Cataracts with potentially associated drugs were identified in 215 patients. Most patients were female (n=81, 38%), followed by male patients (n=76, 35%) and then of unspecified gender (n=58, 27%). 142 drugs were listed as primary suspect drugs. Lamotrigine (n=24, 6%), followed by prednisolone (n=17, 4%) were the most commonly implicated primary suspect drugs. The signal detection analysis indicated positive signals for 23 drugs. Of these, five were from the corticosteroid drug class and five were associated with fetal exposure to anticonvulsants. Zolmitriptan yielded the strongest positive signal (OR = 5507 [2005-15126], p<0.001), followed="" by="" perampanel="" (or="1799" [949-3407],="">< />
Conclusion
Twenty-three potential drug-adverse event pairs were identified for pediatric cataract patients. Novel associations with zolmitriptan and perampanel, alongside previously documented associations with corticosteroids and other anticonvulsants, were observed.
Presenting Author
Ayesh Ali, BEng
Co-Authors
Qais Dihan BSc, Muhammad Chauhan MS, MD, Taher Eleiwa MD, PhD, Andrew Brown OD, Amr Hassan MD, Reem ElSheikh MD, Ahmed Sallam FRCOphth, MD, PhD, Abdelrahman Elhusseiny MD
Purpose
This was a cross-sectional comparative study. We evaluated the ability of three large language models (LLMs) (ChatGPT-3.5, ChatGPT-4, and Google Bard) to generate novel patient education materials (PEMs) and improve the readability of existing PEMs on pediatric cataract.
Methods
We compared LLMs� responses to three prompts. Prompt A requested a handout on pediatric cataract that was �easily understandable by an average American.� Prompt B requested the handout be at a �sixth-grade reading level, using the Simple Measure of Gobbledygook (SMOG) readability formula.� Prompt C rewrote existing PEMs �to a sixth-grade reading level using the SMOG readability formula�. Responses quality (DISCERN; 1-5), understandability and actionability (Patient Education Materials Assessment Tool (?70%: understandable, ?70%: actionable)), accuracy (Likert misinformation; 1-5) and readability (SMOG, Flesch-Kincaid Grade Level (FKGL); grade level <7: highly="" readable)="" were="">
Results
All LLM-generated responses were of high-quality (median DISCERN ?4), understandability (?70%), and accuracy (Likert=1). All LLM-generated responses were not actionable (<70%). chatgpt-3.5="" and="" chatgpt-4="" prompt="" b="" responses="" were="" more="" readable="" than="" prompt="" a="" responses=""><0.001). chatgpt-4="" generated="" more="" readable="" responses="" (lower="" smog="" and="" fkgl="" scores;="" 5.59�0.5?and="" 4.31�0.7,="" respectively)="" than="" the="" other="" two="" llms=""><0.001) and="" consistently="" rewrote="" them="" to="" or="" below="" the="" specified="" sixth-grade="" reading="" level="" (smog:="">
Conclusion
LLMs, particularly ChatGPT-4, proved valuable in generating high-quality, readable, accurate PEMs and in improving the readability of existing materials on pediatric cataract.
Presenting Author
Ayesh Ali, BEng
Co-Authors
Abdelrahman Elhusseiny MD, Paul Phillips MD
Purpose
To identify potential teratogenic medications causing Optic Nerve Hypoplasia (ONH) and/or Septo-Optic Dysplasia (SOD), by screening the FDA Adverse Events Reporting System (FAERS) Database.
Methods
The FAERS database was queried from 2004 to Q2 2024 for ONH and SOD as adverse events. A qualitative assessment considered patient demographics and a disproportionality analysis covered pharmacovigilance signal detection and drug-event reporting frequencies. To ascertain if these reports yielded statistically significant signals, we used the proportional reporting ratio (PRR), Chi-squared with Yates' correction (?2), reporting odds ratio (ROR), empirical Bayes geometric mean (EBGM), and information component (IC).
Results
146 reports were identified. Only the 88 cases reporting prenatal drug exposure were included. 27 reports were males and 62 primary suspect drugs were implicated. Diazepam was the most commonly reported (n=26,15%) followed by methadone hydrochloride (n=22,13%). Positive signals were identified for these drugs: diazepam, methadone hydrochloride, propranolol hydrochloride, ondansetron hydrochloride and nitrazepam. Nitrazepam had the strongest positive signal (n=4, PRR=800, ?2=2435, ROR 95% CI:838 [307-2287], EBGM [EBGM05]:14 [3], IC [IC05]:3.16 [1.42]). Then followed by methadone hydrochloride (n=22, PRR=190, ?2=3945, ROR 95% CI: 253 [156-409], EBGM [EBGM05]: 100 [65], IC [IC05]: 5.19 [4.54]).
Conclusion
Positive signals and hence possible teratogenic causal relationship was found between five medications, 4 of which were novel associations not covered in past literature. Nitrazepam and methadone hydrochloride demonstrated the strongest positive signal. Further investigation is required to confirm this relationship and drug safety profiles.
To log in, click the teal "Login" button in the upper right-hand corner of this page. If you are logged in but still do not have access, please check your 2025 Annual Meeting registration.
