Evaluation of an Artificial Intelligence System for Retinopathy of Prematurity Screening in Nepal and Mongolia.

Purpose: To evaluate the performance of a deep learning (DL) algorithm for retinopathy of prematurity (ROP) screening in Nepal and Mongolia. Design: Retrospective analysis of prospectively collected clinical data. Participants: Clinical information and fundus images were obtained from infants in 2 ROP screening programs in Nepal and Mongolia. Methods: Fundus images were obtained using the Forus 3nethra neo (Forus Health) in Nepal and the RetCam Portable (Natus Medical, Inc.) in Mongolia. The overall severity of ROP was determined from the medical record using the International Classification of ROP (ICROP). The presence of plus disease was determined independently in each image using a reference standard diagnosis. The Imaging and Informatics for ROP (i-ROP) DL algorithm was trained on images from the RetCam to classify plus disease and to assign a vascular severity score (VSS) from 1 through 9. Main Outcome Measures: Area under the receiver operating characteristic curve and area under the precision-recall curve for the presence of plus disease or type 1 ROP and association between VSS and ICROP disease category. Results: The prevalence of type 1 ROP was found to be higher in Mongolia (14.0%) than in Nepal (2.2%; P < 0.001) in these data sets. In Mongolia (RetCam images), the area under the receiver operating characteristic curve for examination-level plus disease detection was 0.968, and the area under the precision-recall curve was 0.823. In Nepal (Forus images), these values were 0.999 and 0.993, respectively. The ROP VSS was associated with ICROP classification in both datasets (P < 0.001). At the population level, the median VSS was found to be higher in Mongolia (2.7; interquartile range [IQR], 1.3-5.4]) as compared with Nepal (1.9; IQR, 1.2-3.4; P < 0.001). Conclusions: These data provide preliminary evidence of the effectiveness of the i-ROP DL algorithm for ROP screening in neonatal populations in Nepal and Mongolia using multiple camera systems and are useful for consideration in future clinical implementation of artificial intelligence-based ROP screening in low- and middle-income countries.

External Validation of a Retinopathy of Prematurity Screening Model Using Artificial Intelligence in 3 Low- and Middle-Income Populations.

Importance: Retinopathy of prematurity (ROP) is a leading cause of preventable blindness that disproportionately affects children born in low- and middle-income countries (LMICs). In-person and telemedical screening examinations can reduce this risk but are challenging to implement in LMICs owing to the multitude of at-risk infants and lack of trained ophthalmologists. Objective: To implement an ROP risk model using retinal images from a single baseline examination to identify infants who will develop treatment-requiring (TR)-ROP in LMIC telemedicine programs. Design, Setting, and Participants: In this diagnostic study conducted from February 1, 2019, to June 30, 2021, retinal fundus images were collected from infants as part of an Indian ROP telemedicine screening program. An artificial intelligence (AI)-derived vascular severity score (VSS) was obtained from images from the first examination after 30 weeks' postmenstrual age. Using 5-fold cross-validation, logistic regression models were trained on 2 variables (gestational age and VSS) for prediction of TR-ROP. The model was externally validated on test data sets from India, Nepal, and Mongolia. Data were analyzed from October 20, 2021, to April 20, 2022. Main Outcomes and Measures: Primary outcome measures included sensitivity, specificity, positive predictive value, and negative predictive value for predictions of future occurrences of TR-ROP; the number of weeks before clinical diagnosis when a prediction was made; and the potential reduction in number of examinations required. Results: A total of 3760 infants (median [IQR] postmenstrual age, 37 [5] weeks; 1950 male infants [51.9%]) were included in the study. The diagnostic model had a sensitivity and specificity, respectively, for each of the data sets as follows: India, 100.0% (95% CI, 87.2%-100.0%) and 63.3% (95% CI, 59.7%-66.8%); Nepal, 100.0% (95% CI, 54.1%-100.0%) and 77.8% (95% CI, 72.9%-82.2%); and Mongolia, 100.0% (95% CI, 93.3%-100.0%) and 45.8% (95% CI, 39.7%-52.1%). With the AI model, infants with TR-ROP were identified a median (IQR) of 2.0 (0-11) weeks before TR-ROP diagnosis in India, 0.5 (0-2.0) weeks before TR-ROP diagnosis in Nepal, and 0 (0-5.0) weeks before TR-ROP diagnosis in Mongolia. If low-risk infants were never screened again, the population could be effectively screened with 45.0% (India, 664/1476), 38.4% (Nepal, 151/393), and 51.3% (Mongolia, 266/519) fewer examinations required. Conclusions and Relevance: Results of this diagnostic study suggest that there were 2 advantages to implementation of this risk model: (1) the number of examinations for low-risk infants could be reduced without missing cases of TR-ROP, and (2) high-risk infants could be identified and closely monitored before development of TR-ROP.

Development of Screening Criteria for Retinopathy of Prematurity in Ulaanbaatar, Mongolia, Using a Web-based Data Management System.

PURPOSE: To describe a process for identifying birth weight (BW) and gestational age (GA) screening guidelines in Mongolia. METHODS: This was a prospective cohort study in a tertiary care hospital in Ulaanbataar, Mongolia, of 193 premature infants with GA of 36 weeks or younger and/or BW of 2,000 g or less) with regression analysis to determine associations between BW and GA and the development of retinopathy of prematurity (ROP). RESULTS: As BW and GA decreased, the relative risk of developing ROP increased. The relative risk of developing any stage of ROP in infants born at 29 weeks or younger was 2.91 (95% CI: 1.55 to 5.44; P < .001] compared to older infants. The relative risk of developing any type of ROP in infants with BW of less than 1,200 g was 2.41 (95% CI: 1.35 to 4.29; P = .003] and developing type 2 or worse ROP was 2.05 (95% CI: 0.99 to 4.25; P = .05). CONCLUSIONS: Infants in Mongolia with heavier BW and older GA who fall outside of current United States screening guidelines of GA of 30 weeks or younger and/or BW of 1,500 g or less developed clinically relevant ROP. [J Pediatr Ophthalmol Strabismus. 2020;57(5):333-339.].

Training of Residents and Fellows in Retinopathy of Prematurity Around the World: An International Web-Based Survey.

PURPOSE: To characterize retinopathy of prematurity (ROP) training practices in international residency and fellowship programs. METHODS: A publicly available online-based platform (http://www.SurveyMonkey.com) was used to develop a 28-question multiple-choice survey that targeted ROP screening and treatment methods. The authors solicited training programs in the Philippines, Thailand, and Taiwan. RESULTS: Programs from three countries participated in the survey, and a total of 95 responses collected from residents, fellows, and attending ophthalmologists were analyzed. A descriptive analysis demonstrated that 45 participants (47%) reported 1% to 33% of ROP screenings were performed under direct supervision of attending ophthalmologists, and 35 (37%) reported the use of formal assessments. The majority of participants (Country A: 87%, Country B: 71%, and Country C: 75%) estimated 1% to 33% of their practice was spent screening for ROP. Notably, 44 participants (46%) reported performing zero laser photocoagulation treatments for ROP during training (Country A: 65%, Country B: 38%, and Country C: 38%). CONCLUSIONS: International ophthalmology trainees perform a limited number of ROP examinations and laser interventions. ROP screenings are often unsupervised and lead to no formal evaluation by an attending ophthalmologist. Limited ROP training among ophthalmologists may lead to misdiagnosis and ultimately mismanagement of a patient. Loss of vision and exposure to unwarranted treatments are among the implications of such errors. The findings highlight the need to improve ROP training in international ophthalmology residency and fellowship programs. [J Pediatr Ophthalmol Strabismus. 2019;56(5):282-287.].

Nonresponse and Recurrence of Retinopathy of Prematurity After Intravitreal Ranibizumab Treatment.

BACKGROUND AND OBJECTIVE: To evaluate the surgical outcome of type 1 retinopathy of prematurity (ROP) after intravitreal ranibizumab (IVR) (Lucentis; Genentech, South San Francisco, CA) treatment. PATIENTS AND METHODS: This was a prospective case series. Premature infants with treatment-requiring ROP who received IVR injections from 2013 to 2015 were included. RESULTS: Twenty-two eyes of 12 children were included in the study. Complete resolution of ROP after a single IVR was noted in 73% of eyes. Retreatment was needed in 27% of eyes due to nonresponse to IVR (18%) or recurrence of ROP (9%). After that, all treated eyes (100%) demonstrated regressed ROP with attached retina. The median visual acuity was 0.3 LogMAR (range: 0 LogMAR to 0.8 LogMAR) with a mean follow-up of 25.2 months ± 6.8 months. CONCLUSIONS: IVR is effective and well-tolerated for patients with treatment-requiring ROP. However, nonresponse to IVR or recurrence of ROP after IVR was noted in 27% of treated eyes and required additional treatment. [Ophthalmic Surg Lasers Imaging Retina. 2016;47:1095-1105.].

Retinoblastoma in Mongolia: Clinical characteristics and survival from 1987 to 2014.

BACKGROUND/PURPOSE: This study aims to describe the clinical characteristics and treatment outcome of retinoblastoma in Mongolian children. METHODS: Data of all children diagnosed with retinoblastoma at the National Center for Maternal and Child Health of Mongolia from 1987 to January 2014 were reviewed retrospectively. The ICRB classification was used. Survival characteristics of the cohort were analyzed. RESULTS: Retinoblastoma was diagnosed in 79 eyes of 64 cases during the study period. Median age of diagnosis was 24.5 ± 15.8 months. There were no differences in sex ratio, and 15 cases (23%) were bilateral. Forty-three (67%) patients were from rural areas. The more frequent clinical presentations were leukocoria in 50 (78%) patients, strabismus in 24 (38%) patients, and glaucoma in 21 (33%) patients. Sixty-one (95%) patients were diagnosed with Classification D or worse when presented to us. Due to late diagnosis in the majority of cases, unilateral and bilateral enucleations were performed in 48 (61%) eyes and 24 (30%) eyes, respectively; exenteration was done in three (4%) eyes. Fifty-two (81%) patients received chemotherapy and 13 (8.3%) patients underwent external beam radiation after enucleation. At the time of last follow-up, 52 (81%) patients were alive, five (8%) patients were dead, and seven (11%) patients had lost to follow-up or unknown vital status. The mean follow-up period was 121.5 months (range, 12-360 months). In five cases with immunohistochemistry analysis in the eye specimen, neuron-specific enolase-, Ki-67 protein-, and B-cell lymphoma 2-positive cells were found in all five (100%) cases and Rb protein was detected in three (60%) cases. CONCLUSION: Retinoblastoma in Mongolia is frequently diagnosed at late stages and has a poor outcome. These data show the importance of early pediatric eye examinations and better treatment of retino-blastoma in children younger than 3 years in Mongolia.