EyeScreen: Development and Potential of a Novel Machine Learning Application to Detect Leukocoria.

Purpose: Early diagnosis and treatment of retinoblastoma are of paramount importance for a positive clinical outcome. The most common sign of retinoblastoma is leukocoria, or white pupil. Effective, easy-to-perform, community-based screening is needed to improve outcomes in lower-income regions. The EyeScreen (developed by Joshua Meyer from the University of Michigan) Android (Google LLC) smartphone application is an important step toward addressing this need. The purpose of this study was to examine the potential of the novel use of low-cost technologies-a cell phone application and machine learning-to identify leukocoria. Design: A cell phone application was developed and refined with the feedback from on-site, single-population use in Ethiopia. Application performance was evaluated in this technology validation study. Participants: One thousand four hundred fifty-seven participants were recruited from ophthalmology and pediatric clinics in Addis Ababa, Ethiopia. Methods: Photographs obtained with inexpensive Android smartphones running the EyeScreen Application were used to train an ImageNet (ResNet) machine learning model and to measure the performance of the app. Eighty percent of the images were used in training the model, and 20% were reserved for testing. Main Outcome Measures: Performance of the model was measured in terms of sensitivity, specificity, receiver operating characteristic (ROC) curve, and precision-recall curve. Results: Analyses of the participant images resulted in the following at the participant level: sensitivity, 87%; specificity, 73%; area under the ROC curve, 0.93; and area under the precision-recall curve, 0.77. Conclusions: EyeScreen has the potential to serve as an effective screening tool in the areas of the world most affected by delayed retinoblastoma diagnosis. The relatively high initial performance of the machine learning model with small training datasets in this early-phase study can serve as a proof of concept for future use of machine learning and artificial intelligence in ophthalmic applications.

Telehealth-based Eye Care During the COVID-19 Pandemic: Utilization, Safety, and the Patient Experience.

PURPOSE: To assess the initial utilization, safety, and patient experience with tele-ophthalmology during the COVID-19 pandemic. DESIGN: Cross-sectional study. METHODS: We conducted a telephone survey and interview of a random sample of patients who received different modalities of care (in-person, telephone, videocall, or visits deferred) during Michigan's shelter-in-place order beginning March 23, 2020. The survey assessed patient safety, patient satisfaction with care, perceptions of telehealth-based eye care, and worry about eyesight. Data were analyzed via frequency measures (eg, means and standard deviations), χ2 tests, ANOVA, and paired t tests. Interviews were analyzed using grounded theory. RESULTS: A total of 3,274 patients were called and 1,720 (53%) agreed to participate. In-person participants were significantly older than telephone (P = .002) and videocall visit (P = .001) participants. Significantly more white participants had in-person visits than minority participants (P = .002). In-person visit participants worried about their eyesight more (2.7, standard deviation [SD] = 1.2) than those who had telephone (2.5, SD = 1.3), videocall (2.4, SD = 1.1), or deferred visits (2.4, SD = 1.2) (P = .004). Of all telephone or videocall visits, 1.5% (n = 26) resulted in an in-person visit within 1 day, 2.9% (n = 48) within 2-7 days, and 2.4% (n = 40) within 8-14 days after the virtual visit demonstrating appropriate triage to telemedicine-based care. Patients frequently cited a desire for augmenting the telephone or videocall visits with objective test data. CONCLUSIONS: When appropriately triaged, tele-ophthalmology appears to be a safe way to reduce the volume of in-person visits to promote social distancing in the clinic. A hybrid model of eye care combining ancillary testing with a video or phone visit represents a promising model of care.

Oncogenic Role of THOR, a Conserved Cancer/Testis Long Non-coding RNA.

Large-scale transcriptome sequencing efforts have vastly expanded the catalog of long non-coding RNAs (lncRNAs) with varying evolutionary conservation, lineage expression, and cancer specificity. Here, we functionally characterize a novel ultraconserved lncRNA, THOR (ENSG00000226856), which exhibits expression exclusively in testis and a broad range of human cancers. THOR knockdown and overexpression in multiple cell lines and animal models alters cell or tumor growth supporting an oncogenic role. We discovered a conserved interaction of THOR with IGF2BP1 and show that THOR contributes to the mRNA stabilization activities of IGF2BP1. Notably, transgenic THOR knockout produced fertilization defects in zebrafish and also conferred a resistance to melanoma onset. Likewise, ectopic expression of human THOR in zebrafish accelerated the onset of melanoma. THOR represents a novel class of functionally important cancer/testis lncRNAs whose structure and function have undergone positive evolutionary selection.