The implementation of clinical-decision support algorithms for medical imaging faces challenges with reliability and interpretability. Here, we establish a diagnostic tool based on a deep-learning framework for the screening of patients with common treatable blinding retinal diseases. Our framework utilizes transfer learning, which trains a neural network with a fraction of the data of conventional approaches. Applying this approach to a dataset of optical coherence tomography images, we demonstrate performance comparable to that of human experts in classifying age-related macular degeneration and diabetic macular edema. We also provide a more transparent and interpretable diagnosis by highlighting the regions recognized by the neural network. We further demonstrate the general applicability of our AI system for diagnosis of pediatric pneumonia using chest X-ray images. This tool may ultimately aid in expediting the diagnosis and referral of these treatable conditions, thereby facilitating earlier treatment, resulting in improved clinical outcomes. VIDEO ABSTRACT.

Probing a wide range of cellular phenotypes in neurodevelopmental disorders using patient-derived neural progenitor cells (NPCs) can be facilitated by 3D assays, as 2D systems cannot entirely recapitulate the arrangement of cells in the brain. Here, we developed a previously unidentified 3D migration and differentiation assay in layered hydrogels to examine how these processes are affected in neurodevelopmental disorders, such as Rett syndrome. Our soft 3D system mimics the brain environment and accelerates maturation of neurons from human induced pluripotent stem cell (iPSC)-derived NPCs, yielding electrophysiologically active neurons within just 3 wk. Using this platform, we revealed a genotype-specific effect of methyl-CpG-binding protein-2 (MeCP2) dysfunction on iPSC-derived neuronal migration and maturation (reduced neurite outgrowth and fewer synapses) in 3D layered hydrogels. Thus, this 3D system expands the range of neural phenotypes that can be studied in vitro to include those influenced by physical and mechanical stimuli or requiring specific arrangements of multiple cell types.3D hydrogels | neuronal migration and maturation | 3D RTT modeling N euronal migration and maturation is a key step in brain development. Defects in this process have been implicated in many disorders, including autism (1) and schizophrenia (2). Thoroughly understanding how neural progenitor cell (NPC) migration is affected in neurodevelopmental disorders requires a means of dissecting the process using cells with genetic alterations matching those in patients. Existing in vitro assays of migration generally involve measurement of cell movement across a scratch or gap or through a membrane toward a chemoattractant in 2D culture systems. Although widely used, such assays may not accurately reveal in vivo differences, as neuronal migration is tightly regulated by physical and chemical cues in the extracellular matrix (ECM) that NPCs encounter as they migrate.In vitro 3D culture systems offer a solution to these limitations (3-7). Compared with 2D culture, a 3D arrangement allows neuronal cells to interact with many more cells (4); this similarity to the in vivo setting has been shown to lengthen viability, enhance survival, and allow formation of longer neurites and more dense networks in primary neurons in uniform matrices or aggregate culture (8, 9). Indeed, 3D culture systems have been used to study nerve regeneration, neuronal and glial development (10-12), and amyloid-β and tau pathology (13). Thus, measuring neuronal migration through a soft 3D matrix would continue this trend toward using 3D systems to study neuronal development and pathology.We sought to develop a 3D assay to examine potential migration and neuronal maturation defects in Rett syndrome (RTT), a genetic neurodevelopmental disorder that affects 1 in 10,000 children in the United States and is caused by mutations in the X-linked methyl-CpG-binding protein-2 (MECP2) gene (14). Studies using induced pluripotent stem cells (iPSCs) from RTT patients in traditiona...

Methylation of the regulatory region of the elongation of very‐long‐chain fatty acids‐like 2 (ELOVL2) gene, an enzyme involved in elongation of long‐chain polyunsaturated fatty acids, is one of the most robust biomarkers of human age, but the critical question of whether ELOVL2 plays a functional role in molecular aging has not been resolved. Here, we report that Elovl2 regulates age‐associated functional and anatomical aging in vivo, focusing on mouse retina, with direct relevance to age‐related eye diseases. We show that an age‐related decrease in Elovl2 expression is associated with increased DNA methylation of its promoter. Reversal of Elovl2 promoter hypermethylation in vivo through intravitreal injection of 5‐Aza‐2’‐deoxycytidine (5‐Aza‐dc) leads to increased Elovl2 expression and rescue of age‐related decline in visual function. Mice carrying a point mutation C234W that disrupts Elovl2‐specific enzymatic activity show electrophysiological characteristics of premature visual decline, as well as early appearance of autofluorescent deposits, well‐established markers of aging in the mouse retina. Finally, we find deposits underneath the retinal pigment epithelium in Elovl2 mutant mice, containing components found in human drusen, a pathologic hallmark of age related macular degeneration. These findings indicate that ELOVL2 activity regulates aging in mouse retina, provide a molecular link between polyunsaturated fatty acids elongation and visual function, and suggest novel therapeutic strategies for the treatment of age‐related eye diseases.