ABSTRACT
Recent advances in the healthcare industry have led to an abundance of unstructured data, making it challenging to perform tasks such as efficient and accurate information retrieval at scale. Our work offers an all-in-one scalable solution for extracting and exploring complex information from large-scale research documents, which would otherwise be tedious. First, we briefly explain our knowledge synthesis process to extract helpful information from unstructured text data of research documents. Then, on top of the knowledge extracted from the documents, we perform complex information retrieval using three major components- Paragraph Retrieval, Triplet Retrieval from Knowledge Graphs, and Complex Question Answering (QA). These components combine lexical and semantic-based methods to retrieve paragraphs and triplets and perform faceted refinement for filtering these search results. The complexity of biomedical queries and documents necessitates using a QA system capable of handling queries more complex than factoid queries, which we evaluate qualitatively on the COVID-19 Open Research Dataset (CORD-19) to demonstrate the effectiveness and value-add. © 2022 IEEE.
ABSTRACT
Purpose : In recent years, innovations in tele-ophthalmology have shown promise in providing quality ophthalmic care to patients in low-access settings and high-risk environments such as the COVID-19 pandemic. Emergency departments and urgent centers may benefit from tele-ophthalmology applications;the tele-images can be sent to the covering ophthalmologist or, to another ED for evaluation of images prior to transferring the patient. In this study, we aimed to assess the ability of resident physicians to identify features of posterior-pole retinal pathology using teleophthalmology. Methods : Retrospective study on 16 patients (32 eyes;30 with retinal pathology and 2 controls) who presented to a retina clinic at an academic medical center. Automated OCT-B images with 3D topographic maps and fundus photographs of the posterior pole using a Topcon Maestro 3D OCT-1 unit were taken. Images were transmitted remotely to a resident physician who attempted to identify retina pathology using fundus photography and OCT. The same images were consequently evaluated by a retina specialist for grading. We then tested the concordance between diagnoses rendered via tele-OCT by the resident physician and the gold standard clinical examination (performed by the retina specialist) using Cohen's Kappa statistic (κ). Results : An overall average of 79.9% concordance for 69 potential findings was obtained between the retina attending's diagnosis with clinical examination and the resident physician's diagnosis using tele-OCT/fundus images based on Cohen's Kappa statistic (κ). The concordance was lower in eyes with vitreous hemorrhage most likely due to the inferior quality fundus and OCT-B images. The resident exam also identified the presence of any macular pathology in all 30 eyes with macular pathology and correctly identified the controls, indicating 100% sensitivity for identifying abnormal findings using tele OCT/fundus images. Conclusions : This study verifies the utility of resident screening of tele-OCT fundus and OCT-B images to identify retinal pathology. Tele-ophthalmology likely has a useful role in triaging retinal pathology whose outcomes could be affected by timely intervention. Many unnecessary emergency transfers may be avoided if the on-call ophthalmology residents are able to review the fundus and OCT images before hand.