We would like to thank Purohit et al.’s for their novel study demonstrating the feasibility of portable optical coherence tomography (OCT) system in children with craniosynostosis [1]. OCT was previously considered a very non-portable imaging modality, making it extremely difficult for pediatric and immobile patients. However with the usage of the portable OCT (Spectralis Flex, Heidelberg, Germany), unilateral OCT imaging was successful in 88% of children with craniosynostosis and bilateral imaging was successful in 78% [1]. OCT imaging was used in this study as a non-invasive method to assess intracranial hypertension, which is a promising method of intracranial pressure (ICP) assessment. Prior research has shown that optic nerve head central thickness assessed on OCT, correlated with ICP (left eye R = 0.73; P = 0.002; right eye r = 0.60; P = 0.02) [2]. However, further research is required to improve the accuracy of this non-invasive method of ICP assessment for widespread clinical use.

Novel deep learning techniques may be one solution to overcome longstanding limitations in OCT imaging. In higher-grade papilledema, OCT segmentation layers have been shown to frequently fail [3]. Edema and the elevation of the optic nerve in papilledema has also been shown to affect the identification of anatomical landmarks in OCTs, which can affect measurements [4]. A possible solution to overcome these limitations of OCTs is with generative adversarial networks (GANs). GANs are a recently developed deep learning method, that has rapidly revolutionized our ability to address limitations in imaging. GANs have previously been used to significantly improve the performance of OCTs by: improving image quality, removing artifacts and speckle noise, removing shadows from blood vessels [5]. By applying GANs to portable imaging technology, we can continue to increase the accuracy of accessible imaging for clinical use. Our group is currently developing GANs to improve the performance of ophthalmic imaging used to assess astronauts during long-duration spaceflight. Improving the diagnostic performance of OCTs, an imaging modality onboard the International Space Station, remains a high priority. We believe this deep learning technique can be applied with novel portable imaging modalities such as portable OCT for terrestrial ophthalmic diseases. We applaud the authors for this novel study.

Log in or create a free account to read this content

Gain free access to this article, as well as selected content from this journal and more on nature.com

Access through your institution

or

Sign in or create an account
Continue with Google
Continue with ORCiD