Feasibility assessment of infectious keratitis depicted on slit-lamp and smartphone photographs using deep learning

BACKGROUND This study aims to investigate how infectious keratitis depicted on slit-lamp and smartphone photographs can be reliably assessed using deep learning. MATERIALS AND METHODS We retrospectively collected a dataset consisting of 5,673 slit-lamp photographs and 400 smartphone photographs acquired on different subjects. Based on multiple clinical tests (e.g., cornea scraping), these photographs were diagnosed and classified into four categories, including normal (i.e., no keratitis), bacterial keratitis (BK), fungal keratitis (FK), and herpes simplex virus stromal keratitis (HSK). We preprocessed these slit-lamp images into two separate subgroups: (1) global images and (2) regional images. The cases in each group were randomly split into training, internal validation, and independent testing sets. Then, we implemented a deep learning network based on the InceptionV3 by fine-tuning its architecture and used the developed network to classify these slit-lamp images. Additionally, we investigated the performance of the InceptionV3 model in classifying infectious keratitis depicted on smartphone images. We, in particular, clarified whether the computer model trained on the global images outperformed the one trained on the regional images. The quadratic weighted kappa (QWK) and the receiver operating characteristic (ROC) analysis were used to assess the performance of the developed models. RESULTS Our experiments on the independent testing sets showed that the developed models achieved the QWK of 0.9130 (95% CI: 88.99-93.61%) and 0.8872 (95% CI: 86.13-91.31%), and 0.5379 (95% CI, 48.89-58.69%) for the global images, the regional images, and the smartphone images, respectively. The area under the ROC curves (AUCs) were 0.9588 (95% CI: 94.28-97.48%), 0.9425 (95% CI: 92.35-96.15%), and 0.8529 (95% CI: 81.79-88.79%) for the same test sets, respectively. CONCLUSION The deep learning solution demonstrated very promising performance in assessing infectious keratitis depicted on slit-lamp photographs and the images acquired by smartphones. In particular, the model trained on the global images outperformed that trained on the regional images.

[1]  M. Burton,et al.  Distinguishing fungal and bacterial keratitis on clinical signs , 2015, Community eye health.

[2]  Hao Chen,et al.  Automated segmentation of the optic disc from fundus images using an asymmetric deep learning network , 2021, Pattern Recognit..

[3]  T. Liesegang.  Classification of herpes simplex virus keratitis and anterior uveitis. , 1999, Cornea.

[4]  Jianqiang Li,et al.  Exploiting ensemble learning for automatic cataract detection and grading , 2016, Comput. Methods Programs Biomed..

[5]  L. E. Clemens,et al.  Designed Host Defense Peptides for the Treatment of Bacterial Keratitis , 2017, Investigative ophthalmology & visual science.

[6]  A. Roychoudhury,et al.  Epithelial remodeling as basis for machine-based identification of keratoconus. , 2014, Investigative ophthalmology & visual science.

[7]  Alexandru Lavric,et al.  KeratoDetect: Keratoconus Detection Algorithm Using Convolutional Neural Networks , 2019, Comput. Intell. Neurosci..

[8]  S. Sivaprasad,et al.  Diagnostic Accuracy of Community-Based Diabetic Retinopathy Screening With an Offline Artificial Intelligence System on a Smartphone. , 2019, JAMA ophthalmology.

[9]  Abhishek Das,et al.  Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization , 2016, 2017 IEEE International Conference on Computer Vision (ICCV).

[10]  S. Kulkarni,et al.  Herpes Simplex Virus: The Hostile Guest That Takes Over Your Home , 2020, Frontiers in Microbiology.

[11]  Subhashini Venugopalan,et al.  Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. , 2016, JAMA.

[12]  Aini Hussain,et al.  Automated pterygium detection method of anterior segment photographed images , 2018, Comput. Methods Programs Biomed..

[13]  Xinyi Wu,et al.  Hyphae Detection in Fungal Keratitis Images With Adaptive Robust Binary Pattern , 2018, IEEE Access.

[14]  Lizhen Cui,et al.  Automatic diagnosis of fungal keratitis using data augmentation and image fusion with deep convolutional neural network , 2020, Comput. Methods Programs Biomed..

[15]  Lamprini Papaioannou,et al.  Corneal Collagen Cross-Linking for Infectious Keratitis: A Systematic Review and Meta-Analysis , 2016, Cornea.

[16]  Z. Nagy,et al.  Accuracy of machine learning classifiers using bilateral data from a Scheimpflug camera for identifying eyes with preclinical signs of keratoconus , 2016, Journal of cataract and refractive surgery.

[17]  Saurabh Ghosh,et al.  A 10-year analysis of microbiological profiles of microbial keratitis: the North East England Study , 2018, Eye.

[18]  J. Rose-Nussbaumer,et al.  Update on the Management of Infectious Keratitis. , 2017, Ophthalmology.

[19]  N. Shoji,et al.  Keratoconus detection using deep learning of colour-coded maps with anterior segment optical coherence tomography: a diagnostic accuracy study , 2019, BMJ Open.

[20]  J. Mehta,et al.  The Asia Cornea Society Infectious Keratitis Study: A Prospective Multicenter Study of Infectious Keratitis in Asia. , 2018, American journal of ophthalmology.

[21]  Qiang Zhu,et al.  Deep Sequential Feature Learning in Clinical Image Classification of Infectious Keratitis , 2020, Engineering.

[22]  Michael W. Belin,et al.  Enhanced Tomographic Assessment to Detect Corneal Ectasia Based on Artificial Intelligence. , 2018, American journal of ophthalmology.

[23]  Sina Farsiu,et al.  Novel Image-Based Analysis for Reduction of Clinician-Dependent Variability in Measurement of the Corneal Ulcer Size , 2017, Cornea.

[24]  Carina Koppen,et al.  Computer aided diagnosis for suspect keratoconus detection , 2019, Comput. Biol. Medicine.