Automated urinary sediment detection for Fabry disease using deep-learning algorithms

Hidetaka Uryu; Ohsuke Migita; Minami Ozawa; Chikako Kamijo; Saki Aoto; Kohji Okamura; Fuyuki Hasegawa; Torayuki Okuyama; Motomichi Kosuga; Kenichiro Hata

doi:10.1016/j.ymgmr.2022.100921

Automated urinary sediment detection for Fabry disease using deep-learning algorithms

Mol Genet Metab Rep. 2022 Sep 28:33:100921. doi: 10.1016/j.ymgmr.2022.100921. eCollection 2022 Dec.

Authors

Hidetaka Uryu^{1

2}, Ohsuke Migita^{3

4

5}, Minami Ozawa⁵, Chikako Kamijo³, Saki Aoto¹, Kohji Okamura⁶, Fuyuki Hasegawa^{7

8}, Torayuki Okuyama⁹, Motomichi Kosuga^{10

11

9}, Kenichiro Hata^{4

12}

Affiliations

¹ Medical Genome Center, National Research Institute for Child Health and Development, Tokyo 157-8535, Japan.
² Department of Leukemia, University of Texas MD Anderson Cancer Center, Houston, TX 77030-4009, USA.
³ Department of Laboratory Medicine, St. Marianna University School of Medicine, Kanagawa 216-8511, Japan.
⁴ Department of Maternal-Fetal Biology, National Research Institute for Child Health and Development, Tokyo 157-8535, Japan.
⁵ Department of Pediatrics, St. Marianna University School of Medicine, Kanagawa 216-8511, Japan.
⁶ Department of Systems BioMedicine, National Research Institute for Child Health and Development, Tokyo 157-8535, Japan.
⁷ Biobank, National Center for Child Health and Development Research Center, Tokyo 157-8535, Japan.
⁸ Center for Maternal-Fetal, Neonatal and Reproductive Medicine, National Center for Child Health and Development, Tokyo 157-8535, Japan.
⁹ Center for Lysosomal Storage Diseases, National Center for Child Health and Development, Tokyo 157-8535, Japan.
¹⁰ Division of Medical Genetics, National Center for Child Health and Development, Tokyo 157-8535, Japan.
¹¹ Department of Clinical Laboratory Medicine, National Center for Child Health and Development, Tokyo 157-8535, Japan.
¹² Department of Molecuar and Medical Genetics, Gunma University Graduate School of Medicine, Gunma 371-8511, Japan.

Abstract

Fabry disease is a congenital lysosomal storage disease, and most of these cases develop organ damage in middle age. There are some promising therapeutic options for this disorder, which can stabilize the progression of the disease. However, a long delay in diagnosis prevents early intervention, resulting in treatment failure. Because Fabry disease is a rare disease, it is not well recognized and disease specific screening tests are rarely performed. Hence, a novel approach to for detecting patients with a widely practiced clinical test is crucial for the early detection of the disease. Recently, decision support systems based on artificial intelligence (AI) have been developed in many clinical fields. However, the construction of these models requires datasets from a large number of samples; this aspect is one of the main obstacles in AI-based approaches for rare diseases. In this study, with a novel image amplification method to construct the dataset for AI-model training, we built the deep neural-network model to detect Fabry cases from their urine samples. Sensitivity, specificity, and the AUC of the models on validation dataset were 0.902 (95% CI, 0.900-0.903), 0.977 (0.950-0.980), and 0.968 (0.964-0.972), respectively. This model could also extract disease-specific findings that are interpretable with human recognition. These results indicate that we can apply novel AI models for rare diseases based on this image amplification method we developed. We expect this approach could contribute to the diagnosis of Fabry disease.

Synopsis: This is the first reported AI-based decision support system to detect undiagnosed Fabry cases, and our new image amplification method will contribute to the AI models for other rare disorders.

Keywords: AI, artificial intelligence; AUC, area under the curve; AdHE, adaptive histogram equalization; Artificial intelligence; CNN, convolutional neural network; CntStr, contrast stretching; Deep learning; ERT, enzyme replacement therapy; Fabry disease; Image augmentation; InceptResNet, InceptionResNetV2; Mulberry cells; OrdHE, ordinary histogram equalization; ROC, receiver operating characteristic; Xcep, Xception; alpha-Gal A, α- galactosidase A.