Dataset Information

Deep phenotype unsupervised machine learning revealed the significance of pachychoroid features in etiology and visual prognosis of age-related macular degeneration.

ABSTRACT: Unsupervised machine learning has received increased attention in clinical research because it allows researchers to identify novel and objective viewpoints for diseases with complex clinical characteristics. In this study, we applied a deep phenotyping method to classify Japanese patients with age-related macular degeneration (AMD), the leading cause of blindness in developed countries, showing high phenotypic heterogeneity. By applying unsupervised deep phenotype clustering, patients with AMD were classified into two groups. One of the groups had typical AMD features, whereas the other one showed the pachychoroid-related features that were recently identified as a potentially important factor in AMD pathogenesis. Based on these results, a scoring system for classification was established; a higher score was significantly associated with a rapid improvement in visual acuity after specific treatment. This needs to be validated in other datasets in the future. In conclusion, the current study demonstrates the usefulness of unsupervised classification and provides important knowledge for future AMD studies.

SUBMITTER: Hosoda Y

PROVIDER: S-EPMC7595218 | biostudies-literature | 2020 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep phenotype unsupervised machine learning revealed the significance of pachychoroid features in etiology and visual prognosis of age-related macular degeneration.

Hosoda Yoshikatsu Y Miyake Masahiro M Yamashiro Kenji K Ooto Sotaro S Takahashi Ayako A Oishi Akio A Miyata Manabu M Uji Akihito A Muraoka Yuki Y Tsujikawa Akitaka A

Scientific reports 20201028 1

Unsupervised machine learning has received increased attention in clinical research because it allows researchers to identify novel and objective viewpoints for diseases with complex clinical characteristics. In this study, we applied a deep phenotyping method to classify Japanese patients with age-related macular degeneration (AMD), the leading cause of blindness in developed countries, showing high phenotypic heterogeneity. By applying unsupervised deep phenotype clustering, patients with AMD ...[more]

PMID: 33116208

Similar Datasets

Project description:To determine the clinical properties of pachychoroid neovasculopathy (PNV) that differ from conventional neovascular age-related macular degeneration (nAMD) and suggest that they are different clinical entities. To accomplish this, we reviewed the medical records of 100 consecutive patients diagnosed with nAMD. All of the patients were Japanese, and their mean age was 75.5 years. There were 72 men and 28 women. For the bilateral cases, only the right eye was analyzed. An eye was diagnosed with PNV when a macular neovascularization (MNV) was detected just above the dilated choroidal vessels. The Indocyanine green angiographic (ICGA) and en face optical coherence tomographic (OCT) images were used to assess the vertical symmetry of the medium and large choroidal vessels. The subfoveal choroidal thickness (SCT) was also measured manually in the OCT images. After reclassification, there were 29 (29%) patients with typical nAMD (25 with type 1 MNV, 4 with type 2 MNV), 43 (43%) with PNV, 21 (21%) with polypoidal choroidal vasculopathy, and 7 (7%) with retinal angiomatous proliferation. Of the 43 PNV, 17 (39.5%) had polypoidal lesions and 26 (60.5%) had no polypoidal lesions. The percentage of eyes with vertical asymmetry of the medium and large choroidal vessels was significantly greater in the 35 PNV (81.4%) than in the 16 non-PNV (28.1%; P < 0.01) cases. The mean SCT was significantly thicker in the PNV eyes than in the non-PNV eyes (298 ± 96 μm vs. 228 ± 82 μm; P < 0.01). The response of PNV to anti-vascular endothelial growth factor treatments was better than that of non-PNV eyes [higher dry macula rate after the loading period (90.9% vs. 59.1%), fewer total number of injections (11.0 ± 2.9 vs. 13.4 ± 3.2), and longer treatment intervals for the anti-VEGF therapy (8.4 ± 3.1 vs. 13.4 ± 3.2 weeks) at 2 years (all P < 0.01)]. These differences in the morphology and response to anti-VEGF treatments suggest that PNV is a separate clinical entity to conventional nAMD.

Dataset Information

Deep phenotype unsupervised machine learning revealed the significance of pachychoroid features in etiology and visual prognosis of age-related macular degeneration.

Publications

Deep phenotype unsupervised machine learning revealed the significance of pachychoroid features in etiology and visual prognosis of age-related macular degeneration.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets