scholarly journals An artificial intelligence-based deep learning algorithm for the diagnosis of diabetic neuropathy using corneal confocal microscopy: a development and validation study

Diabetologia ◽  
2019 ◽  
Vol 63 (2) ◽  
pp. 419-430 ◽  
Author(s):  
Bryan M. Williams ◽  
Davide Borroni ◽  
Rongjun Liu ◽  
Yitian Zhao ◽  
Jiong Zhang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Confocal Microscopy ◽  
Diabetic Neuropathy ◽  
Nerve Fibre ◽  
Learning Algorithm ◽  
Fibre Length ◽  
Branch Points ◽  
Deep Learning Algorithm ◽  
Corneal Confocal Microscopy ◽  
Corneal Nerve

Abstract Aims/hypothesis Corneal confocal microscopy is a rapid non-invasive ophthalmic imaging technique that identifies peripheral and central neurodegenerative disease. Quantification of corneal sub-basal nerve plexus morphology, however, requires either time-consuming manual annotation or a less-sensitive automated image analysis approach. We aimed to develop and validate an artificial intelligence-based, deep learning algorithm for the quantification of nerve fibre properties relevant to the diagnosis of diabetic neuropathy and to compare it with a validated automated analysis program, ACCMetrics. Methods Our deep learning algorithm, which employs a convolutional neural network with data augmentation, was developed for the automated quantification of the corneal sub-basal nerve plexus for the diagnosis of diabetic neuropathy. The algorithm was trained using a high-end graphics processor unit on 1698 corneal confocal microscopy images; for external validation, it was further tested on 2137 images. The algorithm was developed to identify total nerve fibre length, branch points, tail points, number and length of nerve segments, and fractal numbers. Sensitivity analyses were undertaken to determine the AUC for ACCMetrics and our algorithm for the diagnosis of diabetic neuropathy. Results The intraclass correlation coefficients for our algorithm were superior to those for ACCMetrics for total corneal nerve fibre length (0.933 vs 0.825), mean length per segment (0.656 vs 0.325), number of branch points (0.891 vs 0.570), number of tail points (0.623 vs 0.257), number of nerve segments (0.878 vs 0.504) and fractals (0.927 vs 0.758). In addition, our proposed algorithm achieved an AUC of 0.83, specificity of 0.87 and sensitivity of 0.68 for the classification of participants without (n = 90) and with (n = 132) neuropathy (defined by the Toronto criteria). Conclusions/interpretation These results demonstrated that our deep learning algorithm provides rapid and excellent localisation performance for the quantification of corneal nerve biomarkers. This model has potential for adoption into clinical screening programmes for diabetic neuropathy. Data availability The publicly shared cornea nerve dataset (dataset 1) is available at http://bioimlab.dei.unipd.it/Corneal%20Nerve%20Tortuosity%20Data%20Set.htm and http://bioimlab.dei.unipd.it/Corneal%20Nerve%20Data%20Set.htm.

scholarly journals Corneal confocal microscopy identifies corneal nerve fibre loss and increased dendritic cells in patients with long COVID

2021 ◽  
pp. bjophthalmol-2021-319450
Author(s):  
Gulfidan Bitirgen ◽  
Celalettin Korkmaz ◽  
Adil Zamani ◽  
Ahmet Ozkagnici ◽  
Nazmi Zengin ◽  
...  
Keyword(s):  
Confocal Microscopy ◽  
Nerve Fibre ◽  
Fibre Length ◽  
Neurological Symptoms ◽  
University Hospital ◽  
Nerve Branch ◽  
Cross Sectional ◽  
Corneal Confocal Microscopy ◽  
Corneal Nerve ◽  
Fibre Loss

Background/AimsLong COVID is characterised by a range of potentially debilitating symptoms which develop in at least 10% of people who have recovered from acute SARS-CoV-2 infection. This study has quantified corneal sub-basal nerve plexus morphology and dendritic cell (DC) density in patients with and without long COVID.MethodsForty subjects who had recovered from COVID-19 and 30 control participants were included in this cross-sectional comparative study undertaken at a university hospital. All patients underwent assessment with the National Institute for Health and Care Excellence (NICE) long COVID, Douleur Neuropathique 4 (DN4) and Fibromyalgia questionnaires, and corneal confocal microscopy (CCM) to quantify corneal nerve fibre density (CNFD), corneal nerve branch density (CNBD), corneal nerve fibre length (CNFL), and total, mature and immature DC density.ResultsThe mean time after the diagnosis of COVID-19 was 3.7±1.5 months. Patients with neurological symptoms 4 weeks after acute COVID-19 had a lower CNFD (p=0.032), CNBD (p=0.020), and CNFL (p=0.012), and increased DC density (p=0.046) compared with controls, while patients without neurological symptoms had comparable corneal nerve parameters, but increased DC density (p=0.003). There were significant correlations between the total score on the NICE long COVID questionnaire at 4 and 12 weeks with CNFD (ρ=−0.436; p=0.005, ρ=−0.387; p=0.038, respectively) and CNFL (ρ=−0.404; p=0.010, ρ=−0.412; p=0.026, respectively).ConclusionCorneal confocal microscopy identifies corneal small nerve fibre loss and increased DCs in patients with long COVID, especially those with neurological symptoms. CCM could be used to objectively identify patients with long COVID.

scholarly journals Artificial intelligence utilising corneal confocal microscopy for the diagnosis of peripheral neuropathy in diabetes mellitus and prediabetes

Diabetologia ◽  
2021 ◽  
Author(s):  
Frank G. Preston ◽  
Yanda Meng ◽  
Jamie Burgess ◽  
Maryam Ferdousi ◽  
Shazli Azmi ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Peripheral Neuropathy ◽  
Confocal Microscopy ◽  
Large Scale ◽  
Data Augmentation ◽  
Learning Algorithm ◽  
Deep Learning Algorithm ◽  
Corneal Confocal Microscopy ◽  
Corneal Nerves ◽  
Corneal Nerve

Abstract Aims/hypothesis We aimed to develop an artificial intelligence (AI)-based deep learning algorithm (DLA) applying attribution methods without image segmentation to corneal confocal microscopy images and to accurately classify peripheral neuropathy (or lack of). Methods The AI-based DLA utilised convolutional neural networks with data augmentation to increase the algorithm’s generalisability. The algorithm was trained using a high-end graphics processor for 300 epochs on 329 corneal nerve images and tested on 40 images (1 image/participant). Participants consisted of healthy volunteer (HV) participants (n = 90) and participants with type 1 diabetes (n = 88), type 2 diabetes (n = 141) and prediabetes (n = 50) (defined as impaired fasting glucose, impaired glucose tolerance or a combination of both), and were classified into HV, those without neuropathy (PN−) (n = 149) and those with neuropathy (PN+) (n = 130). For the AI-based DLA, a modified residual neural network called ResNet-50 was developed and used to extract features from images and perform classification. The algorithm was tested on 40 participants (15 HV, 13 PN−, 12 PN+). Attribution methods gradient-weighted class activation mapping (Grad-CAM), Guided Grad-CAM and occlusion sensitivity displayed the areas within the image that had the greatest impact on the decision of the algorithm. Results The results were as follows: HV: recall of 1.0 (95% CI 1.0, 1.0), precision of 0.83 (95% CI 0.65, 1.0), F1-score of 0.91 (95% CI 0.79, 1.0); PN−: recall of 0.85 (95% CI 0.62, 1.0), precision of 0.92 (95% CI 0.73, 1.0), F1-score of 0.88 (95% CI 0.71, 1.0); PN+: recall of 0.83 (95% CI 0.58, 1.0), precision of 1.0 (95% CI 1.0, 1.0), F1-score of 0.91 (95% CI 0.74, 1.0). The features displayed by the attribution methods demonstrated more corneal nerves in HV, a reduction in corneal nerves for PN− and an absence of corneal nerves for PN+ images. Conclusions/interpretation We demonstrate promising results in the rapid classification of peripheral neuropathy using a single corneal image. A large-scale multicentre validation study is required to assess the utility of AI-based DLA in screening and diagnostic programmes for diabetic neuropathy. Graphical abstract

scholarly journals Corneal confocal microscopy identifies small fibre damage and progression of diabetic neuropathy

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Shaishav Dhage ◽  
Maryam Ferdousi ◽  
Safwaan Adam ◽  
Jan Hoong Ho ◽  
Alise Kalteniece ◽  
...  
Keyword(s):  
Confocal Microscopy ◽  
Diabetic Neuropathy ◽  
Nerve Fibre ◽  
Fibre Length ◽  
Autonomic Nerve ◽  
Perception Threshold ◽  
Fibre Density ◽  
Symptom Profile ◽  
Corneal Confocal Microscopy ◽  
Patients With Diabetes

AbstractAccurately quantifying the progression of diabetic peripheral neuropathy is key to identify individuals who will progress to foot ulceration and to power clinical intervention trials. We have undertaken detailed neuropathy phenotyping to assess the longitudinal utility of different measures of neuropathy in patients with diabetes. Nineteen patients with diabetes (age 52.5 ± 14.7 years, duration of diabetes 26.0 ± 13.8 years) and 19 healthy controls underwent assessment of symptoms and signs of neuropathy, quantitative sensory testing, autonomic nerve function, neurophysiology, intra-epidermal nerve fibre density (IENFD) and corneal confocal microscopy (CCM) to quantify corneal nerve fibre density (CNFD), branch density (CNBD) and fibre length (CNFL). Mean follow-up was 6.5 years. Glycated haemoglobin (p = 0.04), low-density lipoprotein-cholesterol (LDL-C) (p = 0.0009) and urinary albumin creatinine ratio (p < 0.0001) improved. Neuropathy symptom profile (p = 0.03), neuropathy disability score (p = 0.04), vibration perception threshold (p = 0.02), cold perception threshold (p = 0.006), CNFD (p = 0.03), CNBD (p < 0.0001), CNFL (p < 0.0001), IENFD (p = 0.04), sural (p = 0.02) and peroneal motor nerve conduction velocity (p = 0.03) deteriorated significantly. Change (∆) in CNFL correlated with ∆CPT (p = 0.006) and ∆Expiration/Inspiration ratio (p = 0.002) and ∆IENFD correlated with ∆CNFD (p = 0.005), ∆CNBD (p = 0.02) and ∆CNFL (p = 0.01). This study shows worsening of diabetic neuropathy across a range of neuropathy measures, especially CCM, despite an improvement in HbA1c and LDL-C. It further supports the utility of CCM as a rapid, non-invasive surrogate measure of diabetic neuropathy.

scholarly journals Corneal confocal microscopy detects small fibre neurodegeneration in Parkinson's disease using automated analysis

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Sze Hway Lim ◽  
Maryam Ferdousi ◽  
Alise Kalteniece ◽  
Lewis Kass-Iliyya ◽  
Ioannis N. Petropoulos ◽  
...  
Keyword(s):  
Parkinson’S Disease ◽  
Parkinson's Disease ◽  
Confocal Microscopy ◽  
Nerve Fibre ◽  
Fibre Length ◽  
Mean Difference ◽  
Nerve Branch ◽  
Corneal Confocal Microscopy ◽  
Corneal Nerve ◽  
Branch Density

AbstractWe studied the utility of corneal confocal microscopy (CCM) in detecting a reduction in corneal nerve parameters in a large cohort of patients with Parkinson’s disease (PD) compared to controls using a fully automated potentially scalable method of analysis. We also assessed if CCM parameters are related to the severity and sub-type of PD. 98 participants with PD and 26 healthy controls underwent CCM with automated corneal nerve quantification, MDS-UPDRS III, Hoehn and Yahr scale, Montreal Cognitive Assessment, Parkinson’s Disease Questionnaire-39 and PD subtype assessment. Corneal nerve fibre density (mean difference: − 5.00 no/mm2, 95% confidence interval (CI) [− 7.89, − 2.12], p = 0.001), corneal nerve branch density (mean difference: − 10.71 no/mm2, 95% CI [− 16.93, − 4.48], p = 0.003), corneal total branch density (mean difference: − 14.75 no/mm2, 95% CI [− 23.58, − 5.92], p = 0.002), and corneal nerve fibre length (mean difference: − 2.57 mm/mm2, 95% CI [− 4.02, − 1.12], p = 0.001) were significantly lower in PD participants compared to controls. There was no correlation between corneal nerve parameters and duration, severity or subtype of PD, cognitive function or quality of life. CCM with automated corneal nerve analysis identifies nerve fibre damage and may act as a biomarker for neurodegeneration in PD.

Bilateral morphometric analysis of corneal sub-basal nerve plexus in patients undergoing unilateral cataract surgery: a preliminary in vivo confocal microscopy study

2020 ◽  
pp. bjophthalmol-2019-315449 ◽  
Author(s):  
Giuseppe Giannaccare ◽  
Federico Bernabei ◽  
Marco Pellegrini ◽  
Fabio Guaraldi ◽  
Federica Turchi ◽  
...  
Keyword(s):  
Confocal Microscopy ◽  
Cataract Surgery ◽  
Nerve Fibre ◽  
Fibre Length ◽  
Nerve Plexus ◽  
In Vivo Confocal Microscopy ◽  
Corneal Nerve ◽  
Branch Density ◽  
Unilateral Cataract

AimsTo evaluate bilateral morphometric changes of corneal sub-basal nerve plexus (CSNP) occurring after unilateral cataract surgery by in vivo confocal microscopy (IVCM) images analysed with automated software.MethodsIVCM was performed before (V0) and 1 month after surgery (V1) in both operated eyes (OEs) and unoperated eyes (UEs) of 30 patients. Thirty age and sex-matched subjects acted as controls. Corneal nerve fibre density (CNFD), corneal nerve branch density (CNBD), corneal nerve fibre length (CNFL), corneal nerve total branch density (CTBD), corneal nerve fibre area (CNFA), corneal nerve fibre width, corneal nerve fractal dimension (CNFrD) and dendritic cells density were calculated.ResultsMean CNFD, CNBD, CNFL, CTBD, CNFA and CNFrD significantly decreased at V1 versus V0 in both eyes (respectively, 15.35±7.00 vs 21.21±6.56 n/mm2 in OEs and 20.11±6.69 vs 23.20±7.26 in UEs; 13.57±12.16 vs 26.79±16.91 n/mm2 in OEs and 24.28±14.88 vs 29.76±15.25 in UEs; 9.67±3.44 mm/mm2 vs 13.49±3.42 in OEs and 12.53±3.60 vs 14.02±3.82 in UEs; 22.81±18.77 vs 42.25±24.64 n/mm2 in OEs and 38.06±20.52 vs 43.93±22.27 in UEs; 0.0040±0.0021 vs 0.0058±0.0020 mm2/mm2 in OEs and 0.0049±0.0016 vs 0.0057±0.0019 in UEs; 1.418±0.058 vs 1.470±0.037 in OEs and 1.466±0.040 vs 1.477±0.036 in UEs; always p<0.049).ConclusionPatients undergoing cataract surgery exhibit bilateral alterations of CSNP. This finding could have broad implications in the setting of sequential cataract surgery.

scholarly journals Using corneal confocal microscopy to compare Mecobalamin intramuscular injections vs oral tablets in treating diabetic peripheral neuropathy: a RCT

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Yuanjin Zhang ◽  
Dongsheng Fan ◽  
Yixuan Zhang ◽  
Shuo Zhang ◽  
Haikun Wang ◽  
...  
Keyword(s):  
Peripheral Neuropathy ◽  
Confocal Microscopy ◽  
Nerve Fibre ◽  
Diabetic Peripheral Neuropathy ◽  
Controlled Study ◽  
Oral Tablet ◽  
Injection Treatment ◽  
Intramuscular Injections ◽  
Corneal Confocal Microscopy ◽  
Corneal Nerve

AbstractThis randomized controlled study used corneal confocal microscopy (CCM) to compare the efficacy of Mecobalamin intramuscular injections vs oral tablets in treating mild to moderate diabetic peripheral neuropathy (DPN) by detecting early nerve fiber repair. Enrolled patients were randomized approximately 1:1 to receive Mecobalamin intramuscular injections (0.5 mg/day, 3 times/week) or Mecobalamin oral tablets (1.5 mg/day) for 8 weeks. Primary outcome was change of inferior whorl length (IWL) from baseline. Secondary outcomes included changes of corneal nerve fibre length (CNFL), corneal nerve fibre density (CNFD), corneal nerve branch density (CNBD) and the Survey of Autonomic Symptoms (SAS). 15 (93.75%) patients in the injection group and 17 (89.47%) patients in the tablet group completed the study. The injection treatment significantly improved patients’ IWL from baseline (21.64 ± 3.00 mm/mm2 vs 17.64 ± 4.83 mm/mm2, P < 0.01) while the tablet treatment didn’t. Additionally, the injection treatment led to significantly improved CNFL, CNBD and SAS from baseline (all P < 0.05) while the tablet treatment did not. No patient experienced any adverse events. In conclusion, CCM is sensitive enough to detect the superior efficacy of 8-week Mecobalamin intramuscular injection treatment for DPN compared to the oral tablet treatment.ClinicalTrials.gov registration number: NCT04372316 (30/04/2020).

scholarly journals Combining In Vivo Corneal Confocal Microscopy with Deep Learning-based Analysis Reveals Sensory Nerve Fiber Loss in Acute SIV Infection

2020 ◽  
Author(s):  
Megan E. McCarron ◽  
Rachel L. Weinberg ◽  
Jessica M. Izzi ◽  
Suzanne E. Queen ◽  
Stuti L. Misra ◽  
...  
Keyword(s):  
Fractal Dimension ◽  
Deep Learning ◽  
Confocal Microscopy ◽  
Nerve Fiber ◽  
Fiber Length ◽  
Sensory Nerve ◽  
Corneal Confocal Microscopy ◽  
Siv Infection ◽  
Corneal Nerve

AbstractPurposeTo characterize corneal subbasal nerve plexus morphologic features using in vivo corneal confocal microscopy (IVCM) in normal and SIV-infected macaques and to implement automated assessments using novel deep learning-based methods customized for macaque studies.MethodsIn vivo corneal confocal microscopy images were collected from both male and female age-matched specific-pathogen free rhesus and pigtailed macaques housed at the Johns Hopkins University breeding colony using the Heidelberg HRTIII with Rostock Corneal Module. We also obtained repeat IVCM images of 12 SIV-infected animals including pre-infection and 10 day post-SIV infection time-points. All IVCM images were analyzed using a novel deep convolutional neural network architecture developed specifically for macaque studies.ResultsDeep learning-based segmentation of subbasal nerves in IVCM images from macaques demonstrated that corneal nerve fiber length (CNFL) and fractal dimension measurements did not differ between species, but pigtailed macaques had significantly higher baseline corneal nerve fiber tortuosity than rhesus macaques (P = 0.005). Neither sex nor age of macaques was associated with differences in any of the assessed corneal subbasal nerve parameters. In the SIV/macaque model of HIV, acute SIV infection induced significant decreases in both corneal nerve fiber length and fractal dimension (P= 0.01 and P= 0.008 respectively).ConclusionsThe combination of IVCM and objective, robust, and rapid deep-learning analysis serves as a powerful noninvasive research and clinical tool to track sensory nerve damage, enabling early detection of neuropathy. Adapting the deep-learning analyses to human corneal nerve assessments will refine our ability to predict and monitor damage to small sensory nerve fibers in a number of clinical settings including HIV, multiple sclerosis, Parkinson’s disease, diabetes, and chemotherapeutic neurotoxicity.

scholarly journals A novel investigation method for axonal damage in neuromyelitis optica spectrum disorder: In vivo corneal confocal microscopy

2021 ◽  
Vol 7 (1) ◽  
pp. 205521732199806
Author(s):  
Ayşe Altıntaş ◽  
Ayse Yildiz-Tas ◽  
Sezen Yilmaz ◽  
Betul N Bayraktutar ◽  
Melis Cansu Comert ◽  
...  
Keyword(s):  
Confocal Microscopy ◽  
Nerve Fibre ◽  
Disease Severity ◽  
Neuromyelitis Optica ◽  
Spectrum Disorder ◽  
Neuromyelitis Optica Spectrum Disorder ◽  
Rnfl Thickness ◽  
Corneal Confocal Microscopy ◽  
Corneal Nerve

Background Neuromyelitis optica spectrum disorder (NMOSD) is an inflammatory autoimmune disorder that damages optic nerves, brainstem, and spinal cord. In vivo corneal confocal microscopy (IVCM) is a noninvasive technique that provides corneal images with dendritic cells (DCs) and corneal subbasal nerve plexus (SBP), which arises from the trigeminal nerve. Objective We investigated corneal SBP changes in NMOSD and proposed IVCM as a potential new disease severity biomarker for NMOSD. Methods Seventeen age-sex matched NMOSD patients and 19 healthy participants underwent complete neurologic and ophthalmologic examinations. The duration of disease, first symptom, presence of optic neuritis attack, antibody status, Expanded Disability Status Scale(EDSS) score and disease severity score(DSS) were recorded. Retinal nerve fibre layer (RNFL) thickness was measured with optical coherence tomography, and corneal SBP images were taken with IVCM. Results NMOSD patients had significantly reduced corneal nerve fibre lenght-density and corneal nerve branch lenght-density compared with controls, while DC density was increased. NMOSD patients also showed significantly reduced RNFL thickness compared with controls. EDSS,DSS levels were inversely correlated with IVCM parameters. Conclusion We observed significant corneal nerve fibre loss in NMOSD patients in relation to disease severity. IVCM can be a candidate noninvasive imaging method for axonal damage assessment in NMOSD that warrants further investigation.

Exploring the optimal diagnostic threshold value of corneal nerve fibre length (CNFL) for diabetic neuropathy (DN) identification

2017 ◽  
Vol 41 (5) ◽  
pp. S62
Author(s):  
Leif E. Lovblom ◽  
Vera Bril ◽  
Andrej Orszag ◽  
Katie Edwards ◽  
Nicola Pritchard ◽  
...  
Keyword(s):  
Diabetic Neuropathy ◽  
Nerve Fibre ◽  
Fibre Length ◽  
Threshold Value ◽  
Diagnostic Threshold ◽  
Corneal Nerve
Sign in / Sign up

Export Citation Format

Share Document