Sensitivity and specificity of machine learning classifiers for glaucoma diagnosis using Spectral Domain OCT and standard automated perimetry

Silva, Fabrício R; Vidotti, Vanessa Gonçalves; Cremasco, Fernanda; Dias, Marcelo Fernandes Pacheco; Gomi, Edson Satoshi; Costa, Vital Paulino

doi:10.1590/s0004-27492013000300008

Cited by 53 publications

(50 citation statements)

References 29 publications

(44 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Optical coherence tomography (OCT) is a useful tool that can be used to evaluate the differences between NMO- ON and MS-ON [38,39]. OCT showed more severe retinal damage after ON episodes in NMO compared to relapsingremitting MS [40,41].…”

Section: Discussionmentioning

confidence: 99%

Features of anti-aquaporin 4 antibody-seropositive Chinese patients with neuromyelitis optica spectrum optic neuritis

Wang

et al. 2015

J Neurol

View full text Add to dashboard Cite

The detection of anti-aquaporin-4 autoantibody (AQP-4 Ab) is crucial to detect patients who will develop neuromyelitis optica (NMO); however, there are few studies on the AQP-4 Ab serostatus of patients with neuromyelitis optica spectrum ON. We analyzed the clinical and paraclinical features of neuromyelitis optica spectrum ON patients in China according to the patients' AQP4-Ab serostatus. 125 patients with recurrent and bilateral ON with simultaneous attacks were divided into AQP-4 Ab-seropositive and -seronegative groups. Demographic, clinical, serum autoantibody data, connective tissue disorders (CTDs), visual performance were compared. A Visual Acuity (VA) of less than 0.1 during acute ON attacks occurred more frequently in the seropositive group (p = 0.023); however, there was not a significant difference between groups on VA recovery after the first attack. The seropositive group experienced the worst outcome during the last attack (p = 0.017). Other co-existing autoimmunity antibodies (p < 0.001) and CTDs (p < 0.001) were more prevalent in seropositive patients. There were no significant differences on VA recovery and RNFLT combined with other autoantibodies or CTDs. The two groups did not differ significantly with regard to time to relapse, annualized relapse rates, time of diagnosis NMO, or RNFL. There were no significant differences on VA recovery and RNFLT combined with other autoantibodies or CTDs. RNFLT was thinner in NMO seropositive patients. Although AQP-4 Ab expression predicted poor visual outcome, positive patients were usually associated with mild symptoms at first onset. Anti-SSA/SSB antibody or Sjögren syndrome may be associated with AQP-4 Ab in neuromyelitis optica spectrum ON.

show abstract

Section: Discussionmentioning

confidence: 99%

Features of anti-aquaporin 4 antibody-seropositive Chinese patients with neuromyelitis optica spectrum optic neuritis

Wang

et al. 2015

J Neurol

View full text Add to dashboard Cite

show abstract

“…All OCT examinations were carried out using a high-definition spectral-domain optical coherence tomography (HD-OCT) (Cirrus HD-OCT; Carl Zeiss Meditec Inc., Dublin, CA). Both the Macular Cube 512×128 scan and RNFL measurement by the Optic Disc Cube 200×200 protocol were performed on all eyes [29]. All the data were analyzed using the Statistical Package for Social Sciences version 19.0 (SPSS Inc, Chicago, IL, USA).…”

Section: Methodsmentioning

confidence: 99%

Frequency of Autoantibodies and Connective Tissue Diseases in Chinese Patients with Optic Neuritis

Zhang

et al. 2014

PLoS ONE

View full text Add to dashboard Cite

BackgroundOptic neuritis (ON) is often associated with other clinical or serological markers of connective tissue diseases (CTDs). To date, the effects of autoantibodies on ON are not clear.PurposeTo assess the prevalence, clinical patterns, and short outcomes of autoantibodies and Sjögren’s syndrome (SS) involvement in Chinese ON patients and evaluate the relationship between ON, including their subtypes, and autoantibodies.MethodsA total of 190 ON patients were divided into recurrent ON (RON), bilateral ON (BON), and isolated monocular ON (ION). Demographic, clinical, and serum autoantibodies data were compared between them with and without SS involvement. Serum was drawn for antinuclear antibody (ANA), extractable nuclear antigen antibodies (SSA/SSB), rheumatoid factor (RF), anticardiolipin antibodies (ACA), and anti-double-stranded DNA antibody (A-ds DNA), anticardiolipin antibody (ACLs), anti-β2-glycoprotein I (β2-GPI) and Aquaporin-4 antibodies (AQP4-Ab). Spectral-domain optical coherence tomography (SD-OCT) was used to evaluate the atrophy of the optic nerve.Results68 patients (35.79%) had abnormal autoantibodies, 26(13.68%) patients met diagnostic criteria for CTDs, including 15(7.89%) patients meeting the criteria for SS. Antibodies including SSA/SSB 23 (30.26%) (p1 and p 2<0.001) and AQP4–Ab10 (13.16%) (p1 = 0.044, p2 = 0.01) were significantly different in patients in the RON group when compared with those in the BON (P1 = RON VS ION) and ION (p2 = RON VS ION) groups. SS was more common in RON patients (p1 = 0.04, p2 = 0.028). There was no significant difference between SSA/SSB positive and negative patients in disease characteristics or severity. Similar results were obtained when SS was diagnosed in SSA/SSB positive patients.ConclusionRON and BON were more likely associated with abnormal autoantibodies; furthermore, AQP4 antibody, SSA/SSB and SS were more common in the RON patients. AQP4 antibodydetermination is crucial in RON patients who will develop NMO. However, when compared with other autoantibodies, SSA/SSB detected in patients was not significantly associated with disease characteristics or severity.

show abstract

“…They got 0.877 of area under the ROC value using RF. Recently, Silva et al [8] tested almost all of the classifiers using Spectral Domain optical coherence tomography (OCT) and standard automated perimetry. They got 0.946 as the best aROC value using RF.…”

Section: Introductionmentioning

confidence: 99%

Development of machine learning models for diagnosis of glaucoma

2017

View full text Add to dashboard Cite

The study aimed to develop machine learning models that have strong prediction power and interpretability for diagnosis of glaucoma based on retinal nerve fiber layer (RNFL) thickness and visual field (VF). We collected various candidate features from the examination of retinal nerve fiber layer (RNFL) thickness and visual field (VF). We also developed synthesized features from original features. We then selected the best features proper for classification (diagnosis) through feature evaluation. We used 100 cases of data as a test dataset and 399 cases of data as a training and validation dataset. To develop the glaucoma prediction model, we considered four machine learning algorithms: C5.0, random forest (RF), support vector machine (SVM), and k-nearest neighbor (KNN). We repeatedly composed a learning model using the training dataset and evaluated it by using the validation dataset. Finally, we got the best learning model that produces the highest validation accuracy. We analyzed quality of the models using several measures. The random forest model shows best performance and C5.0, SVM, and KNN models show similar accuracy. In the random forest model, the classification accuracy is 0.98, sensitivity is 0.983, specificity is 0.975, and AUC is 0.979. The developed prediction models show high accuracy, sensitivity, specificity, and AUC in classifying among glaucoma and healthy eyes. It will be used for predicting glaucoma against unknown examination records. Clinicians may reference the prediction results and be able to make better decisions. We may combine multiple learning models to increase prediction accuracy. The C5.0 model includes decision rules for prediction. It can be used to explain the reasons for specific predictions.

show abstract

Sensitivity and specificity of machine learning classifiers for glaucoma diagnosis using Spectral Domain OCT and standard automated perimetry

Cited by 53 publications

References 29 publications

Features of anti-aquaporin 4 antibody-seropositive Chinese patients with neuromyelitis optica spectrum optic neuritis

Features of anti-aquaporin 4 antibody-seropositive Chinese patients with neuromyelitis optica spectrum optic neuritis

Frequency of Autoantibodies and Connective Tissue Diseases in Chinese Patients with Optic Neuritis

Development of machine learning models for diagnosis of glaucoma

Contact Info

Product

Resources

About