Background:Soft drusen and basal linear deposit(BLinD)are two forms of the same extracellular lipid rich material that together make up an Oil Spill on Bruch’s membrane(BrM).Drusen are focal and can be recognized cli...Background:Soft drusen and basal linear deposit(BLinD)are two forms of the same extracellular lipid rich material that together make up an Oil Spill on Bruch’s membrane(BrM).Drusen are focal and can be recognized clinically.In contrast BLinD is thin and diffusely distributed,and invisible clinically,even on highest resolution OCT,but has been detected on en face hyperspectral autofluorescence(AF)imaging ex vivo.We sought to optimize histologic hyperspectral AF imaging and image analysis for recognition of drusen and sub-RPE deposits(including BLinD and basal laminar deposit),for potential clinical application.Methods:Twenty locations specifically with drusen and 12 additional locations specifically from fovea,perifovea and mid-periphery from RPE/BrM flatmounts from 4 AMD donors underwent hyperspectral AF imaging with 4 excitation wavelengths(λex 436,450,480 and 505 nm),and the resulting image cubes were simultaneously decomposed with our published non-negative matrix factorization(NMF).Rank 4 recovery of 4 emission spectra was chosen for each excitation wavelength.Results:A composite emission spectrum,sensitive and specific for drusen and presumed sub-RPE deposits(the SDr spectrum)was recovered with peak at 510-520 nm in all tissues with drusen,with greatest amplitudes at excitationsλ_(ex)436,450 and 480 nm.The RPE spectra of combined sources Lipofuscin(LF)/Melanolipofuscin(MLF)were of comparable amplitude and consistently recapitulated the spectra S1,S2 and S3 previously reported from all tissues:tissues with drusen,foveal and extra-foveal locations.Conclusions:A clinical hyperspectral AF camera,with properly chosen excitation wavelengths in the blue range and a hyperspectral AF detector,should be capable of detecting and quantifying drusen and sub-RPE deposits,the earliest known lesions of AMD,before any other currently available imaging modality.展开更多
Background:Age-related macular degeneration(AMD)and diabetic retinopathy(DR)are among the leading causes of blindness in the United States and other developed countries.Early detection is the key to prevention and eff...Background:Age-related macular degeneration(AMD)and diabetic retinopathy(DR)are among the leading causes of blindness in the United States and other developed countries.Early detection is the key to prevention and effective treatment.We have built an artificial intelligence-based screening system which utilizes a cloud-based platform for combined large scale screening through primary care settings for early diagnosis of these diseases.Methods:iHealthScreen Inc.,an independent medical software company,has developed automated AMD and DR screening systems utilizing a telemedicine platform based on deep machine learning techniques.For both diseases,we prospectively imaged both eyes of 340 unselected non-dilated subjects over 50 years of age.For DR specifically,152 diabetic patients at New York Eye and Ear faculty retina practices,ophthalmic and primary care clinics in New York city with color fundus cameras.Following the initial review of the images,308 images with other confounding conditions like high myopia and vascular occlusion,and poor quality were excluded,leaving 676 eligible images for AMD and DR evaluation.Three ophthalmologists evaluated each of the images,and after adjudication,the patients were determined referrable or non-referable for AMD DR.Concerning AMD,172 were labeled referable(intermediate or late),and 504 were non-referable(no or early).Concurrently,regarding DR,33 were referable(moderate or worse),and 643 were non-referable(none or mild).All images were uploaded to iHealthScreen’s telemedicine platform and analyzed by the automated systems for both diseases.The system performances are tested on per eye basis with sensitivity,specificity,accuracy,and kappa scores with respect to the professional graders.Results:In identifying referable DR,the system achieved a sensitivity of 97.0%and a specificity of 96.3%,and a kappa score of 0.70 on this prospective dataset.For AMD,the sensitivity was 86.6%,the specificity of 92.1%,and a kappa score of 0.76.Conclusions:The AMD and DR screening tools achieved excellent performance operating together to identify two retinal diseases prospectively in mixed datasets,demonstrating the feasibility of such tools in the early diagnosis of eye diseases.These early screening tools will help create an even more comprehensive system capable of being trained on other retinal pathologies,a goal within reach for public health deployment.展开更多
文摘Background:Soft drusen and basal linear deposit(BLinD)are two forms of the same extracellular lipid rich material that together make up an Oil Spill on Bruch’s membrane(BrM).Drusen are focal and can be recognized clinically.In contrast BLinD is thin and diffusely distributed,and invisible clinically,even on highest resolution OCT,but has been detected on en face hyperspectral autofluorescence(AF)imaging ex vivo.We sought to optimize histologic hyperspectral AF imaging and image analysis for recognition of drusen and sub-RPE deposits(including BLinD and basal laminar deposit),for potential clinical application.Methods:Twenty locations specifically with drusen and 12 additional locations specifically from fovea,perifovea and mid-periphery from RPE/BrM flatmounts from 4 AMD donors underwent hyperspectral AF imaging with 4 excitation wavelengths(λex 436,450,480 and 505 nm),and the resulting image cubes were simultaneously decomposed with our published non-negative matrix factorization(NMF).Rank 4 recovery of 4 emission spectra was chosen for each excitation wavelength.Results:A composite emission spectrum,sensitive and specific for drusen and presumed sub-RPE deposits(the SDr spectrum)was recovered with peak at 510-520 nm in all tissues with drusen,with greatest amplitudes at excitationsλ_(ex)436,450 and 480 nm.The RPE spectra of combined sources Lipofuscin(LF)/Melanolipofuscin(MLF)were of comparable amplitude and consistently recapitulated the spectra S1,S2 and S3 previously reported from all tissues:tissues with drusen,foveal and extra-foveal locations.Conclusions:A clinical hyperspectral AF camera,with properly chosen excitation wavelengths in the blue range and a hyperspectral AF detector,should be capable of detecting and quantifying drusen and sub-RPE deposits,the earliest known lesions of AMD,before any other currently available imaging modality.
基金This project was supported by NIH SBIR Grant:1R44EY031202.
文摘Background:Age-related macular degeneration(AMD)and diabetic retinopathy(DR)are among the leading causes of blindness in the United States and other developed countries.Early detection is the key to prevention and effective treatment.We have built an artificial intelligence-based screening system which utilizes a cloud-based platform for combined large scale screening through primary care settings for early diagnosis of these diseases.Methods:iHealthScreen Inc.,an independent medical software company,has developed automated AMD and DR screening systems utilizing a telemedicine platform based on deep machine learning techniques.For both diseases,we prospectively imaged both eyes of 340 unselected non-dilated subjects over 50 years of age.For DR specifically,152 diabetic patients at New York Eye and Ear faculty retina practices,ophthalmic and primary care clinics in New York city with color fundus cameras.Following the initial review of the images,308 images with other confounding conditions like high myopia and vascular occlusion,and poor quality were excluded,leaving 676 eligible images for AMD and DR evaluation.Three ophthalmologists evaluated each of the images,and after adjudication,the patients were determined referrable or non-referable for AMD DR.Concerning AMD,172 were labeled referable(intermediate or late),and 504 were non-referable(no or early).Concurrently,regarding DR,33 were referable(moderate or worse),and 643 were non-referable(none or mild).All images were uploaded to iHealthScreen’s telemedicine platform and analyzed by the automated systems for both diseases.The system performances are tested on per eye basis with sensitivity,specificity,accuracy,and kappa scores with respect to the professional graders.Results:In identifying referable DR,the system achieved a sensitivity of 97.0%and a specificity of 96.3%,and a kappa score of 0.70 on this prospective dataset.For AMD,the sensitivity was 86.6%,the specificity of 92.1%,and a kappa score of 0.76.Conclusions:The AMD and DR screening tools achieved excellent performance operating together to identify two retinal diseases prospectively in mixed datasets,demonstrating the feasibility of such tools in the early diagnosis of eye diseases.These early screening tools will help create an even more comprehensive system capable of being trained on other retinal pathologies,a goal within reach for public health deployment.
