Automated Detection of Macular Drusen Using Geometric Background Leveling and Threshold Selection

Abstract

Background: Age-related macular degeneration (ARMD) is the most prevalent cause of visual loss in patients older than 60 years in the United States. Observation of drusen is the hallmark finding in the clinical evaluation of ARMD. Objectives: To segment and quantify drusen found in patients with ARMD using image analysis and to compare the efficacy of image analysis segmentation with that of stereoscopic manual grading of drusen. Design: Retrospective study. Setting: University referral center. Patients: Photographs were randomly selected from an available database of patients with known ARMD in the ongoing Columbia University Macular Genetics Study. All patients were white and older than 60 years. Interventions: Twenty images from 17 patients were selected as representative of common manifestations of drusen. Image preprocessing included automated color balancing and, where necessary, manual segmentation of confounding lesions such as geographic atrophy (3 images). The operator then chose among 3 automated processing options suggested by predominant drusen type. Automated processing consisted of elimination of background variability by a mathematical model and subsequent histogram-based threshold selection. A retinal specialist using a graphic tablet while viewing stereo pairs constructed digital drusen drawings for each image. Main Outcome Measures: The sensitivity and specificity of drusen segmentation using the automated method with respect to manual stereoscopic drusen drawings were calculated on a rigorous pixel-by-pixel basis. Results: The median sensitivity and specificity of automated segmentation were 70% and 81%, respectively. After preprocessing and option choice, reproducibility of automated drusen segmentation was necessarily 100%. Conclusions: Automated drusen segmentation can be reliably performed on digital fundus photographs and result in successful quantification of drusen in a more precise manner than is traditionally possible with manual stereoscopic grading of drusen. With only minor preprocessing requirements, this automated detection technique may dramatically improve our ability to monitor drusen in ARMD.

Quantitative Comparison of Drusen Segmented on SD-OCT versus Drusen Delineated on Color Fundus Photographs

“…Furthermore, we chose to use a `composite CFP' drusen map, defining drusen and non-drusen areas as sites where any two of three graders agreed, to minimize the potential bias introduced by any one grader. When checking this composite grading to a previously published method of automated segmentation of drusen on CFP,21 the findings were remarkably similar.…”

Quantitative Comparison of Drusen Segmented on SD-OCT versus Drusen Delineated on Color Fundus Photographs

“…Furthermore, we chose to use a `composite CFP' drusen map, defining drusen and non-drusen areas as sites where any two of three graders agreed, to minimize the potential bias introduced by any one grader. When checking this composite grading to a previously published method of automated segmentation of drusen on CFP,21 the findings were remarkably similar.…”

Dynamic soft drusen remodelling in age-related macular degeneration

“…For drusen, this method has inter-observer reproducibility and correlates well with retinal experts' drawings 12. However, highly reflectant structures such as the nerve fibre layer bundles at the arcades, peripapillary atrophy and retinal pigment epithelium (RPE) hypopigmentation, exudates and scars are more likely be mistaken for drusen by the automated method than by an expert grader, requiring post-processing steps 12. For this reason we have developed the user-interactive method, in which the user initially labels regions of interest in drusen images for input to the model that exclude unwanted reflectant structures.…”

Automated drusen detection in retinal images using analytical modelling algorithms

“…Smith et al . [ 5 ] evaluated their work with a dataset of twenty images examined by one expert, obtaining a sensitivity of 0.7 and a specificity of 0.8. Therefore, we can consider AD3RI more accurate namely because it was tested against a set of eight experts in order to achieve a more reliable ground-truth.…”

Quantitative Comparison of Drusen Segmented on SD-OCT versus Drusen Delineated on Color Fundus Photographs

“…Furthermore, we chose to use a `composite CFP' drusen map, defining drusen and non-drusen areas as sites where any two of three graders agreed, to minimize the potential bias introduced by any one grader. When checking this composite grading to a previously published method of automated segmentation of drusen on CFP,21 the findings were remarkably similar.…”

Dynamic soft drusen remodelling in age-related macular degeneration

“…For drusen, this method has inter-observer reproducibility and correlates well with retinal experts' drawings 12. However, highly reflectant structures such as the nerve fibre layer bundles at the arcades, peripapillary atrophy and retinal pigment epithelium (RPE) hypopigmentation, exudates and scars are more likely be mistaken for drusen by the automated method than by an expert grader, requiring post-processing steps 12. For this reason we have developed the user-interactive method, in which the user initially labels regions of interest in drusen images for input to the model that exclude unwanted reflectant structures.…”

Automated drusen detection in retinal images using analytical modelling algorithms

“…Smith et al . [ 5 ] evaluated their work with a dataset of twenty images examined by one expert, obtaining a sensitivity of 0.7 and a specificity of 0.8. Therefore, we can consider AD3RI more accurate namely because it was tested against a set of eight experts in order to achieve a more reliable ground-truth.…”

Quantitative Comparison of Drusen Segmented on SD-OCT versus Drusen Delineated on Color Fundus Photographs

“…Furthermore, we chose to use a `composite CFP' drusen map, defining drusen and non-drusen areas as sites where any two of three graders agreed, to minimize the potential bias introduced by any one grader. When checking this composite grading to a previously published method of automated segmentation of drusen on CFP,21 the findings were remarkably similar.…”

Dynamic soft drusen remodelling in age-related macular degeneration

“…For drusen, this method has inter-observer reproducibility and correlates well with retinal experts' drawings 12. However, highly reflectant structures such as the nerve fibre layer bundles at the arcades, peripapillary atrophy and retinal pigment epithelium (RPE) hypopigmentation, exudates and scars are more likely be mistaken for drusen by the automated method than by an expert grader, requiring post-processing steps 12. For this reason we have developed the user-interactive method, in which the user initially labels regions of interest in drusen images for input to the model that exclude unwanted reflectant structures.…”

Automated drusen detection in retinal images using analytical modelling algorithms

“…Smith et al . [ 5 ] evaluated their work with a dataset of twenty images examined by one expert, obtaining a sensitivity of 0.7 and a specificity of 0.8. Therefore, we can consider AD3RI more accurate namely because it was tested against a set of eight experts in order to achieve a more reliable ground-truth.…”