Evaluation of a deep learning system for the joint automated detection of diabetic retinopathy and age-related macular degeneration.

PURPOSE: To validate the performance of a commercially available, CE-certified deep learning (DL) system, RetCAD v.1.3.0 (Thirona, Nijmegen, The Netherlands), for the joint automatic detection of diabetic retinopathy (DR) and age-related macular degeneration (AMD) in colour fundus (CF) images on a dataset with mixed presence of eye diseases. METHODS: Evaluation of joint detection of referable DR and AMD was performed on a DR-AMD dataset with 600 images acquired during routine clinical practice, containing referable and non-referable cases of both diseases. Each image was graded for DR and AMD by an experienced ophthalmologist to establish the reference standard (RS), and by four independent observers for comparison with human performance. Validation was furtherly assessed on Messidor (1200 images) for individual identification of referable DR, and the Age-Related Eye Disease Study (AREDS) dataset (133 821 images) for referable AMD, against the corresponding RS. RESULTS: Regarding joint validation on the DR-AMD dataset, the system achieved an area under the ROC curve (AUC) of 95.1% for detection of referable DR (SE = 90.1%, SP = 90.6%). For referable AMD, the AUC was 94.9% (SE = 91.8%, SP = 87.5%). Average human performance for DR was SE = 61.5% and SP = 97.8%; for AMD, SE = 76.5% and SP = 96.1%. Regarding detection of referable DR in Messidor, AUC was 97.5% (SE = 92.0%, SP = 92.1%); for referable AMD in AREDS, AUC was 92.7% (SE = 85.8%, SP = 86.0%). CONCLUSION: The validated system performs comparably to human experts at simultaneous detection of DR and AMD. This shows that DL systems can facilitate access to joint screening of eye diseases and become a quick and reliable support for ophthalmological experts.

Morphological and topographical appearance of microaneurysms on optical coherence tomography angiography.

AIMS: To investigate retinal microaneurysms in patients with diabetic macular oedema (DME) by optical coherence tomography angiography (OCTA) according to their location and morphology in relationship to their clinical properties, leakage on fundus fluorescein angiography (FFA) and retinal thickening on structural OCT. METHODS: OCTA and FFA images of 31 eyes of 24 subjects were graded for the presence of microaneurysms. The topographical and morphological appearance of microaneurysms on OCTA was evaluated and classified. For each microaneurysm, the presence of focal leakage on FFA and associated retinal thickening on OCT was determined. RESULTS: Of all microaneurysms flagged on FFA, 295 out of 513 (58%) were also visible on OCTA. Microaneurysms with focal leakage and located in a thickened retinal area were more likely to be detected on OCTA than not leaking microaneurysms in non-thickened retinal areas (p=0.001). Most microaneurysms on OCTA were seen in the intermediate (23%) and deep capillary plexus (22%). Of all microaneurysms visualised on OCTA, saccular microaneurysms were detected most often (31%), as opposed to pedunculated microaneurysms (9%). Irregular, fusiform and mixed fusiform/saccular-shaped microaneurysms had the highest likeliness to leak and to be located in thickened retinal areas (p<0.001, p<0.001 and p=0.001). CONCLUSIONS: Retinal microaneurysms in DME could be classified topographically and morphologically by OCTA. OCTA detected less microaneurysms than FFA, and this appeared to be dependent on leakage activity and retinal thickening. Morphological appearance of microaneurysms (irregular, fusiform and mixed saccular/fusiform) was associated with increased leakage activity and retinal thickening.

Preoperative infra- and suprarenal aortic pulsatile distension is comparable between relatively young and older patients with an abdominal aortic aneurysm.

BACKGROUND: Young healthy individuals have a large aortic pulsatile distension during the cardiac cycle. In patients with an abdominal aortic aneurysm (AAA), aortic distension during the cardiac cycle is associated with stent graft migration. However, whether the pulsatile distension is larger in relatively young patients with an AAA compared to older AAA patients is unknown. This study investigated whether preoperative pulsatile aneurysm neck distension is related to age. METHODS: From our database of endovascular aneurysm repair (EVAR), we selected 25 consecutive male patients ≤65 years of age (group 1) and 25 consecutive patients >65 years of age (group 2). All patients had a preoperative electrocardiogram-triggered computed tomography angiography scan consisting of 8 phases. Aortic area and diameter changes per heartbeat were measured at 2 levels: (A) 3 cm above and (B) 1 cm below the most distal renal artery. RESULTS: In group 1 compared to group 2, distension during the cardiac cycle at level A was 1.6 ± 0.4 versus 1.5 ± 0.4 mm (P = 0.62), and the aortic area increase was 45.4 ± 19.6 versus 41.7 ± 20.8 mm(2) (P = 0.52). Aortic distension at level B was 1.4 ± 0.3 versus 1.5 ± 0.4 mm (P = 0.79), and the area increase was 35.5 ± 12.0 versus 35.0 ± 15.5 mm(2) (P = 0.90). CONCLUSIONS: Preoperative pulsatile aneurysm neck distension did not differ between younger and older patients; therefore, we do not expect young patients to have more pulsatile distension-related complications after EVAR.