Deep learning versus human graders for classifying diabetic retinopathy severity in a nationwide screening program.

Deep learning algorithms have been used to detect diabetic retinopathy (DR) with specialist-level accuracy. This study aims to validate one such algorithm on a large-scale clinical population, and compare the algorithm performance with that of human graders. A total of 25,326 gradable retinal images of patients with diabetes from the community-based, nationwide screening program of DR in Thailand were analyzed for DR severity and referable diabetic macular edema (DME). Grades adjudicated by a panel of international retinal specialists served as the reference standard. Relative to human graders, for detecting referable DR (moderate NPDR or worse), the deep learning algorithm had significantly higher sensitivity (0.97 vs. 0.74, p < 0.001), and a slightly lower specificity (0.96 vs. 0.98, p < 0.001). Higher sensitivity of the algorithm was also observed for each of the categories of severe or worse NPDR, PDR, and DME (p < 0.001 for all comparisons). The quadratic-weighted kappa for determination of DR severity levels by the algorithm and human graders was 0.85 and 0.78 respectively (p < 0.001 for the difference). Across different severity levels of DR for determining referable disease, deep learning significantly reduced the false negative rate (by 23%) at the cost of slightly higher false positive rates (2%). Deep learning algorithms may serve as a valuable tool for DR screening.

Erratum: Author Correction: Deep learning versus human graders for classifying diabetic retinopathy severity in a nationwide screening program.

[This corrects the article DOI: 10.1038/s41746-019-0099-8.].

Valsalva-Like Retinopathy Secondary to Pancytopenia following Induction of Etoposide and Ifosfamide.

Etoposide and ifosfamide are chemotherapeutic agents used frequently in the treatment of sarcomas and hematologic malignancies. Ocular side effects are rarely reported. We describe a case of a patient on etoposide and ifosfamide who presented with unilateral vision loss, anemia, and thrombocytopenia. The patient was found to have a large subinternal limiting membrane hemorrhage in the right eye that is thought to be related to his anemia and thrombocytopenia. The hemorrhage resolved spontaneously after 10 days. This case illustrates how bone marrow suppression by chemotherapeutic agents may indirectly contribute to retinal hemorrhages resulting in at least transient vision loss.

Intraocular penetration of systemic antibiotics in eyes with penetrating ocular injury.

PURPOSE: To determine whether penetrating scleral or corneal injury can enhance intraocular penetration of systemic moxifloxacin, vancomycin, and ceftazidime. METHODS: Thirty rabbits were divided into 3 groups for each antibiotic and then further subdivided to receive either scleral or corneal injury to the right eye. The left eye served as a control. Intravenous antibiotics were given following injury, and eyes were subsequently enucleated. Vitreous antibiotic concentration was determined by high-performance liquid chromatography analysis. Plasma concentration was measured for comparison. RESULTS: Intravitreal moxifloxacin concentration was unchanged by injury. Minimum inhibitory concentration (MIC90) was achieved in the vitreous against the most common gram-positive endophthalmitis-causing organisms. Intravitreal vancomycin levels were not enhanced by injury and did not reach the MIC90 for gram-positive organisms commonly causing intraocular infection. Intravitreal ceftazidime was increased in the injured eyes, 67% and 73% higher in scleral and corneal injury eyes. It reached MIC90 of many gram-negative bacteria. CONCLUSIONS: Intravitreal antibiotic penetration of systemic antibiotics with or without penetrating ocular injury varies depending on the antibiotic. For prevention or treatment of gram-positive-bacteria-causing endophthalmitis, intravitreal vancomycin is necessary and provides the most reliable coverage. Systemic ceftazidime can be used for many gram-negative bacteria, but intravitreal injection is recommended for better coverage, especially for more-potent organisms. Systemic moxifloxacin can be considered for most gram-positive and -negative infections due to its excellent intraocular penetration and broad coverage, but the patient's previous history of its topical use and increasing resistance patterns must be considered.

Nd:YAG membranotomy for preretinal hemorrhage secondary to valsalva retinopathy.

PURPOSE: To report four cases of premacular hemorrhage secondary to valsalva retinopathy treated with Nd:YAG membranotomy and discuss techniques as well as the literature. DESIGN: Retrospective case series. METHODS: A retrospective review was conducted for four patients with vision obstructing hemorrhage secondary to valsalva retinopathy. These patients were all treated with Nd:YAG membranotomy. RESULTS: Four patients with premacular hemorrhage secondary to valsalva retinopathy were treated with Nd:YAG laser creating a membranotomy to drain the hemorrhage. Power settings ranged from 1.7 to 3.8 mJ. Visual acuity at presentation ranged from 20/400 (1 patient) to count fingers (3 patients). Visual acuity improved in three out of four patients after laser treatment. Final visual acuity ranged from 20/20 to 20/30 in these three patients. One patient was lost to follow up after performing laser membranotomy and therefore visual acuity after treatment was not obtained. No complications were noted. CONCLUSION: Nd:YAG membranotomy is a non-invasive, office-based treatment option that may be successfully used to treat premacular hemorrhage secondary to valsalva retinopathy.