Polypoidal Choroidal Vasculopathy: Polyp Evolution, Disease Classification

Several researchers discussed polypoidal choroidal vasculopathy (PCV) during AAO 2019.

In one scientific poster presentation, “Automated Classification of Polypoidal Choroidal Vasculopathy and Other Subtypes of AMD Using Bimodal Deep Convolutional Neural Networks,” the authors of this retrospective study utilized a bimodality deep convolutional neural network (DCNN) to recognize PCV and other age-related macular degeneration (AMD) subtypes using color fundus and optical coherence tomography (OCT). Between 2013 and 2018, 904 AMD and PCV patients and 195 controls were assessed; 2,389 anonymous fundus and OCT images were evaluated, all of which were pre-categorized into four groups: normal, dry AMD, wet AMD, or PCV. Additional comparisons were drawn using unimodal DCNN-OCT/Fundus and bimodal DCNN-Combo. The researchers employed a human-machine comparison. Final analysis included 223 images from 80 years (20 eyes per group). DCNN-Combo outperformed the other techniques, with an accuracy of 87.4%, sensitivity of 88.8% and specificity of 95.6%. DCNN-Combo performed slightly better than the best expert (Machine κ, 0.828; Human1 κ, 0.810).

In a second scientific poster presentation, “Evolution of Polypoidal Lesions Among Patients of a Multicenter Randomized Controlled Clinical Trial on Polypoidal Choroidal Vasculopathy,” researchers examined polyp evolution in symptomatic macular PCV patients and assessed differences between baseline and new polyps. This randomized, controlled, multicenter trial included 30 patients whose indocyanine green angiograms (ICGAs) were reviewed at baseline and again after three, six, 12, and 24 months. A reading center was used to determine polyp evaluation and how it was associated with treatment. Patients treated with combination photodynamic therapy and ranibizumab had significantly higher complete polyp regression than patients receiving ranibizumab monotherapy (80.0% vs. 35.7% at 12 months; P=0.016). The combination therapy group also had lower persistent baseline polyps than the monotherapy group (6.7% vs. 50.0%; P<0.001). Most new polyps developed from month six and continued to form a more significant proportion of all polyps over the duration of treatment.