Diagnostic Performance of Fundus Autofluorescence for
Detecting Polypoidal Choroidal Vasculopathy
Prakhunhungsit S, MD¹, Rodanant N, MD¹, Phasukkijwatana N, MD, PhD¹, Narongkiatikhun S, MD¹,
Thoongsuwan S, MD¹
Affiliation : ¹ Department of Ophthalmology, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand
Objective: To investigate the efficacy of fundus autofluorescence (FAF) for differentiating polypoidal choroidal vasculopathy
(PCV) from neovascular age-related macular degeneration (nAMD), and to compare FAF and optical coherence tomography
(OCT) findings.
Materials and Methods: One hundred forty-one PCV and nAMD patients were retrospectively reviewed. The FAF findings were categorized into ring pattern and patch pattern. The OCT characteristics were grouped into 1) steep pigment epithelial detachment (sPED), 2) notched PED (nPED), 3) double-layer sign (DLS), and 4) hyporeflective lumen with PED.
Results: Seventy-six PCV patients were PCV (male 44.7%) and 65 patients were nAMD (male 53.8%). The sensitivity and specificity of the ring pattern in the FAF findings were 45.7% and 76.9% while the patch pattern was 59.2% and 30.8%, respectively. The PPV and NPV were 68.1% and 56.8% for the ring pattern, and 59.2% and 39.2% for the patch pattern. The ring pattern was found more frequently in the PCV group (n=32, 68%) than in the nAMD group (n=15, 32%; OR 2.8 [1.33 to 5.90]; p=0.006). Significant associations of the FAF and OCT findings were found in the groups of ring pattern and sPED (OR 6.28 [2.89 to 13.68]; p<0.001), and of patch pattern and DLS (OR 7.00 [1.56 to 31.33]; p=0.004).
Conclusion: The sensitivity and specificity of the FAF findings were low, which precludes the use of FAF as a sole diagnostic tool for PCV. However, the significant associations between the FAF and SD-OCT emphasize the use for a multimodal approach to the non-invasive diagnosis of PCV.
Keywords : Fundus autofluorescence, Polypoidal choroidal vasculopathy, Sensitivity, Specificity, Positive predictive value,
Negative predictive value
All Articles
Download