Background  

Evaluating the presence of leakage on fluorescein angiography (FA) in patients with age-related macular degeneration (AMD) retreated with photodynamic therapy (PDT) can be difficult. New diagnostic tools such as optical coherence tomography (OCT) might help to optimize PDT management.

Methods  

Thirty AMD patients scheduled for regular follow-up FA in conjunction with PDT treatment were also scanned with OCT. Follow-up data at 9 months were retrieved from the patients’ medical records. Inter-observer agreement [kappa (κ) coefficient] for the presence of leakage on FA, for OCT parameters for leakage, and agreement between FA and OCT evaluations were calculated. The indication for retreatment was evaluated using the leakage analysis based on FA alone, OCT alone, and both examinations combined, and compared to the actual follow-up of the patients at 9 months.

Results  

Agreement between the two observers for the presence of leakage on FA was moderate (κ=0.51). OCT agreement between the two observers for the presence of leakage was good (κ=0.85). Agreement between FA and OCT for the presence of leakage was poor (κ=0.16). Follow-up data at 9 months on all patients were analyzed. Seven out of 30 patients were not retreated at the time of examination, and four of these patients (57%) remained stable without further treatment. Twenty-three patients did receive a PDT treatment at the time of examination; and eight of these patients did not show leakage on OCT, and five of these patients (62%) remained stable without additional treatment. In contrast, only three out of 15 patients (20%) with leakage on both FA and OCT remained stable during this 9 month follow-up period.

Conclusions  

Inter-observer agreement for the presence of leakage was moderate for FA and good for OCT. There was considerable disagreement between leakage as judged by OCT and by FA. OCT could be of help in the decision regarding PDT retreatment. Assuming that 57% of the patients without leakage either on FA or OCT would remain stable without retreatment, the rate of probable ineffective retreatment could be reduced from 35% to 20%.