Session Title: Free Paper Session 4: AMD II
Session Date/Time: Thursday 07/09/2017 | 11:00-12:30
Paper Time: 11:30
Venue: Room 111
First Author: : P.Nderitu UK
Co Author(s): : R. Muniraju
To evaluate the diagnostic utility of fluorescein fundus angiography (FFA) and indocyanine green (ICG) angiography and assess the correlation with spectral domain ocular coherence tomography (SD-OCT) in the diagnosis of neovascular age-related macular degeneration (nAMD) in a large university district general hospital and compare findings to the reported literature.
A university London district general hospital serving over 400,000 patients and receiving referrals from general practice, community optometry services, acute eye services and internal referrals.
A retrospective review of SD-OCT, FFA and ICG of all patients who were tested between the June and Nov 2016 (5-month period). All FFA and ICG images were reported by ophthalmology consultants routinely as part of the nAMD service and findings were abstracted directly from the electronic medical records (EMR). SD-OCT findings, which were reported by medical retina and AMD consultants, specialist nurses and optometrists were also extracted as reported in the EMR. SD-OCT images were acquired using the Topcon Triton (Plus) DRI 3D OCT or the Ziess Cirrus HD-OCT 4000. FFA and ICG images were acquired on the Zeiss Visucam 500.
There were 186 FFAs performed of which 48 also had an ICG. Choroidal neovascularisation (CNV) was suspected in 59.3% of all FFAs (n=62) and 33% of all ICGs (n=12); 99.1% of suspected CNV patients had an SD-OCT. The mean age of suspected CNV patients was 81 years and the mean log minimal angle of resolution (logMAR) best-corrected visual acuity was 0.55. Reported SD-OCT findings in suspected CNV patients were pigment epithelial detachment (PED) 53%, subretinal/intraretinal fluid (SRF/IRF) 93% and subretinal membrane (SRM) 14%. In those who had an ICG, reported SD-OCT findings were PED (85%), SRF/IRF (92%) and SRM (8%). In patients with SRF/IRF on SD-OCT, FFA leak was present in 67% and absent in 19% of cases. Where there was no SRF/IRF on OCT, FFA leak was present in only 4% of cases. CNV classifications were only reported in 36% of cases of which 62% were occult or minimally classic and 38% were classic or predominantly classic. SD-OCT sensitivity and specificity for CNV diagnosis was 92% and 64% respectively with a positive predictive value (PPV) of 63% and a negative predictive value (NPV) of 93% compared to FFA. SD-OCT had a false positive rate of 36% for CNV diagnosis.
Current studies indicate that SD-OCT has >95% sensitivity and 80-92% specificity for detecting CNV in nAMD with a false positive rate of ~15% when interpreted by consultant ophthalmologists. In comparison, our study from a multidisciplinary nAMD team demonstrated similar SD-OCT sensitivity for the diagnosis of CNV in nAMD but with lower specificity and greater false positives. SD-OCT reporting of SRMs was under represented, as was the categorization of CNV into its various subtypes. Absence of SRF/IRF on SD-OCT was a good predictor of absence of FFA leak. SD-OCT is a useful adjunct in the diagnosis of nAMD, but due to the high false positives, it cannot currently replace FFA in the investigation of nAMD. Consistent reporting of all CNV features present on SD-OCT, especially when SD-OCTs are reported by a multidisciplinary team may help increase the test sensitivity and specificity in diagnosing nAMD. However, in the absence of signs of nAMD on SD-OCT, the likelihood of CNV being the underlying diagnosis at FFA is low. Therefore, SD-OCT is helpful in guiding decision making in cases requiring a more urgent FFA and it provides helpful information that adds to the diagnosis of CNV in suspected nAMD.