TY - JOUR
T1 - Spectral-domain optical coherence tomography angiography of choroidal neovascularization
AU - De Carlo, Talisa E.
AU - Bonini Filho, Marco A.
AU - Chin, Adam T.
AU - Adhi, Mehreen
AU - Ferrara, Daniela
AU - Baumal, Caroline R.
AU - Witkin, Andre J.
AU - Reichel, Elias
AU - Duker, Jay S.
AU - Waheed, Nadia K.
N1 - Publisher Copyright:
© 2015 American Academy of Ophthalmology.
PY - 2015/6/1
Y1 - 2015/6/1
N2 - Purpose To describe the characteristics as well as the sensitivity and specificity of detection of choroidal neovascularization (CNV) on optical coherence tomography angiography (OCTA) using spectral-domain optical coherence tomography. Design Observational, retrospective study. Participants Seventy-two eyes of 61 subjects (48 eyes of 43 subjects with CNV, 24 eyes of 18 subjects without CNV). Methods Patients imaged using the prototype AngioVue OCTA system (Optovue, Inc, Fremont, CA) between August 2014 and October 2014 at New England Eye Center were assessed. Patients in whom CNV was identified on OCTA were evaluated to define characteristics of CNV on OCTA: size using greatest linear dimension (small, <1 mm; medium, 1-2 mm; large, >2 mm), appearance (well-circumscribed, poorly circumscribed), and presence of subretinal and intraretinal fluid. Concurrently, an overlapping second cohort of patients who underwent same-day OCTA and fluorescein angiography (FA) for suspected CNV was evaluated to estimate sensitivity and specificity of OCTA in detecting CNV using FA as ground truth. Main Outcome Measures Choroidal neovascularization appearance, CNV size, and presence of subretinal and intraretinal fluid. Results In 48 eyes, CNV was visualized on OCTA. Thirty-one eyes had CNV associated with neovascular age-related macular degeneration. Size of CNV was small in 23% (7/31), medium in 42% (13/31), and large in 35% (11/31). Poorly circumscribed vessels, subretinal fluid, and intraretinal fluid each were seen in 71% (22/31). Seven eyes had CNV associated with central serous chorioretinopathy. Size of CNV was small in 71% (5/7) and large in 29% (2/7). Seventy-one percent (5/7) had well-circumscribed vessels, 86% (6/7) had subretinal fluid, and 14% (1/7) had intraretinal fluid. Thirty eyes with OCTA and same-day FA were evaluated to determine sensitivity and specificity of CNV detection on OCTA. Sensitivity was 50% (4/8) and specificity was 91% (20/22). Conclusions Using OCTA allows the clinician to visualize CNV noninvasively and may provide a method for identifying and guiding treatment of CNV. The specificity of CNV detection on OCTA compared with FA seems to be high. Future studies with larger sample sizes are needed to elaborate better on the sensitivity and specificity of CNV detection and to illustrate clinical usefulness.
AB - Purpose To describe the characteristics as well as the sensitivity and specificity of detection of choroidal neovascularization (CNV) on optical coherence tomography angiography (OCTA) using spectral-domain optical coherence tomography. Design Observational, retrospective study. Participants Seventy-two eyes of 61 subjects (48 eyes of 43 subjects with CNV, 24 eyes of 18 subjects without CNV). Methods Patients imaged using the prototype AngioVue OCTA system (Optovue, Inc, Fremont, CA) between August 2014 and October 2014 at New England Eye Center were assessed. Patients in whom CNV was identified on OCTA were evaluated to define characteristics of CNV on OCTA: size using greatest linear dimension (small, <1 mm; medium, 1-2 mm; large, >2 mm), appearance (well-circumscribed, poorly circumscribed), and presence of subretinal and intraretinal fluid. Concurrently, an overlapping second cohort of patients who underwent same-day OCTA and fluorescein angiography (FA) for suspected CNV was evaluated to estimate sensitivity and specificity of OCTA in detecting CNV using FA as ground truth. Main Outcome Measures Choroidal neovascularization appearance, CNV size, and presence of subretinal and intraretinal fluid. Results In 48 eyes, CNV was visualized on OCTA. Thirty-one eyes had CNV associated with neovascular age-related macular degeneration. Size of CNV was small in 23% (7/31), medium in 42% (13/31), and large in 35% (11/31). Poorly circumscribed vessels, subretinal fluid, and intraretinal fluid each were seen in 71% (22/31). Seven eyes had CNV associated with central serous chorioretinopathy. Size of CNV was small in 71% (5/7) and large in 29% (2/7). Seventy-one percent (5/7) had well-circumscribed vessels, 86% (6/7) had subretinal fluid, and 14% (1/7) had intraretinal fluid. Thirty eyes with OCTA and same-day FA were evaluated to determine sensitivity and specificity of CNV detection on OCTA. Sensitivity was 50% (4/8) and specificity was 91% (20/22). Conclusions Using OCTA allows the clinician to visualize CNV noninvasively and may provide a method for identifying and guiding treatment of CNV. The specificity of CNV detection on OCTA compared with FA seems to be high. Future studies with larger sample sizes are needed to elaborate better on the sensitivity and specificity of CNV detection and to illustrate clinical usefulness.
UR - http://www.scopus.com/inward/record.url?scp=84930042719&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84930042719&partnerID=8YFLogxK
U2 - 10.1016/j.ophtha.2015.01.029
DO - 10.1016/j.ophtha.2015.01.029
M3 - Article
C2 - 25795476
AN - SCOPUS:84930042719
SN - 0161-6420
VL - 122
SP - 1228
EP - 1238
JO - Ophthalmology
JF - Ophthalmology
IS - 6
ER -