Authors: Raviv Katz1, Ana Rita Santos2,3, Dalila Alves2, Inês Laíns1,2, Jay C. Wang1, John B. Miller1, João Figueira2,4,6, Rufino Silva2,4,5,6

  1. Retina Service, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Department of Ophthalmology Boston, MA, 243 Charles Street, Boston, MA 02114.
  2. Association for Innovation and Biomedical Research on Light, Coimbra, Portugal, Coimbra, Portugal.
  3. Department of Orthoptics, School of Health, Polytechnic of Porto, Porto, Portugal
  4. Department of Ophthalmology. Centro Hospitalar e Universitário de Coimbra (CHUC). Portugal
  5. Coimbra Institute for Clinical and Biomedical Research. Faculty of Medicine.University of Coimbra (iCBR- FMUC)
  6. Faculty of Medicine, University of Coimbra, Coimbra, Portugal.

Purpose: To evaluate the associations between anti-VEGF therapy and central choroidal thickness (CCT), choroidal vascular density (CVD) and choroidal vascular volume (CVV), in patients with Diabetic Macular Edema (DME) using Swept Source OCT (SS-OCT), and to correlate these findings with treatment visual outcomes.

Methods: Prospective longitudinal study, including consecutive patients with treatment-naive DME. All patients received monthly intravitreal injections (IVT) of ranibizumab for 3 months (loading dose) followed by a treat-and-extend regimen during a 12 months period. BCVA and 3D horizontal volume macular SS-OCT scans (Topcon® DRI OCT-1 Atlantis) were obtained before 1st injection (M0), 1 month after loading dose (M3), and at 6 (M6) and 12 months (M12) after 1st injection. CCT was calculated as the mean central 1mm value of ETDRS grid. Enface SS-OCT images of choroidal vasculature were binarized to calculate CVD and CVV. CVD was defined as the percent area occupied by choroidal vessels in macular region (6-mm diameter circle centred on fovea) and CVV was calculated by multiplying the average CVD by the macular area and choroidal thickness. Treatment visual outcome was defined as BCVA improvement after M3 and categorized into 2 groups: Good Responders(≥5 letters) and Poor Responders(<5 letters).

Results: Twenty-three naïve DME eyes were included. After the loading dose of IVT ranibizumab (M3), 17 eyes (73.9%) were good responders and 7 (30.4%) poor responders. At baseline, good responders showed a thicker choroid compared with poor responders (199.7±79.6μm vs 182.5±60.4μm;p=0.134). Macular CVD and CVV were also significantly higher in good responders (CVD=0.26±0.06 vs 0.21±0.03; CVV=1.73±0.95 vs 1.28±0.48;p=0.151). After treatment, two distinct behaviors were observed: a significant decrease of CCT in good responders (-11.3%;p=0.014) and an increase in poor responders that did not reach statistical significance (+8.5%;p=0.576). CVD and CVV showed analogous changes with statistically significant reductions in good responders and increases in poor responders (+16.2%;p=0.006 and +34.1%;p=0.134,respectively). Reduced CVD at baseline identified well the good responders to anti-VEGF treatment (ROC AUC=0.74;p=0.030).

Conclusions: Choroidal indices such as CVD and CVV, measured at baseline, seems to discriminate the good and poor responders to anti-VEGF therapy in DME patients, and may be robust predictors of treatment response.

Financial Disclosures: No relevant financial disclosures for any of the authors in this abstract.