Congreso Argentino de Oftalmologia, May 2009, Buenos Aires Structural & Functional Optical Coherence Tomography for Glaucoma Diagnosis David Huang, MD, PhD Assoc. Prof. of Ophthalmology & Biomedical Engineering Doheny Eye Institute, University of Southern California Financial Interests: Optovue, Inc.: stock options, patent royalty, travel, grant Carl Zeiss Meditec, Inc.: patent royalty
NIH / NEI Grant R01 EY013516 www.aigstudy.net This work is part of the Advanced Imaging for Glaucoma (AIG) study, a multi-center, prospective longitudinal study to develop and use advanced imaging technologies to improve the detection and management of glaucoma
The Rationale for Quantitative Imaging in Glaucoma Diagnosis
Visual field has poor repeatability OHTS: 85.9% of abnormal and reliable fields were not confirmed on retest!
3 consecutive fields are required to reliably confirm glaucoma! The proportion of VF test results that were normal subsequent to a VF POAG end point in eyes whose abnormality was confirmed by 2 consecutive, abnormal, reliable test results was significantly higher (73 [66%] of 110) compared with eyes whose abnormality was confirmed by 3 consecutive, abnormal, reliable test results. (46 [12%] of 381) (P=.01). Keltner et al. for the Ocular Hypertension Treatment Study Group, Arch Ophthalmol 123:1201 (2005).
Structural loss precedes functional loss Disc change precedes VF loss in most cases
photo A photo B Subtle disc change Optic Disc Photos photo C or photo D No disc change x 30 years (without treatment) Spaeth GL, Franca Lopes J, Junk AK, et al. Systems for Staging the Amount of Optic Nerve Damage in Glaucoma: A Critical Review and New Material. Surv Ophthalmol. 2006; 51: 293-315.
Optic disc change using Stereo disc photos is subjective and qualitative with only moderate intraobserver and interobserver reproducibility Azuara-Blanco A, Katz JL, Spaeth GL, et al. Clinical agreement among glaucoma experts in the detection of glaucomatous changes of the optic disk using simultaneous stereoscopic photographs. Am J Ophthalmol. 2003; 136: 949-950.
Optic Nerve Head Photo or Drawing: documented in only ½ of initial glaucoma evaluations (395 patients in managed care setting) Fremont MA, Lee PP, Mangione CM, et al. Patterns of Care for Open Angle Glaucoma in Managed Care. Arch Ophthalmol. 2003; 121: 777-783.
Quantitative Imaging may detect glaucoma at an earlier stage N T MD -1.2 db PSD 1.75 db N T MD -1.73 db PSD 1.62 db N T MD -1.77 db PSD 1.71 db VF Normal OCT Abn. GDX Abn. HRT Abn
Quantitative Imaging may improve the detection of glaucoma progression 80 70 60 Baseline 50 40 Baseline 30 6 month 20 12 month 10 0 OCT NFLT GDx TSNIT HRT Rim Area 1 year Quantitative imaging parameters Disc photograph unchanged
Why use OCT? (rather than other imaging modalities)
Let s compare diagnostic accuracy Stratus TD-OCT GDx-ECC Scanning Laser Polarimetry HRT2 Scanning Laser Tomography
www.aigstudy.net NIH/NEI R01 EY013516 Normal Glaucoma # subjects 99 134 # eyes 198 205 Age (Year) 55+/-9 63+/-9 Female (%) 68% 59% Age bias removed by mixed effects model
Stratus OCT had significantly better diagnostic accuracy (best combination of continuous variables) Continuous scale AROC P. v. OCT Stratus: overall, Inferior or superior quadrant RNFL 0.92 GDx-ECC NFI 0.87 0.006 HRT2 C/D area ratio 0.83 0.0008
Stratus OCT had significantly better diagnostic accuracy (best combination of discrete variables) Ordinal scale* AROC P. v. OCT Stratus: overall, Inferior or superior quadrant RNFL 0.88 GDx-ECC NFI 0.80 <0.0001 HRT2 Overall moorfields 0.82 0.015 Classifications are normal, <5%,2%,1%, and <0.5%
Previous literature comparisons Paper AROC # Subjects Stratus GDx-VCC HRT2 N G Pueyo et al. J. Glaucoma 2007 Overall RNFL 0.91 NFI 0.88* MRA 0.90 66 73 Pueyo et al. ARCH SOC ESP OFTALMOL 2006 Overall RNFL 0.93 NFI 0.88 Mikelberg 0.90 66 74 Medeiros et al. Arch Ophthalmol. 2004 Inferior RNFL 0.92 NFI 0.91 LDF 0.86 66 75 Zangwill et al. Arch Ophthalmol. 2001 5 o clock RNLF 0.87 LDF 0.84 MHC N/I 0.86 50 41 NFI = nerve fiber index; MRA = Moorefields regression analysis; LDF = linear discriminant function; MHC = mean height contour, N/I = nasal/inferior
How is Fourier-domain OCT better than time-domain OCT?
A generational leap 2006 26,000 Fourier domain Speed (A-scans /sec) 400 100 Zeiss OCT1/2 1996 Zeiss Stratus 2002 Time domain 16 10 5 Resolution (μm) RTVue has 65x speed & 2x resolution of Stratus
FD OCT Simultaneous 2048 pixels at a time Small blood vessels IS/OS TD OCT Sequential 1 pixel at a time Motion artifact 1024 A-scans in 0.04 sec 512 A-scans in 1.28 sec Higher speed, higher definition and higher signal.
TD-OCT susceptible to eye movements 768 pixels (Ascans) captured in 1.92 seconds is slower than eye movements Stabilizing the retina reveals true scan path (white circles) 1 1. Koozekanani, Boyer and Roberts. Tracking the Optic Nervehead in OCT Video Using Dual Eigenspaces and an Adaptive Vascular Distribution Model ; IEEE Transactions on Medical Imaging, Vol. 22, No. 12, 2003
Scan location and eye movements affect results Properly centered Poorly centered: too inferior Poorly centered: too superior T S N I T T S N I T T S N I T Normal Double Hump Inferior RNFL Loss Superior RNFL Loss
FD-OCT can scan more points in less time sampling greater area with less motion error RTVue FD-OCT Optic Nerve Head Map (ONH) 9510 a-scans 0.39 sec
New advances from AIG: translating higher speed into better diagnostic accuracy
Glaucoma affects 3 areas in the posterior segment of the eye Cupping Nerve fiber thinning Ganglion cell loss
Ganglion cell layer contains the cell bodies of the retinal nerve fibers In the macula, the ganglion cells take up greater volume than the nerve fibers.
Macular retinal thinning is a relatively insensitive diagnostic parameter for glaucoma Guedes V, et al, Ophthalmology 2003 Bagga H GD, et al, Am J of Ophthalmol 2004 Medeiros FA, et al, Am J of Ophthalmol 2005
Macular retinal thickness on Stratus OCT Zeiss Stratus OCT 400 A-scan per-second 10 micron axial resolution Stratus FMTM 6 mm 768 a-scans, 2 sec } GCC Ishikawa H, et al., IOVS 2005 Tan O, et al., ARVO 2004; Ganglion cell complex thickness measurement by computer Tan O, et al., Ophthalmology, 2008;115:949-56.
Retinal thickness mapping is not sensitive for detecting glaucoma because glaucoma preferentially affects the inner retinal layers Ganglion Cell Complex (GCC) Ganglion cells: Axons = nerve fiber layer Body = ganglion cell layer Dendrites = inner plexiform layer Solution: automated computer measurement of GCC
The Ganglion Cell Complex (GCC) becomes thinner in glaucomatous eyes Normal GCC Glaucoma with thinner GCC GCC
Glaucoma: Macular Ganglion Cell Mapping RTVue FD-OCT, 26,000 A-scan per-second 5 micron axial resolution Ganglion Cell Complex (GCC) 7 mm scan area 14,944 a-scans, 0.58 sec NFL GCL IPL }GCC }Retina GCC = Ganglion Cell Complex mgcc thickness map micron
GCC Deviation Map % loss = actual scan value normal value normal value color coded map Percent loss value at each pixel location relative to normal based on age-adjusted normative database of over 300 healthy eyes Blue = thinning 20-30% relative to normal Black = 50% loss or greater Tan O, et al., Ophthalmology 2009; in press
GCC Significance Map color coded map shows regions where the change from normal reaches statistical significance Green = values within normal range (p-value 5% to 95%) Yellow = borderline results (p-value < 5%) Red = outside normal limits (p-value <1%) Tan O, et al., Ophthalmology 2009; in press
Normal Pre-Perimetric glaucoma (PPG) Perimetric Glaucoma (PG) # eyes 125 76 112 # subjects 65 52 79 Age (Yrs) 53±9 60±10 61±8 Tan O, et al., Ophthalmology 2009; in press
FD-OCT provided improved diagnostic accuracy for assessment of macular ganglion cells and nerve fiber layer loss in glaucoma OCT system Thickness Parameter (µm) AROC Group N vs PG N vs PPG ppnfl-avg 0.93 0.85 RTVue FD-OCT mgcc-avg mgcc-glv 0.90 0.91 0.78 0.79 mgcc-flv 0.92 0.73 Stratus TD-OCT cpnfl-avg mr-avg 0.92 0.85 0.80 0.76 Tan O, et al., Ophthalmology 2009; in press P=0.01
High-speed FD-OCT allows correlation of glaucoma disease patterns Pre-Perimetric Glaucoma Optic Nerve RNFL scan Macular GCC Significance Map N T Optic Nerve Photo Pattern Deviation Visual field
High-speed FD-OCT allows correlation of glaucoma disease patterns Perimetric Glaucoma Optic Nerve RNFL scan Macular GCC Significance Map Optic Nerve Photo Visual field
FD-OCT improved the repeatability of macular ganglion cell measurement and the potential to track glaucoma over time OCT system Thickness Parameter Standard Deviation of Repeat Measurements Group N PPG PG RTVue FD-OCT mgcc-avg (μm) 1.03 1.06 0.98 mr-avg (μm) 1.20 1.06 1.36 Stratus TD-OCT mr-avg (μm) 2.17 3.69 3.01 Tan O, et al., Ophthalmology 2009; in press
FD-OCT improved the repeatability of macular ganglion cell complex compared to TD-OCT circumpapillary nerve fiber layer measurements, thus improving the potential to track glaucoma over time OCT system Thickness Parameter CV (%) Group N PPG PG RTVue FD-OCT mgcc-avg 1.09 1.23 1.25 Stratus TD-OCT cpnfl-avg 1.72 1.75 2.86 Tan O, et al., Ophthalmology 2009; in press
GCC Progression Analysis
Fourier-domain OCT provides greater diagnostic power and repeatability for glaucoma Higher performance Better diagnosis Time-domain OCT Fourier-domain OCT
New Frontier Functional OCT: Doppler Measurement of Retinal Blood Flow
Current techniques do not allow practical measurement of total retinal blood flow Fluorescein Angiography Laser doppler flowmeter Ultrasound
Retinal blood flow measurement with Doppler OCT may help us understand the role of perfusion in the causation and treatment of glaucoma,other optic neuropathies and retinal diseases
Measurement of both Doppler shift and incidence angle are needed to compute flow in a vessel v + Δv v Probe beam Δv = 2V cos( α ) / λ α V
Measuring Total Flow Velocity Dual Beam Dave DP, Milner TE, Opt Lett 2000;25:1523 Pedersen CJ, et al. Opt Lett 2007;32:506-8 *Werkmeister RM et al. Opt Lett 2008;33:2967 Dual Plane 3D *Wehbe, H.M., et al. Opt. Express 15, 15193-15206 (2007) Michaely, R., et al. J. Biomed Optics, 12, 041213-1~7 (2007) Makita, S., Fabritius, T., and Yasuno, Y. Opt. Lett. 33, 836-838 (2008)
Doppler FD-OCT: Initial Experimental System Axial resolution: 6.1 μm A-scan rate: 17 khz Lateral resolution: ~20 μm Sensitivity: 110dB Integration Time: 50 μs FD-OCT system provided by Duke University (Izatt) and Bioptigen Inc. under contract to Advanced Imaging for Glaucoma grant Maximum detectable speed: 2.8 mm/s Minimum detectable speed: 16.3 μm/s
Flow direction relative to OCT beam is measured by 2 parallel cross-sections Double circular scan Flow profile and direction determined on parallel sections* *Y. Wang, B. Bower, J. Izatt, O. Tan, D. Huang, In vivo total retinal blood flow measurement by Fourier-domain Doppler optical coherence tomography, Journal of Biomedical Optics 2007;12:041215-22.
Double circular scan transects all retinal branch vessels 4 times per second Inner Circle Outer Circle
Total retinal blood flow is measured by summing flow in all branch veins Wang Y, Bower BA, Izatt JA, Tan O, Huang D. Retinal blood flow measurement by circumpapillary Fourier domain Doppler optical coherence tomography. J Biomed Opt 13, 064003, 2008 r 1 =1.7 mm r 2 =1.9 mm
Retinal perfusion is integrated over 2 seconds speed (m m /s) 7 6 5 4 3 2 No.2 No. 1 16 17 1 2 3 4 5 6 7 8 9 1 15 10 0 0 0.5 1 1.5 2 Time (second) 14 13 12 11 Hear beat Rate 70/min Reproducibility = 10% pooled coefficient of variation Wang Y, et al., Br J Ophthalm 2009;93:634-637.
Glaucoma reduces retinal blood flow Group Number of Subjects Total retinal blood flow (μl/min) Average Range Normal 8 45.64 40.73-52.91 Perimetric Glaucoma 10 33.54 23.6-40.88 p < 0.0003 Wang, Y., Tan, O., Huang, D. SPIE Proceeding, 7168, (2009)
Visual field pattern standard deviation was highly correlated with retinal blood flow 20 R 2 =0.77 p=0.0009 PSD 10 0-10 15 20 25 30 35 40 45 50 Total Blood Flow (μl/min) Wang, Y., Tan, O., Huang, D. SPIE Proceeding, 7168, (2009)
Superior Branch Vein Occlusion & PDR Amani Fawzi, MD Venous Flow Case Normal mean±sd (Range) Superior (μl/min) 8.2 23.5 ± 3.0 (19.2-27.5) Inferior (μl/min) 12.1 22.2 ± 2.6 (19.6-27.4)
Amani Fawzi, MD Diabetic Retinopathy Venous Flow Diabetic NR (2 eyes) DR (5 eyes) Normal mean±sd (Range) Superior (μl/min) 18.9 ± 4.3 (15.9-21.9) 8.2 ± 6.9 (4.6-20.4) 23.5 ± 3.0 (19.2-27.5) Inferior (μl/min) 23.0 ± 2.3 (21.4-24.6) 8.6 ± 3.6 (3.7-12.8) 22.2 ± 2.6 (19.6-27.4)
Non-arteritic anterior ischemic optic neuropathy Venous Flow Superior (μl/min) Inferior (μl/min) Case Normal mean±sd (Range) 12.99 23.5 ± 3.0 (19.2-27.5) 19.47 22.2 ± 2.6 (19.6-27.4) Alfredo Sadun, MD, PhD
The dual circular scan OCT Doppler retinal blood flow measurement technology has been licensed by USC to Optovue for further development Optovue RTVue 26 khz, 6 circles/sec
FD-OCT provides more information than other advanced imaging technologies FD-OCT SLT (HRT) SLP (GDx) ppnfl thickness Macular GCC + + + Disc & Cup + + Total retinal blood flow * Angle + Cornea + *Under development, not yet released
FD-OCT may have a growing role in glaucoma diagnosis
Acknowledgements www.aigstudy.net NIH/NEI Grant R01 EY013516
www.coollab.net David Huang, MD, PhD Ou Tan, PhD Yimin Wang, PhD Doheny Eye Institute Maolong Tang, PhD Yan Li, PhD Xinbo Zhang, PhD Sylvia Ramos, COA Jose Luiz B. Ramos, MD Camila Salaroli, MD Julie Schallhorn, MS Timothy Hsia, MS