16.1 OCT Principles, Scan Acquisition & Artifact Recognition

OCT Principles, Scan Acquisition & Artifact Recognition

Optical Coherence Tomography (OCT) has revolutionized vitreoretinal clinical practice by providing non-invasive, in vivo optical biopsies of the retina, retinal pigment epithelium (RPE), and choroid with near-histological axial resolution. As the front-line imaging specialists in retinal clinics, Certified Retina Technicians are responsible for operating advanced OCT systems, selecting appropriate scan protocols, optimizing image quality, and recognizing technical artifacts that can mimic pathology or compromise therapeutic decisions. A rigorous understanding of low-coherence interferometry, the operational differences between Time-Domain, Spectral-Domain, and Swept-Source platforms, and systematic artifact troubleshooting is fundamental to providing reliable diagnostic data.

---

Physical & Optical Principles of OCT

Unlike ophthalmic ultrasound, which measures the echo transit time of acoustic sound waves, OCT measures the echo time delay and backscatter intensity of light waves. Because light travels at approximately 3 × 10⁸ meters per second in air, direct electronic timing of reflected light across retinal micrometers is physically impossible. OCT overcomes this limitation by implementing low-coherence interferometry using a Michelson interferometer setup.

Low-Coherence Interferometry & The Michelson Interferometer

A low-coherence light source—typically a superluminescent diode (SLD) or a tunable laser—emits near-infrared broadband light. The incident light beam is directed toward a 50/50 beam splitter, which divides the light into two distinct optical pathways:

1. Sample Arm: The light beam is focused through the ocular optics (cornea, crystalline lens, vitreous) onto the patient's retinal tissue. Different microstructural layers (nerve fibers, plexiform synapses, photoreceptors, and RPE) reflect and backscatter light according to their local refractive index and tissue micro-architecture.
2. Reference Arm: The light beam is directed down an external optical path of known, calibrated length to a reference mirror.
3. Interference & Recombination: Light reflected back from the reference mirror and backscattered from the retinal tissue recombines at the beam splitter and strikes an optical photodetector.

Because the light source has a low coherence length (broad spectral bandwidth), constructive and destructive interference fringes occur only when the optical path length of the sample arm matches the optical path length of the reference arm within the source coherence length (typically a few micrometers). By measuring these optical interference fringes, the system accurately calculates the exact depth (Z-axis) and optical reflectivity of each microscopic retinal interface.

A-Scans, B-Scans & 3D Volumetric Cubes

• A-Scan (Axial Scan): A single one-dimensional reflectivity profile measured along the optical axis at a specific retinal point, recording tissue depth versus reflectivity.
• B-Scan (Cross-Sectional Image): Formed by rapidly sweeping the incident laser beam laterally across the retina (X-axis or Y-axis) while acquiring hundreds or thousands of adjacent A-scans, assembling a two-dimensional cross-sectional greyscale or pseudocolor image.
• 3D Volumetric Cube: Formed by acquiring a dense raster series of parallel B-scans across an entire macular or peripapillary region (e.g., 6×6 mm grid), permitting volumetric reconstruction, topographic thickness mapping, and coronal (en face) slicing.

---

Evolution of OCT Platforms: TD-OCT, SD-OCT & SS-OCT

Retinal optical coherence tomography has evolved through three major technological generations, each characterized by substantial increases in acquisition speed, axial resolution, and tissue penetration depth.

1. Time-Domain OCT (TD-OCT)

Introduced clinically in the late 1990s (exemplified by the Stratus OCT), TD-OCT relied on a mechanically moving reference mirror:

• Operating Principle: To measure reflectivity at varying retinal depths, the reference mirror was physically moved forward and backward along the optical axis. Interference occurred only when the mirror's mechanical position matched a specific reflective retinal layer.
• Acquisition Speed: Limited by physical mirror inertia to approximately 400 A-scans per second.
• Axial Resolution: Approximately 10 µm.
• Clinical Limitations: Due to slow acquisition speeds, volume acquisitions were prone to severe motion artifacts from normal microsaccades and blinking. TD-OCT scans were restricted to sparse radial line patterns, and full-thickness 3D retinal mapping was impossible.

2. Spectral-Domain / Fourier-Domain OCT (SD-OCT)

Developed in the mid-2000s (exemplified by the Zeiss Cirrus HD-OCT, Heidelberg Spectralis, and Topcon Maestro), SD-OCT revolutionized retinal diagnostics by eliminating mechanical mirror movement:

• Operating Principle: The reference mirror remains completely stationary. The light source is a broadband superluminescent diode (SLD) operating at a central wavelength of 840 nm to 870 nm. When recombined with backscattered sample light, the resulting interference spectrum contains all depth information simultaneously. This combined light is dispersed by a diffraction grating into its individual spectral wavelengths and projected onto a high-speed line-scan CCD or CMOS spectrometer.
• Fourier Transformation: A mathematical inverse Fourier transform is applied to the acquired spectral interference pattern, instantly resolving the entire axial depth profile (A-scan) in a single exposure.
• Acquisition Speed: 20,000 to 85,000 A-scans per second (50 to 200 times faster than TD-OCT).
• Axial Resolution: 3 to 5 µm, allowing crisp distinction of photoreceptor layers (ellipsoid zone, external limiting membrane) and individual retinal sub-bands.
• Signal-to-Noise Ratio (SNR): SD-OCT provides a 20 to 30 dB improvement in sensitivity over TD-OCT, enabling dense 3D volume cube captures, real-time eye tracking, and multiple-frame noise averaging.

3. Swept-Source OCT (SS-OCT)

Swept-Source OCT represents the latest frontier in cross-sectional posterior segment imaging (exemplified by the Topcon DRI OCT Triton and Zeiss Plex Elite 9000):

• Operating Principle: Rather than using a broad-spectrum diode and a dispersive spectrometer, SS-OCT employs a short-cavity frequency-swept tunable laser that rapidly sweeps through a narrow band of near-infrared wavelengths over time. Recombined light is detected by a high-speed dual-balanced photodetector without requiring a camera spectrometer.
• Wavelength: Operates at a longer central wavelength of 1,050 nm (compared to 840 nm in SD-OCT).
• Acquisition Speed: 100,000 to 200,000+ A-scans per second, dramatically reducing acquisition time and motion artifacts.
• Axial Resolution: Approximately 5 µm.
• Deep Penetration Advantages: The 1,050 nm wavelength exhibits significantly less optical scattering in media opacities (such as nuclear sclerotic cataracts, corneal haze, and vitreous hemorrhages) and reduced absorption by RPE melanin. As a result, SS-OCT achieves superior, uniform visualization from the vitreous down through the full thickness of the choroid and into the retrobulbar sclera without signal roll-off.

---

Standard Clinical Scan Protocols & Geometries

Use the ordered scan protocol and current clinic workflow, adapting acquisition technique within delegation and training to the patient's fixation capability and the stated diagnostic objective:

1. High-Definition Line Scan (HD Raster / Cross-Hair):
   • Geometry: A single high-density horizontal or vertical line scan (e.g., 6 mm or 9 mm line consisting of 1,024 to 4,096 A-scans).
   • Averaging: Utilizes high-level frame averaging (Automatic Real-Time / ART mode averaging 9 to 100 frames) to eliminate speckle noise and deliver publication-grade resolution of subtle foveal microstructures, such as the foveal pit, external limiting membrane, and ellipsoid zone.
2. Dense Volume Cube Scans (Macular 3D Grid):
   • Geometry: Parallel series of continuous B-scans over a rectangular grid (e.g., 512×128 or 200×200 lines covering 6×6 mm centered at the fovea).
   • Clinical Utility: Renders comprehensive topographical macular thickness maps, ETDRS 9-zone subfield thickness grids, and permits automated volumetric calculation of intraretinal or subretinal fluid.
3. Radial Scans (Starburst Pattern):
   • Geometry: A sequence of 6, 12, or 24 equidistant linear B-scans intersecting at a single central point (like spokes on a wheel).
   • Clinical Utility: Ideal for evaluating radially symmetric pathology such as full-thickness macular holes, vitelliform lesions, optic nerve head pit architecture, and peripapillary retinal nerve fiber layer (RNFL) insertion.
4. Enhanced Depth Imaging (EDI-OCT):
   • Pioneered by Spaide: In conventional SD-OCT, peak optical sensitivity occurs near the zero-delay line (conventionally placed near the vitreoretinal interface), with progressive sensitivity loss ("roll-off") in deeper layers.
   • Mechanism: EDI-OCT adjusts the physical working distance of the device or shifts the zero-delay line to position it at or beneath the level of the RPE and outer choroid. This inverts the sensitivity gradient, maximizing signal-to-noise ratio in the choriocapillaris, Haller's and Sattler's large choroidal vessels, and the suprachoroidal space.
   • Indications: Measurement of subfoveal choroidal thickness in pachychoroid disorders (central serous chorioretinopathy), extreme myopia, Vogt-Koyanagi-Harada (VKH) disease, and differentiation of choroidal nevi from melanomas.

---

Patient Positioning, Acquisition Technique & Active Eye Tracking

Capturing diagnostic-quality OCT scans requires meticulous attention to patient alignment, instrument settings, and scan optimization:

Patient Positioning & Optical Alignment Protocol

1. Ergonomic Positioning: Adjust the motorized instrument table and chin rest height so the patient sits comfortably upright without hyperextending or straining the neck. Align the patient's lateral canthus with the engraved canthus alignment marker on the chin rest support.
2. Pupillary Centration: Utilize the external infrared iris camera display to guide the joystick. Center the scanning pupil aperture directly within the patient's pupil. While SD-OCT and SS-OCT can image through non-dilated pupils, pharmacologic mydriasis (≥5 mm) is strongly recommended for dense volume scans to prevent pupillary edge vignetting.
3. Refractive Focus & Polarization: Rotate the focusing diopter wheel to correct for the patient's spherical equivalent refractive error until fine retinal details appear crisp on the live infrared fundus view. On systems equipped with polarization controls, adjust the polarization sliders to optimize the brightness and contrast of the interference signal.
4. Z-Offset Alignment: Adjust the working distance until the live B-scan profile is centered vertically within the acquisition frame, keeping the retinal profile well clear of both the upper zero-delay line and lower window borders.

Real-Time Active Eye Tracking (TruTrack)

Normal fixation is interrupted by involuntary physiological eye movements, including high-frequency micro-tremors, drifts, and abrupt microsaccades (occurring 1 to 3 times per second). Without compensation, microsaccades cause spatial misregistration, sheared vessel contours, and false thickness artifacts.

• Dual-Beam cSLO Tracking: Systems such as the Heidelberg Spectralis incorporate simultaneous confocal Scanning Laser Ophthalmoscopy (cSLO) tracking alongside the OCT beam (TruTrack Active Eye Tracking):
  • One beam continuously captures high-speed infrared fundus images (scanning at 870 nm) and identifies prominent anatomical landmarks (such as retinal vessel bifurcations).
  • If the patient blinks or makes a microsaccadic flick, the tracking system instantly pauses the OCT acquisition beam and repositions it precisely at the identical retinal coordinate before resuming capture.
  • Follow-Up (AutoRescan) Mode: Active tracking stores the exact Cartesian coordinates of baseline retinal scans, automatically locking onto the identical anatomical location during subsequent follow-up visits. This enables sub-micron longitudinal tracking of subtle retinal thickness changes and therapeutic responses to anti-VEGF therapy.

Image quality metrics

Signal or quality scores are platform-, software-, scan-, and protocol-specific. Use the current device instructions and clinic acceptance criteria rather than memorizing one cutoff. Inspect the B-scans, centration, clipping, motion, shadow, segmentation, registration, and repeatability even when a displayed score passes. When disease or media opacity prevents a nominal threshold, preserve the best attainable scan, document the limitation, and route it for clinician review instead of deleting potentially useful data.

---

Systematic Recognition & Troubleshooting of OCT Artifacts

Artifacts are common in clinical practice and can lead to misdiagnosis or inappropriate treatment if not recognized and resolved immediately by the technician:

1. Saccadic Motion & Banding Artifacts

• Appearance: Jagged, stepped, or sheared retinal contours on cross-sectional B-scans; "broken" or displaced retinal vessels and doubled vascular bifurcations on en face reconstruction maps.
• Etiology: Rapid patient eye movements (saccades) during volume acquisition.
• Troubleshooting: Activate real-time active eye tracking; ensure patient is fixating on the internal green fixation cross (switch to external fixation lamp if foveal vision is severely impaired); remind patient to hold steady.

2. Blink Artifacts

• Appearance: Distinct, sharp black vertical bands of complete signal extinction across the cross-sectional B-scan; wide horizontal dark gaps devoid of data across en face thickness topography.
• Etiology: Complete or partial eyelid closure during the scan sweep, physically blocking the laser beam.
• Troubleshooting: Instruct the patient to blink firmly several times right before scan activation, then hold both eyes wide open for 3 to 5 seconds. If necessary, have an assistant gently elevate the upper eyelid without pressing on the globe; re-acquire the scan.

3. Optical Shadowing Artifacts

• Appearance: Vertical columns of reduced signal or complete signal void located beneath optically dense structures, obscuring underlying outer retina and choroid.
• Etiology: Caused by dense pre-retinal opacities (vitreous floaters, asteroid hyalosis, vitreous hemorrhage), dense intraretinal lesions (dense superficial flame hemorrhages, hard lipid exudates, clumps of migrated RPE pigment), or normal large retinal blood vessels blocking incident light.
• Troubleshooting: For movable vitreous floaters, ask the patient to look up, look down, and re-center on the fixation cross to displace the floater away from the macular axis. Shadowing from fixed intraretinal hemorrhages is an unavoidable physical phenomenon and must be documented as an intrinsic lesion shadow.

4. Inverted / Mirror Artifacts (Zero-Delay Line Crossing)

• Appearance: The retinal image appears upside-down or folded across itself, with the vitreoretinal interface crossing through the outer retina.
• Etiology: In Fourier-Domain OCT, positive and negative optical path length differences yield mathematically identical interference frequencies, creating a mirrored duplicate image across the theoretical "zero-delay line." If the patient moves too close to the objective lens or the Z-offset is improperly positioned, the retina crosses this zero-delay plane, causing the image to wrap around and fold onto itself.
• Troubleshooting: Immediately adjust the Z-offset tracking control (or joystick working distance) to center the retinal profile safely within the middle vertical third of the acquisition window, well away from the top and bottom margins.

5. Cut-Off / Clipping Artifacts

• Appearance: The superior or inferior portions of the retinal cross-section are truncated or cut off flatly by the top or bottom edges of the scanning frame.
• Etiology: Incorrect Z-axis alignment, severe patient head tilting, or high myopia with a steep posterior staphyloma placing parts of the retina outside the axial detector window.
• Troubleshooting: Re-center the B-scan vertically using the Z-axis joystick; coach the patient to keep their head upright against the forehead band.

6. Automated Segmentation Errors

• Appearance: The automated software detection lines (e.g., green line for the Internal Limiting Membrane [ILM] and red line for the RPE/Bruch's membrane) misidentify the true anatomical borders, dipping into fluid cavities or skipping across elevated lesions.
• Etiology: Pathological disruptions—such as massive intraretinal cystoid edema, dense subretinal fluid, large pigment epithelial detachments, high myopia, or severe epiretinal membranes—alter normal reflectivity transitions, confusing the edge-detection algorithm.
• Clinical Impact: Causes wildly inaccurate macular thickness measurements, false "red disease" (pseudo-thickening) or false "green disease" (missed edema) on ETDRS sector heatmaps.
• Troubleshooting: The technician must review the raw segmentation boundaries on all B-scans across the volume. If automated lines fail, preserve the raw data, document the artifact, and notify the clinician that the numerical map may be invalid. Edit boundaries only when trained and authorized under the current device and clinic workflow.

---

Clinical Table: OCT Technologies, Scan Protocols & Troubleshooting Matrix