Ординатура / Офтальмология / Учебные материалы / Uveitis Text and Imaging Text and Imaging Text and Imaging 2009

co-operation for scan acquisition. It is difficult to perform, needs experienced hands. It is difficult to perform UBM in children and uncooperative patients.

KEY POINTS

1.UBM provides high resolution, cross sectional biomicroscopic images of the anterior structures of the eye.

2.It uses ultrasound of high frequency than conventional ultrasound.

3.This procedure is noncontact, non-invasive, quick to perform.

4.It does not require any pupillary dilatation and clear media to operate.

5.It helps in depicting the anatomic location of the normal as well as abnormal lesion.

6.Quantitative analysis helps us in follow up of the disease to document any worsening i.e. in cilliary body atrophy.

REFERENCES

1.Mundt GH Jr, Hughes WF Jr. Ultrasonics in ocular diagnosis. Am J Ophthalmol 1956;41:488-98.

2.Sherear MD, Foster FS. A 100 MHz PVDF ultrasound microscope with biological applications. Acoust Imaging 1988;16:511-3.

3.Sherar MD, Foster FS. The design and fabrication of high frequency poly(vinylidene fluoride)transducers. Ultrason imaging 1989;11;75-94.

4.Pavlin CJ, Sherar MD, Foster FS. Subsurface ultrasound microscopic imaging of the intact eye. Ophthalmology 1990;97:244-50.

5.Pavlin CJ, Harasiewicz K, Sherear MD, Foster FS, et al. Clinical use of ultrasound biomicroscopy. Ophthalmology 1991;98:287-95.

6.Tran VT, Lehoang P, Herbort C. Value of high-frequency ultrasound biomicroscopy in uveitis. Eye 2001;15:23-30.

7.Bhende M, Biswas J, Gopal L. Ultrasound biomicroscopy in the diagnosis and management of intraocular gnathostomiasis. Am J Ophthalmol 2005;140:140-2.

8.Bhende M, Biswas J, Sharma T, Chopra SK, Gopal L, Shroff CM. Ultrasound biomicroscopy in the diagnosis and management of pars planitis caused by caterpillar hairs. Am J Ophthalmol 2000;130:125-6.

9.Roters S, Szurman P, Engels BF, Bartz-Schmidt KU, Krieglstein GK. Ultrasound biomicroscopy in chronic ocular hypotony: its impact on diagnosis and management. Retina 2002;22:581-8.

10.Garcia-Feijoo J, Martin-Carbajo M, Benitez del Castillo JM, Garcia-Sanchez J. Ultrasound biomicroscopy in pars planitis. Am J Ophthalmol 1996;121:214-5.

11.Greiner KH, Kilmartin DJ, Forrester JV, Atta HR. Grading of pars planitis by ultrasound biomicroscopy—echo- graphic and clinical study. Eur J Ultrasound 2002;15:13944.

12.Ozdal PC, Mansour M, Deschenes J. Ultrasound biomicroscopy of pseudophakic eyes with chronic postoperative inflammation. J Cataract Refract Surg 2003;29:1185-91.

13.Tran VT, Lumbroso L, LeHoang P, Herbort CP. Ultrasound biomicroscopy in peripheral retinovitreal Toxocariasis. Am J Ophthalmol 1999;127:607-9.

14.Schneider C, Arnaud B, Schimtt-Bernard CF. Ocular toxocariasis. Value of local immunodiagnosis. J Fr Ophthalmol 2000;23:1016-9.

INTRODUCTION-HISTORY

Application of Optical Coherence Tomography (OCT), as an imaging modality for non-invasive cross-sectional imaging of the biological tissues was first reported in 1991.1 The technique was developed at the Massachusetts Institute of Technology and Humphary division of Carl Zeiss, Inc developed the commercial product. Because of the easy optical accessibility, it is especially suited for the diseases of the posterior segment of the eye.2-4 Swanson et al2 and Puliafito et al3 reported application in clinical practice for posterior segment of the eye. This technology essentially uses low-coherence interferometry to produce a twodimensional image of the optical scattering from various biological tissues. OCT gives a tomographic crosssectional image of the retina, RPE-choriocapillaris in the axial direction, that complements the standard topographic imaging techniques like fundus photography, fluorescein and indocyanine green angiography. This helps in interpretation of pathology in context of its anatomic location, substantiating the diagnosis, monitoring the course of the disease and evaluating the response to a therapeutic intervention.

PRINCIPLE

The principle of OCT is essentially similar to ultrasound B-scan, except light (low-coherence) is used instead of sound. The images are obtained by measuring the time delay of the light rays from different

microstructures in the eye. A beam splitter is placed in front of the light source that is emitting a continuous beam of low-coherence light. This beam splitter splits the beam into “Probe beam” and “Reference beam”. The probe beam is the one that is directed into the eye and is then reflected back from various tissues and interfaces with different optical properties.5 The time delay by various rays of light getting reflected at variable distances is then measured by applying the principle of low coherence interferometry.6,7 A continuous super-luminescent diode source is used to produce low-coherence probe light (830 nm). The reflected probe beam is now combined with the known reference beam. Matching the path lengths between the reference and probe creates an interference signal. The reference beam gives the distance information and the probe beam gives the structural information. Multiple tissue interfaces give rise to multiple peaks in the interference signal.8-10 Successive longitudinal scans are performed in the transverse direction and two scanning mirrors translate the beam laterally. Both axial and transverse data form a cross-sectional image with a transverse resolution of 13-microns and axial resolution of 8-micron. The colours that we see in the final image are false. Cold colours including blue and black indicate no/ reduced reflectivity, whereas hot colours like white, yellow and red indicate high reflectivity. An infrared camera takes the simultaneous red free fundus photograph and this photograph helps in the proper placement of probe beam during scan

Figure 1: Photograph showing Stratus OCT unit of Humphary /Carl Zeiss

acquisition. An inbuilt + 78 D lens helps in focusing on the retina.

THE STRATUS OCT MACHINE

The machine hardware essentially consists of a patient module with a chin rest, focus adjustment knob, joystick and a computer unit with a flat screen video monitor, keyboard, mouse and printer. The hardware mounts on a motorised table (Figure 1).

The software necessary to operate and analyse is pre-installed. The images and patient data can be archieved in digital versatile disk (DVD).

TECHNIQUE OF ACQUIRING A SCAN11

Switching on the system activates all the components and takes approximately 45 seconds to display ‘Start Window’ The menu and toolbar in the Main Window offers various options including Select Patient, Acquisition Protocol, and Analysis Protocol and so on (Figures 2A-C).

One can select the appropriate category and make data entry for new patient.

DATA ENTRY

For each new patient, the data including the name, gender, ethnicity, date of birth, and physician’s name is entered. Using drop down list to select a category enters the diagnosis (Figure 2D). For entering a new diagnosis to the list, use ‘Define Categories’ option from the list (Figure 2E). A patient in whom the scans are being repeated, the scanning is possible only when

Figures 2A-C: “Start Window” screen showing the various options that are displayed on starting the OCT machine

a patient has been selected from the database (Figure 2F).

PATIENT PREPARATION

It is preferable to dilate pupil before examination as small pupils of less than 3 mm may result in images that are truncated or are of poor quality due to lack of image intensity. The patient is asked to look into the

Figure 2D: Screen showing window for data entry including the name, gender, ethnicity, date of birth, and physician’s name

Figure 2E: ‘Define Categories’ option from the list : for entering a new diagnosis to the list

Figure 2G: Fast scan mode for optimising the alignment and polarisation and placement of the scan on the area of interest

Figure 2H: Selection of acquisition protocol

Figure 2F: Repeat scanning is possible only when a patient has been selected from the database

Figure 2I: Screen for review of last 8 scans for saving the select scan. The “review” option allows last eight scans to be reviewed, analysed and saved

‘Internal Fixation’ target in the ocular lens with the study eye. In patients with poor vision, the external fixation target can be used. The internal fixation, however, is the preferred method as it is more reproducible.

For internal fixation, the patient is asked to look inside the ocular lens. When patient looks into the ocular lens, he sees a rectangular field of red with a green light. He is asked to fixate at the green light. The location of internal fixation target can be readjusted as per requirements. The opposite eye of the patient can be covered as this helps the patient to fixate more steadily. With external fixation method, the patient has to use the fellow eye to fixate on the target that is external to the ocular lens.

The protocol for scan acquisition can be selected as per requirement. The Scan acquisition window gets activated by double click on the Scan button. The scanner, by default, activates in the fast scan mode, also known as scan alignment mode. This mode is useful for optimising the alignment and polarisation and placement of the scan on the area of interest (Figure 2G).

Once the alignment is satisfactory, one can click ‘Scan Mode’ button to change to Slow scan mode also known as scan acquisition mode. It is important to note that scanner must be in the slow scan mode to acquire scans. The desired scans can be reviewed, analysed and saved (Figures 2H and I).

The Stratus OCT assigns the signal strength to frozen/ saved scans and grades them from 0-10; with 0 being the worst and 10 the best. Signal strength of 5 or less is indicative of poor quality and these should not be used for quantitative analysis. Signal strength of 5 or more can be used for such analysis, provided the image brightness is uniform throughout. One should ideally aim for signal strength of 7 or more with uniform image brightness throughout. It is important to check for the message “Analysis Confidence Low” as it indicates unreliable information and warrants the scan to be repeated. The Stratus OCT also analyses the vertical placement and completeness of the scan image, which if not satisfactory, may result in display of messages like “Scan too high”, “Scan too low”, “Scan missing data” and describe what is seen on the screen. These messages appear on the Scan Acquisition window and should not be ignored. “Scan too high”, “Scan too low”, messages indicate the misplacement of the scan and the scan should be repeated using Z-

offset arrow buttons to raise or lower the scan image. “Scan missing data” indicates loss of data due to patient blink, and is unsuitable for analysis. The scan should be repeated after instructing the patient not to blink.

In addition, there is OCT image Tab that adjusts for the OCT image noise and range value. Noise is the level of signal in the background seen as blue or green speckles in the dark background. Range refers to the range of interferometer signal levels. Both noise and range are set by default, but can be adjusted if required. There is also Video and Lamp parameter Tab that enables one to adjust the brightness or contrast of the video image according to one’s preference.

The Stratus OCT offers 19 scan acquisition protocols designed for examination of the Retina or Glaucoma patients. The protocols that are helpful in uveitic diseases are as follows:

1.Line Scan

2.Radial Lines

3.Macular Thickness Map

4.Fast Macular Thickness Map

5.Raster Lines

6.Repeat

LINE SCAN

The line scan gives an option of acquiring multiple line scans without returning to main window. The default angle is 0° and the nasal position is defined as 0°. By default, the length of line scan is 5 mm (Figure 3A). The length of the line scan and the angle can be altered (Figure 3B), though one has to keep in mind that as the scan length increases, the resolution decreases. This protocol enables one to acquire one multiple scans of different parameters.

RADIAL LINES

This scan protocol consists of 6 to 24 equally spaced line scans that can be varied in size and parameters. All the lines pass through a central common axis. The default setting has 6 lines of 6 mm length. However, the length of these line scans can be changed by adjusting the size of the aiming circle. The change can be made only before saving the first scan. The radial lines are useful for acquiring Macular scan and retinal thickness/volume analysis (Figure 4).

MACULAR THICKNESS MAP

This is same as radial lines except that the aiming circle has a fixed diameter of 6 mm. It comprises of a series of 6-24 equally spaced lines through a common central axis. The number of lines can be adjusted before saving the first scan. This acquisition protocol helps in measuring the retinal thickness (Figure 5).

Figure 3A: Horizontal 5 mm OCT line scan passing through the foveal centre

Figure 3B: Horizontal 10 mm OCT line scan passing through the foveal centre

Figure 4: The radial lines for acquiring Macular scan and retinal thickness/volume analysis

FAST MACULAR THICKNESS MAP

This protocol is designed for use with retinal thickness analysis. When done in both the eyes, it can be used for comparative retinal thickness/ volume analysis. It is a quick protocol that takes only 1.92 seconds to acquire six scans of 6 mm length each. The size and number of scans is fixed in this protocol and cannot be altered (Figure 6).

RASTER LINES

This protocol provides an option of acquiring series of line scans that are parallel, equally spaced and are 6-24 in number. These multiple line scans are placed over a rectangular region, the area of which can be

Figure 5: ‘Macular Thickness Map’ acquisition protocol helps in measuring the retinal thickness

Figure 6: ‘Fast Macular Thickness Map’ protocol acquires six scans of 6 mm length each and helps in measuring the retinal thickness

adjusted so as to cover the entire area of pathology. This is especially useful in conditions like choroidal neovascular membranes where one wishes to obtain scans at multiple levels. The default setting has 3 mm square with 6 lines. The scan series proceeds from superior to inferior and the line scan from nasal to temporal (Figure 7).

REPEAT

Repeat protocol enables one to repeat any of the previously saved protocols using same set of parameters that include scan size, angle, placement of fixation LED and landmark. The system repeats all these parameters giving the same settings as in the

previous scan. This protocol is especially helpful when one is monitoring retinal changes. No parameter except placement can be changed.

The landmark can be placed on a point of reference. This helps in reproducibility during repeat scan. The previous image can be displayed for accurate placement of landmark. The landmark is a pulsating point of light that can be moved and placed on a reference point to enable reproducibility of scan pattern placement. In repeat scan, one has to click the ‘show old fundus image’ button and alternate between the old and current fundus video image.

While selecting a protocol, one has to keep in mind the kind of information one wishes to obtain in a given patient. The analysis protocols can be either image processing or quantitative analysis protocols. The image processing protocols can be used with any scan type while the quantitative analysis protocols can be used with certain scan type only. To get the most accurate and meaningful information, one needs to apply an appropriate protocol. Given below is the list of Quantitative Analysis protocols that correlate well with the scan acquisition protocols:

ANALYSIS/INTERPRETATION11

There are two ways of interpreting the OCT scan— Objective and Subjective. For the accurate interpretation of the image, one needs to combine both these modalities.

OBJECTIVE

We are all familiar with the interpretation of fluorescein angiogram where we categorise the pathology as hypoor hyper-fluorescent, or ultrasound B scan where images are either hypo or hyper echoic. Likewise in OCT scan; we look at the reflectivity pattern of the scanned images. The best way, in our experience, to do this is to select the scan group, select the appropriate analysis protocol and go to “Scan selection”. This gives a magnified view of the selected

image for objective assessment. One can modify the image before studying. This basically involves studying the morphology and reflectivity patterns of the acquired scan. It is important to be familiar with the normal OCT scan. For this, on a 10 mm horizontal line scan passing through the foveal centre one can clearly demarcate two major landmarks namely, optic disc and fovea. The optic disc is seen towards the right of the tomogram and is easily identifiable by its contour. The central depression represents the optic head cup and the stalk continuing behind is the anterior part of optic nerve. The fovea is easily identifiable by the characteristic thinning of retinal layers. The vitreous anterior to the retina is non-reflective and is seen as a dark space. The interface between the nonreflective vitreous and backscattering retinal layers is

the vitreo-retinal interface. The retinal nerve-fibre layer is highly reflective and increases in thickness towards the optic nerve. A hyper-reflective layer that represents Retinal pigment epithelium (RPE) and choriocapillaris marks the posterior boundary of the retina. RPE-choriocapillaris are seen together as a highly backscattering single band. The choroid and sclera are not seen well on tomograms as the signal attenuates by the time it reaches these structures. Just anterior to RPE-choriocapillaries complex is a minimally reflective layer that represents photoreceptors. Above this layer of photoreceptors are alternating layers of moderate and low reflectivity that represent different layers of neurosensory retina. (Figure 8). The retinal blood vessels within the neurosensory retina show backscatter and also cast a shadow behind.

Figure 8: OCT Line scan through fovea showing different retinal layers

Figure 9: The hard exudates in retinal layers seen as hyperreflective shadows that completely block the reflections from the underlying retina

Figure 10: OCT Line scan showing blood causing increased scattering

Figure 11: OCT Line scan showing hyper-reflectivity from the Congenital Toxoplasma scar

Following lesions are hyper-reflective.3

Hard Exudates

The hard exudates are seen as hyper-reflective shadows in the neurosensory retina that completely block the reflections from the underlying retina (Figure 9).

Blood

Blood causes increased scattering. In cases of small and thin haemorrhage, hyper-reflectivity is seen, whereas if the haemorrhage is thick, it might block the reflections from the underlying structures (Figure 10).

Scars

All the fibrotic lesions including disciform scars, choroidal rupture scars, healed choroiditis, etc. are hyper-reflective (Figure 11).

Following lesions are hypo-reflective

Serous Fluid

Retinal oedema is the commonest cause of reduced backscattering and one can localise the site of fluid accumulation. The serous fluid that is devoid of any particulate matter produces an optically empty space with no backscattering (Figure 12).

Figure 12: OCT Line scan showing the serous fluid accumulation as an optically empty space in a patient with central serous chorioretinopathy

Figure 13: Single eye retinal thickness analyses one scan group and the output graph depicts retinal thickness in a black line

Hypo-pigmented lesions of RPE

That might represent alteration in the cellular structure.

One must remember that poor quality scans due to opaque media and refractive errors might be falsely interpreted as hypo-reflective. However, hyporeflectivity in these situations is diffuse resulting in overall attenuation of the scan.

SUBJECTIVE /QUANTITATIVE

Retinal Thickness (Single Eye) (Figure 13)

This analyses one scan group and the output graph depicts retinal thickness in a black line. When we move the scroll bar on the fundus image, the corresponding thickness at that A-scan location appears on the graph. Use of calipers gives the reading in microns. This protocol, when applied to ‘Fast Macular Thickness Scan’ also shows the normative data colour code in each A-scan location.

Normative Data

Normative database is applicable to Fast Macular Thickness Map Scan protocol. Normative data is agematched to the patient. Normative data appears only for patients over 18 years of age and only for those patients whose records include their date of birth. This uses different colours including light red-light yellow-

Figure 14: ‘Retinal Map Single eye’ protocol obtains two maps of retinal thickness, one showing the colour code and other giving numerical values in nine map sectors

green-yellow-red colour code that indicate the normal distribution percentiles.

Retinal Map (Single Eye) (Figure 14)

This protocol obtains two maps of retinal thickness, one showing the colour code and other giving numerical values in nine map sectors. The diameter of the map by default is adjusted to 1, 3 or 6 mm centred on the macula. Click on 3.5 mm radio button changes the circle diameters of 1, 2.2 and 3.5 mm. This protocol, when applied to ‘Fast Macular Thickness Scan’ also shows the normative data colour code in each A-scan location.