GO

4. Practical techniques of clinical confocal microscopy

4.1 Technical setup of the confocal microscope
4.2 Technique of biomicroscopy with the confocal microscope
4.2.1  Instrument adjustment and modification
4.2.2 Clinical examination procedure
4.3 Patient data archiving and retrieval
4.4 Dissemination, documentation, publication of microscopic findings

4. PRACTICAL TECHNIQUES OF CLINICAL CONFOCAL MICROSCOPY

4.1 Technical setup of the confocal microscope

When the ophthalmologist decides to use a confocal microscope, several points should be clarified before the microscope setting is designed. From our personal experience, we  recommend separate settings for in vitro examinations (donor corneas, organ culture) and in vivo examinations. This means that two different confocal microscopes will have to be installed,  the in vitro microscope preferably with a vertical  focusing setup;  while the investigations on the living eye are performed with a confocal microscope which incorporates a horizontal motion of the objective. Investigations on living animals, which must be in systemic anesthesia, can be performed with either setup, however, we prefer to use the in vitro version for this purpose.
The system used for the investigation of patients should be installed in a separate room, which is easily darkened and offers a quiet atmosphere without irritating noise or other distractions. Both, patient and investigator, have to concentrate to ensure proper location and coordination of ocular and microscope movements, enabling for the imaging of the selected corneal sites and tissue layers.

 When the microscope is going to be used frequently for the investigation of patients, a concept for the further storage and retrieval of patient's confocal images should be made beforehand. We found it useful to install, within reach of the microscope operating panel, a computer for digitizing and retrieving selected images with a connection to the department network (see below). Additionally, a slit lamp should be available in the same room to (i) analyze the corneal morphology according to the slit lamp image and to (ii) ensure the absence of ocular lesions (punctata, erosions, other lesions) secondary to the confocal microscopy session.

As corneal confocal microscopy is performed as a contact - method employing water immersion objectives which require a gel immersion optical contact to the cornea, appropriate prerequisites for removing gel remnants from the lens surface, efficiently disinfecting and drying the objective without causing damage to the optical surface have to be provided. In our setting, we routinely have at hand disinfection solutions, sterile cotton swabs, sterile BSS, and a wash basin available for the patient (and the doctor).

 The confocal microscopy procedure is performed under low ambient light conditions, where it is not possible to search for items like objectives, eye drops, etc. Special care must be taken to secure objectives when they are being exchanged to  prevent them from dropping. The microscope controls and electric switches should be clearly visible under these conditions, enabling for a secure and immediate operation by the investigator.

 The microscope stand including a headrest and horizontal handles the patient can hold on during the examination should be stable enough to function over a presumed patient weight / size  range from 100 to 300 pounds. In terms of stability, patients that can well control their voluntary movements may run into problems, when a chest / volume problem causes respiratory fluctuations of their corneal apex making even a short microscopy session impossible.

 It clearly is the responsibility of the investigator to ensure full function of the microscope controls and stability of the patient's support before an examination is begun. Failure to do so may result in damage to the patient's eye or in non - optimal picture quality and extended investigation times. Under this aspect, the microscope should be always operated from and returned to full backward  position. Beyond the immediate operational setup, sufficient space and an appropriate facility should supply space for patient related items like eye drops, optical contact eye gel holders, spare parts, video tapes and other requirements of the procedure.

 Confocal microscopy can also be used for the investigation of donor eyes. Due to the requirement of a contact mode for image acquisition, intact donor globes are more suitable for the investigation than excised donor discs. Intact donor tissue can be studied with the clinical type of confocal microscope, if a suitable eye holder is attached to the microscope stand's headrest. To study excised donor discs, a vertical setup with suitable facilities to immerse the objective in the culture medium and to fix the donor tissue in an appropriate position is preferable. Under these latter condition, the maintenance of sterility of the corneal storage is not always guaranteed.
 

4.2  Technique of Biomicroscopy with the Confocal Microscope

4.2.1  Instrument adjustment and modification

 The instrument used for our investigations is a real time flying slit confocal microscope as previously described in detail (Masters and Thaer 1994b, Böhnke and Thaer 1994).  The microscope is equipped with a 100 Watt halogen lamp for illumination and a slit scan-synchronized, high sensitivity video camera with adjustable black level suppression. In clinically normal corneas (fig.3), the halogen lamp will have to be set to full power to supply enough reflected light from the corneal structures. In pathological corneas with scar tissue or other highly reflective contents, the lamp power has to be reduced by about 50 percent. With selected filters, which can be inserted into the optical path of the microscope, the spectrum emitted from the halogen lamp can be confined to selected spectra, either improving optical penetration by selecting longer wavelengths or improving image contrast by blocking out longer wavelengths (the detailed discussion of these features is beyond the scope of this chapter). For studies using fluorescent dyes, a 300 W mercury lamp with an appropriate fluorescence excitation and emission filter set is supplied. If very high light levels are required (e.g. in the investigation of edematous ex vivo corneas), the mercury lamp can be used without the filter set. Depending on the experience of the investigator, the instrument should be readjusted periodically to give an optimum and homogeneous illumination of the optical section, a perfect alignment of the scanning slits, and the lowest possible levels of stray light which may degrade the image contrast. The projected confocal slit width, which can be adjusted in some instruments,  was selected to be 10 microns, which is a compromise between the best resolution, and the best illumination and contrast.

 For corneal imaging, a 25x/0.65 NA, a 40x /0.75 or a 50x /1.0 NA water immersion lens (Leitz, Germany) can be used. The resolution of the different objectives and the thickness of the optical section to some extent are influenced by the light levels used and the geometry and the reflectivity of the structures studied. As a practical guideline, the lateral resolution of the 25x objective can be assumed with 1.4 microns , whereas the thickness of the optical sections is about 16 microns. The lateral resolution of the 50x objective is better than 0.8 microns, the thickness of optical section is about 10 microns.

 To check for homogeneity of field illumination and also to practice coordination of the microscope movements, a piece of black paper is placed vertically in front of the microscope. The microscopy is equipped with a 10x  objective and advanced until the surface structure of the object is visible. The halogen lamp has to be dimmed for this procedure. With this overall weakly reflective object, the technical status of the microscope optics can be easily checked and, if required, also adjusted. By mounting the black paper not strictly en face before the objective, the investigator can practice to judge and establish perpendicularly by using extra drives tilting the microscope's frontal plane.

 To investigate the patient’s cornea, the instrument objective is brought from its most backward position into optical contact with the corneal tissue by a high viscosity acrylic ocular gel. The S-VHS tape recorder is started when an optimum centration has been achieved, as judged from a well centrated light reflex upon passing the epithelial or endothelial layer. The real-time en face sections from all layers of the cornea are then recorded on videotape. The position of the optical plane in the Z-axis is controlled with a manual micrometer drive. Bowman’s membrane and the corneal endothelium are used as additional reference structures for the Z-axis position.

 For a detailed analysis, the recorded video sequences are reviewed frame by frame. For the patients' record, selected frames can be printed with a videoprinter.  Pictures for publication can be directly photographed off the monitor. Alternatively, if the appropriate equipment is available, selected frames can be digitized and directly entered into an image file server to be available either on - line, or be exposed to photographic film with a laser film printer. We recommend that the examiner routinely digitize images from all corneal layers, plus specific findings for a given case.
 

4.2.2 Clinical examination procedure

 Patients are routinely evaluated by clinical slit lamp biomicroscopy before and after the confocal scanning procedure. Before the examination with the confocal microscope, the patients are informed about the nature of the confocal scanning procedure. The investigator should decide beforehand which objectives are going to be used. For a quick routine examination, the 40x objective offers the best overview type of characteristics. However, this objective will miss some details visible with the 50x objective and does not give a generous overview (and a better link to slit lamp morphology) like the 25x objective. In most of our patients, we try to work with all three objectives and select the one or two most appropriate to image the specific pathology in further follow up sessions.

 One drop of  acrylic eye gel is placed on the microscope objective and the instrument moved to a full backward position. After topical anesthesia with one drop of 0.4% Novesine or any other topical anesthetic, the patient’s head is positioned on an adjustable headrest. The confocal microscope with the 40x objective is then placed 1 - 1.5 mm above the apex of the corneal center. The patient is asked to look into the light, so that the optical center of the cornea is aligned with a lateral accuracy of probably less than 1 mm. Then, the microscope is brought into optical contact with the cornea by manual advancement of the micrometer - controlled Z-drive. From this point on,  all further x-y-z movements of the instrument are controlled from the real-time picture displayed on the monitor. The image is recorded on a S-VHS or DV-CAM tape recorder and can be reviewed in slow motion. Simultaneously, the observer comments on the video tape sound track on position and other findings to supplement the information recorded.  Care must be taken not to mechanically damage the anterior segment.
 For the manual imaging with the 40 x and the 50x objective setup , the following pattern is usually employed :

1. Establish optical contact with cornea, focus on basal epithelial layer, centrate  surface parallel sections, start video tape.
2. Move backwards to image superficial epithelial cell layers.
3. Proceed to basal cell layer.
4. Proceed to Bowman’s membrane, look for subepithelial nerve plexus.
5. When first stromal keratocytes are visible, return to (2).
6. Repeat (2)-(5) until enough frames of epithelium have been sampled.
7. From Bowman’s membrane proceed to endothelium at about  0.1 - 0.5  mm/sec with manual micrometer advancement.
8. When endothelium reflex just above the anterior chamber becomes visible,  recentrate the microscope if required, then return to keratocyte layer just  above Descemet’s' membrane.
9. Return to endothelium and anterior chamber, making sure that some frames  of the endothelium have been captured.
10. Repeat (8) and (9).
11. Return to epithelium, reverse movement as in (7).
12. Repeat (2)-(11) at least once, more times if required.
If required, the same procedure (1-12) can be carried out with the 50x and the 25x microscope objectives.
For every eye and objective, a minimum of 0.5 to 1 minute good quality tape recording should be collected. Extended recording times are usually due to a nontrained investigator, which lead to decreased patient comfort and later on to extended tape reviewing times. With sensitive or less cooperative patients, the recording time however may be extended to obtain at least 1000 useful optical sections from all corneal layers. Included in this number usually is a minimum of at least optical sections from all Z-positions of the central corneal stroma. If a specific finding is observed during the examination for example in the epithelium, extra recording time is spent on the region of interest.

 To investigate non - central locations on the cornea, the limbus, or even conjunctival areas, the patient’s direction of gaze is aided with a fixation light for the contralateral eye. For these special locations, the angle of the frontal plane has to be tilted to achieve perpendicularity to the surface of the tissue studied. Occasionally, a somewhat oblique section may also be interesting, as it gives more information on the thickness of some structures imaged in one optical section.

 As the scanning procedure for some purposes will benefit from a transient complete immobilization of the patient's eye, we have recently designed a corneal suction device to fixate to the limbal area with a rather low vacuum. The vacuum is adjusted to a level high enough to ensure immobilization of the eye for about one minute and low enough to allow the patient to break away from the headrest if he desires to do so. With a temporary mechanical fixation of the eye under investigation, an automated scan allowing for the standardized collection of data can be performed. In this procedure, the microscope is centered manually for perpendicularly over the field of interest, and then the focus is advanced manually through the stroma and behind Descemet's membrane to lie in the anterior chamber. Then, the Z-drive is coupled to a stepping  motor, which moves the focal plane backwards through the stroma at a preselected speed of 125 microns per second over a total distance of 1 mm . With a recording speed of 25 frames per second, the average thickness of consecutively imaged sections will be 5 microns. For more detailed information regarding selected layers and lateral locations, a supplementing manual imaging adjustment is usually required. This work is still under way to define the optimum setup conditions.
The video tapes should be evaluated in the slow motion or single frame mode of the video  recorder. From our patient examinations stored on SVHS or DV-CAM tapes, we usually digitize a standard set of frames (from 40x or 50x objective recordings) for every patient, which includes:

1. epithelial surface cells (see fig.4)
2. epithelial wing (intermediate) cells (see fig.5)
3. epithelial basal cells (see fig.6)
4. subepithelial nerve plexus / Bowman's layer (see fig. 7 and 8)
5. first keratocyte layer (see fig. 8 and 9)
6. anterior stroma keratocytes (10 - 100 m below / behind Bowman)  (see fig. 10)
7. intermediate stroma (100 to 350 m depth) (see fig. 10 and 16)
8. posterior stroma (350 to >500 m depth) (see fig.11)
9. most posterior keratocyte layers, just before Descemet's membrane (see fig.11 and 12)
10. endothelium (see fig.12 - 15)

In addition,  we digitize a variable number of frames, which show specific findings for a given case. The digitized files are given identifying numbers, which clearly link them to a record of the database carrying information about the identity of the patient, the location of the recording  and relevant technical information regarding the microscope (see below).  When video images are digitized, we do not apply image enhancement techniques except for a median filter function, which basically makes the image appear smoother and reduces pixel noise.  With the advent of digital video recording technology, the video images can now be stored in a digitized format with an unique identifier of every image on a given tape. Thus, the need to digitize selected frames within due time to prevent deterioration of image quality may be reduced, however, a critical review of a recorded sequence after the recording session is still mandatory.

 In contrast to the extended sessions in the early days of the procedure, a central corneal exam of all layers will now typically consist of 40-100 seconds recording time for every microscope objective used.  Considering patient and instrument handling,  the total exam time does not exceed 10 minutes.  It should be kept in mind, however, that the observer has to review the tape frame by frame later on,  and either make hard copies (photographs) directly from the monitor screen or digitize and store selected images. The total time for review, selection, processing and storage of the confocal images can be much more than one hour per patient. If, on the other hand, only one specific aspect has to be answered, the total investigation time can be considerably shorter.
 

4.3  Patient data archiving and retrieval

Confocal microscopy creates a vast amount of (image) information, which has to be kept organized and retrievable with respect to the patient investigated. The current recording technique on video tape requires a meticulous documentation of the identity of a given video sequence. Currently, we are storing the original recordings in addition to the selected and digitized images. These latter images represent a biased sample due to their nature of having been preselected for digitization. By also preserving the original videotapes, we keep the opportunity to always go back to the original material and perform additional evaluations.

 Confocal microscopy is associated with other techniques of corneal investigation. These include:

1. clinical slit lamp biomicroscopy, with video option mounted on the slit lamp
2. slit lamp photography with high magnification option
3. corneal pachymetry, preferably with non-contact techniques
4. corneal topography with current videokeratoskopes
5. corneal endothelial cell microscopy, preferably with non-contact techniques
6. optical low coherence interferometry techniques, which enable for a precise  measurement of corneal thickness (Hitzenberger et al 1992, Böhnke et al 1997, Böhnke et al 1998) and the imaging of sagittal reflectometric corneal sections (Koop et al. 1996).

 To organize all relevant information, we have installed a database for parallel use with the confocal microscope. Every video sequence is identified regarding its tape number, tape position, recording date, technical microscope data, investigator's name and comments, patient information, study number, and diagnostic group allocation. Additional information on other corneal findings from the same visit and comments on specific findings is stored with a free text description of the video sequence. The record also displays the file numbers of digitized and stored video frames.

Due to the tendency of analogue video - tapes to fade over time, backup copies have to be made from the tapes as well as from the digitized images. With an assumed total recording time of 30 hours per year, the corresponding number of digitized files may range up to 10 TIFF-files per minute of video tape. Consequently, a strategy for data handling and retrieval should be developed beforehand if a frequent use of the microscope in patients is planned. In our unit's setup, the TIFF files are stored on dedicated magnetic and optical hard drives, which are accessible via the department's computer net from anywhere.  With digital video tape recorders the image quality may be more stable over time.
 

4.4  Dissemination, documentation and publication of  microscopic  findings

 The confocal microscopic examination creates a sequence of video-images, which contains more information than a single printed frame can express. As the judgment of the observer - similar to histopathology - is more important than a number of hardcopies attached to the patient's file, we currently prefer to enter a comment on the result of the examination into the patients record.
Confocal microscopy creates three categories of data:

Morphological findings can be communicated as qualitative descriptions, weighing the observations in consideration of current knowledge from established techniques. For this purpose, hardcopies either directly  from the videotape or from digitized images can be used for illustration.

Morphometrical data can be generated from the confocal investigation by applying stereological and other quantitative techniques (Underwood, 1969) . These may include size of basal epithelial cells, patterns of corneal nerves, keratocyte densities in various layers, endothelial cell size, number of stromal inclusions like micro-dots, and stromal light backscatter in Z-scans with or without simultaneous lateral scan movements. For these data, numbers rather than figures can be communicated. One example is the quantification of stromal micro - dot deposits in contact - lens wearers (Böhnke and Masters, 1997), where we changed from a "0" to "4" semiquantitative grading to a numerical value between 0 and 150,000 per mm3 (Cadez et al.1998).

Three-dimensional and kinetic images. may be obtained with 3 -D reconstructions to be viewed with special technical setups. Selected video sequences may also be transposed to slow motion and used for teaching purposes as special video tapes.

GO