• Cornea
• infection
• imaging
• diagnostic tests/investigation

Infectious corneal ulceration is a serious ocular disorder that can result in severe visual disability. Suppurative keratitis and its complications constitute important causes of ocular morbidity. According to several epidemiological studies, it is estimated that nearly 1.5–8 million corneal ulcers occur each year in developing nations.1 Early diagnosis and institution of appropriate therapy are the key factors in preventing visual loss. The treatment of corneal ulcer is aimed at rapid eradication of infecting organisms. Identification of the infecting microorganism, therefore, plays a crucial role in management.

The following are various options available to us for aetiological diagnosis:

1. Clinical examination: A good history and thorough clinical examination looking for characteristic clinical features are important steps in arriving at aetiological diagnosis.2 3 However, characteristic clinical features may not manifest in every case of corneal ulcer. In an elegant study, Thomas et al4 looked at the value of characteristic clinical features as an aid to the aetiological diagnosis of suppurative keratitis and concluded that clinical features of microbial keratitis may vary considerably and that no clinical feature can be considered absolutely pathognomonic of a particular type of aetiological agent. Similar results were seen in another study published by Dahlgren et al.5 Both studies highlighted that clinical examination alone cannot be the basis for deciding how to treat suspected microbial keratitis.4 5 Rather, the microbiological identification of specific microbes can more reliably guide the individualised treatment of corneal ulcer.

2. Microbiology workup: This includes taking corneal scrapings using either a number-15 surgical blade or a Kimura spatula from the base and edges of the ulcer and using the scraped material for preparation of smears and direct inoculation on various culture media that facilitate growth of bacteria, fungi and parasites.6 A microbiology workup is the definitive method of identifying aetiological agents and ascertaining in vitro antimicrobial susceptibility. However, this requires maintaining or accessing a laboratory representing a certain cost in terms of time, finances and human resources. Parasites, fungi and some filamentous bacteria may take several days to weeks to grow in culture, thereby delaying definitive treatment. Corneal scrapings and microbiological examination yield positive results in only 52% to 65% cases.6–9 Further, because most of the antibacterials are bactericidal and attain high concentrations in cornea on topical administration, a large number of bacterial corneal ulcers respond to treatment with broad-spectrum therapy even without identification of the specific causative agent.8–10 Not surprisingly, most community ophthalmologists do not obtain cultures but instead treat these cases empirically with broad-spectrum antimicrobial therapy.7 11

3. Corneal biopsy: Yet another option especially in cases where corneal scrapings fail to provide a useful sample is excising a small piece of corneal tissue from the site of the infiltrate and subjecting it to histopathological examination using routine and specialised stains.12 Although this technique allows direct visualisation of organisms, it has the limitation of being less sensitive in cases of bacterial infection, involving tissue destruction and having a high sampling error.13

In addition to these limitations, corneal scrapings and biopsy are in vitro procedures and provide a static, two-dimensional picture of the disease; we cannot obtain information on the three dimensional extent of infiltration by the infecting organisms. Further, we cannot use these techniques to assess treatment response.

Confocal microscopy offers a method that addresses these disadvantages. It is a non-invasive and in vivo technique for the examination of the cornea wherein by focusing the observation and illumination system on a single point, any light reflected by elements outside the focal point is excluded thus increasing image resolution and contrast.14 Currently, two different models of confocal microscopes are in clinical use: (a) the slit-scanning confocal microscope represented by the Confoscan 4 (Nidek Technologies, Tokyo, Japan) and (b) the laser scanning confocal microscope represented by the Heidelberg Retina Tomograph Rostock Cornea Module (Heidelberg Engineering, Heidelberg, Germany). The confocal microscope can provide images of all corneal layers. The white light confocal microscope operates at a magnification of ×400 with a resolution of 1.5 to 6 μm along the x and y axes and 10 to 20 μm along the z axis.14 15 The resolution and overall image quality is better with the Heidelberg Retina Tomograph Rostock Cornea Module.16 Because of its ability to analyse corneal structures even in the presence of oedema, infiltrates, inflammation and scar, in vivo confocal microscopy (IVCM) is used to study cellular and structural details under normal and pathological conditions.14–18 Diagnosis of microbial keratitis represents one of the most important clinical applications of this modality.17 18 Various studies have clearly demonstrated its value in the diagnosis of Acanthamoeba,17–20 filamentous fungi17 18 21 22 and large filamentous bacteria.17 21 Acanthamoeba cysts are observed as spherical, round, ovoid or pear shaped hyper-reflective structures with a double-wall appearance measuring between 10 and 30 μm. IVCM may be able to provide images of trophozoites, but they seem to be more difficult to distinguish from other hyper-reflective structures. Radial keratoneuritis, a characteristic slit-lamp finding in Acanthamoeba keratitis, can also be seen on IVCM as irregular, thickened nerves.17–20 Filamentous fungi such as Aspergillus and Fusarium are seen as hyper-reflective, thin and branching structures 3–5 μm wide and of variable lengths with branching at a 45- to 90-degree angle.17 18 21 22 Nocardia, a genus of filamentous bacteria, are seen as multiple, thin (<1 μm), short, beaded filamentous structures that demonstrate right-angled branching.17 21 The American Academy of Ophthalmology technology assessment group reviewed available evidence for the use of confocal microscopy in the diagnosis of infectious keratitis and concluded that the available literature supports the use of IVCM as an adjunctive modality in the diagnosis of Acanthamoeba and fungal keratitis.18

It offers several advantages over conventional diagnostic modalities: (a) it is an in vivo technique in contrast to the in vitro nature of other diagnostic modalities; (b) it is a dynamic procedure, whereas others are static in nature; (c) it provides information on the presence of organisms and on their distribution in the cornea and (d) it can be used to monitor response of treatment. Further, it has a very high sensitivity, specificity, positive and negative predictive values, and acceptable inter-observer and intra-observer agreements.23 24

The procedure is, however, not without its limitations. IVCM is a contact diagnostic tool, and any movements of the highly sensitive eyes with infectious keratitis during the procedure can blur the images. To obtain high-resolution images, illumination needs to be optimal and requires adjustment on a case-by-case basis. Most bacteria and viruses cannot be visualised using this technology, and the current resolution is not sufficient to differentiate between bacteria, inflammatory cells and cellular debris.17 18 The double-walled nature of Acathamoeba cysts may not always be apparent and depends on the plane of the image and the obliqueness. In such a scenario, differentiation between cysts, leucocytes, lymphocytes and epithelial cells becomes extremely difficult.17 Further, the linear images produced by corneal nerves, scar tissue and activated fibroblasts can be confused with fungal filaments.17 18 21 22 Therefore, it is obvious that IVCM has a learning curve in terms of performing the procedure and interpretation of images. The article published in this issue by Hau et al very clearly highlights this important aspect of the procedure (see page 982).25 The authors found a twofold difference in sensitivity between the most and the least experienced observer, indicating higher diagnostic accuracy with clinicians experienced in confocal microscopy. This must be kept in mind while using IVCM as a stand-alone tool for the diagnosis of microbial keratitis. Probably a better approach will be to use it as an adjunctive tool along with clinical characteristics. Because IVCM is in vivo and non-invasive and can be repeated, it must also be evaluated for its use in assessing effectiveness of treatment.

In summary, accurate and rapid diagnosis is crucial for the management of infectious keratitis. Towards this goal, no single modality can be used for all cases. By combining clinical evaluation with microbiology or histopathology and confocal microscopy, we can make the best use of these modalities. At this point, confocal microscopy is a useful adjunct, but improvement in clinician training and experience, better standardisation of image interpretation, and higher resolution of images is likely to improve its diagnostic accuracy further. We must also explore its usefulness in assessing response to medical treatment and thereby monitoring therapy.