A Gentle Introduction to Endomicroscopy

Add One

Many diseases, including cancer, can be diagnosed by extracting tissue samples from the patient’s body and studying them under a microscope. This is called histopathology. An alternative technique called endomicroscopy has recently been developed. It allows tissue to be imaged at the microscopic level without removing it from the patient. This post introduces the topic of endomicroscopy – sometimes known as ‘confocal laser endomicroscopey’ or CLE – to the non-specialist.

Why do we need Endomicroscopy?

Histopathology is a vital part of modern medicine, but it’s certainly not without its drawbacks. Biopsies are very invasive – small sections of tissue are physically removed from the patient. There are practical limits to how many of these biopsies can be performed during a procedure, so only a few areas of tissue can be looked at. And because the sample has to be sent to a laboratory to be studied, the time taken to obtain the results can be considerable. So histolopathology cannot be used to guide interventions in ‘real time.’

Endomicroscopy sidesteps some of these limitations by giving us the option to perform ‘virtual’ histology. Instead of removing tissue from the patient through a physical biopsy, the idea is to study it in situ with an ‘optical biopsy’. The operator sees a live video feed from the endomicroscope and can view as many locations as he or she would like. And while we can’t say that this is truly non-invasive – after all, we still have to put the endomicroscope inside the patient – it’s certainly less invasive than biopsy.

How does it work?

Endomicroscopes have to be small enough to operate inside the human body. The most convenient way to use these devices is through the instrument channel of an endoscope; that way the operator can see a macroscopic camera view alongside the microscope image. So it’s the diameter of these instrument channels which actually determines the maximum size of the endomicroscopes, a figure which turns out to be around 3 mm.

In some ways, building a probe of this size isn’t too difficult. A particular type of lens, called a GRIN lens, can easily be fabricated with diameters down to a millimeter. Fibre optics can be used to send light into the patient and to and from the microscope tip, meaning that other components can sit outside of the patient and don’t need to be miniaturized. But there’s another problem.

In turns out that conventional microscopes are totally unsuitable for imaging tissue in the human body. This is because microscopes have a small depth of field, meaning that only a very narrow range of depths are in-focus at any one time. Light returning from all the other depths in the sample will out-of-focus and so appear blurred. This blurred image is super-imposed on the useful, in-focus image, making it difficult to pick out features. This is one of the reasons why biopsy samples are thinly sliced (‘sectioned’) before histopathology imaging: the other layers, which would generate this unwanted background blur, are physically removed.

Since we can’t do any slicing inside the patient, we have to remove this out-of-focus blur using optical methods. One type of microscope that can do this is called the laser scanning confocal microscope. It generates images containing only light from a thin layer of the tissue. Unfortunately, one of the features of a confocal microscope is that the image has to be built up, point-by-point, by scanning a laser beam across the sample. So we also have to find some way of scanning a laser inside the patient.

There are several ways of doing this. The first is to actually build a miniaturized scanning mechanism. Very small laser scanning mirrors can be built using a family of techniques known as MEMS – micro-electrical mechanical systems. Another option is to scan the tip of an optical fibre using piezo-electric translators. The disadvantage of both of these designs is that the probe is ‘active’. This means we have to send high voltages along wires into the patient, and perform extra engineering to ensure that there is no chance of an electric shock. And the more complicated the probe, the more expensive it becomes, especially if it has a limited number of uses.

An alternative approach is to use a fibre imaging bundle. The type of bundles used in endomicroscopy are essentially optical cables with a diameter of around a millimetre. They contain thousands of fibre optics cores, each only a few micrometres in diameter. Bundles can be used to transfer an image from one location to another, with each core acting like a pixel. So the scanning can be performed using bulky scanners outside of the patient, and transferred to the tissue through the imaging bundle. A recent post discusses the use of fibre bundles in microscopy in more detail.

What are the limitations?

The images produced by an endomicroscope may be useful, but they cannot really compete with those generated with histopathology. Histological imaging has a number of intrinsic advantages. The tissue sample can be stained with an array of dyes which make salient features more obvious. Although some fluorescence stains can be used with endomicroscopy, problems with delivery of the dyes as well as potential toxicity problems mean that this is necessarily much more limited. The other advantage of imaging biopsied tissues is that it can be cut and viewed in any orientation. Endomicroscopy is limited to imaging of the tissue surface and layers just beneath.

These are fundamental problems, but there are also some technical limitations with the current generation of endomicroscopes. They cannot compete with the bench-top microscopes used in histology in terms of sensitivity (seeing faint features), resolution (seeing small features) or field of view (seeing lots of features at the same time). They also tend to be quite difficult to use. In order to get stable images, the operator needs to maintain the position of the microscope to within a few tens of micrometers, while also ensuring a constant pressure is applied to the tissue. This is possible, but it is not necessarily easy.

For all these reasons, endomicroscopy is unlikely to replace histology on an appreciable scale any time in the near future.


3 Comments add one

  1. Graeme Mutton says:

    Mike Hughes FYI

    This is an abstract presented at DDW 2012 which shows the value of the Optiscan/Pentax endomicroscope from a multi centre trial.

    In Vivo Endoscope-Based Confocal Laser Endomicroscopy
    (eCLE) Improves Detection of Unlocalized Barrett’s Esophagus-
    Related Neoplasia Over High Resolution White Light Endoscopy:
    an International Multicenter Randomized Controlled Trial
    Marcia I. Canto*1, Sharmila Anandasabapathy2, William R. Brugge3,
    Gary W. Falk4, Kerry B. Dunbar5, Kevin E. Woods3,
    Jose Antonio N . Almario1, Ursula Schell6, Martin Goetz6, Ralf Kiesslich6
    1Medicine, Gastroenterology, Johns Hopkins Medical Institutions,
    Baltimore, MD; 2Gastroenterology, Mount Sinai School of Medicine,
    New York, NY; 3Gastroenterology, Massachusetts General Hospital,
    Boston, MA; 4Gastroenterology, University of Pennsylavnia,
    Philadelphia, PA; 5Gastroenterology, VA North Texas Health Care
    System, Dallas, TX; 6Gastroenterology, University of Mainz, Mainz,
    Confocal laser endomicroscopy (CLE) may increase the detection of
    endoscopically-inapparent Barrett’s esophagus (BE) high grade dysplasia and
    early adenocarcinoma (neoplasia) compared to standard endoscopy. However,
    the role of endoscope-based CLE (eCLE) for prediction of BE neoplasia
    compared to high resolution endoscopy (HRE) is unknown. Aim: To compare
    the diagnostic yield, performance characteristics, and clinical impact of
    HRE_eCLE followed by targeted biopsy (TB) with HRE alone with TB and
    random biopsy (RB) for detection of BE neoplasia. Methods: 5 centers
    participated in this international, multicenter double-blind randomized trial.
    Participating site investigators from the all sites had baseline eCLE experience of
    _ 25 cases and passed a test for interpretation of esophageal eCLE images.
    Patients with proven BE 1-10cm in length were stratified by indication for
    endoscopy (routine surveillance or suspected neoplasia) then randomized to
    either HRE alone (Group A) or HRE _ eCLE (Group B) using a 1:1 ratio. Patients
    with known localized neoplasia, esophageal lesion _ 2cm, prior ablation, or
    fluorescein-allergy were excluded. In Group A patients, HRE with TB of lesions
    followed by a standard RB protocol was performed. In Group B patients, HRE
    followed by fluorescein-aided eCLE with TB of endoscopic lesions and flat BE
    was performed. Before and after eCLE, a diagnosis was determined for each
    imaging site and a plan (do nothing, biopsy, or EMR) was recorded. After eCLE,
    EMR and/or RB protocol were performed. Biopsies were also taken for
    calculation of test characteristics during and after eCLE. Biopsies were interpreted
    blindly by 2 experienced pathologists. Results: Among 200 patients enrolled, 178
    were evaluable at time of analysis (median age 62.6 years, 75% men, mean BE
    length was 2.9cm). 44 patients had suspected neoplasia, 134 were undergoing
    routine surveillance (final neoplasia prevalence 23% and 3.7%, respectively). 3 of
    the endoscopists had performed _ 75 eCLE cases, 2 were less experienced. 46
    lesions in 33 patients were detected by HRE. Using a per biopsy analysis, the
    overall diagnostic yield for neoplasia (45.6% vs. 8.8%, p _ 0.0001) was 5.2-fold
    greater using eCLE, despite fewer biopsies obtained. The difference in diagnostic
    yield was primarily in patients with unlocalized neoplasia (59.5% vs. 12.7%, p
    _0.0001, 5.2-fold difference). Using a per patient analysis, HRE _ eCLE led to
    4-fold increase in diagnostic yield for BE neoplasia (79.3% vs. 20%, p_0.0001).
    In 26 Group B lesions, eCLE changed the HRE diagnosis to a correct one in 15
    (53.6%) and correctly changed the treatment plan in 9 (34.6%). CONCLUSION:
    Compared to HRE with RB, HRE_eCLE_TB improved detection of BE neoplasia
    with significantly fewer biopsies. eCLE led to greater sensitivity for diagnosis of
    neoplasia compared to HRE alone and impacts in vivo decision-making

    • Mike Hughes says:


      I hope my post didn’t sound too negative about endomicroscopy – I am most certainly an advocate of the technology, and I think it has a lot of potential. On the other hand, I also think it’s important to be realistic about what endomicroscopy can offer compared with standard histology.

  2. Graeme Mutton says:

    The process will initially be for targeted biopsy though Optiscan believe their latest system has resolution good enough for optical biopsy (see and treat). With health budgets under pressure that is where we are headed and targeted biopsy a cost saving step along the way. With telepatholgy in the future, I believe Endomicroscopy will provide confirmatory backup for Doctors that see and treat. You would have more appreciation than me on this matter.

Leave a Comment

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>