Ex vivo confocal microscopy: an emerging technique in dermatology

This review aims to give an overview of the current available applications of ex vivo confocal microscopy (EVCM) in dermatology. EVCM is a relatively new imaging technique that allows microscopic examination of freshly excised unfixed tissue. It enables a rapid examination of the skin sample directly in the surgery room and thus represents an alternative to the intraoperative micrographic control of the surgical margins of cutaneous tumors by standard microscopic examination on cryopreserved sections during Mohs surgery. Although this technique has mainly been developed for the margin’s control of basal cell carcinoma, many other skin tumors have been studied, including melanoma. Use of EVCM is continuing to evolve, and many possible applications are under investigation, such as the study of nails and hair diseases and the diagnosis of skin infections.


Ex vivo Confocal Microscopy Examination Microscope
Only one device is commercially available (VivaScope 2500(r), new version of VivaScope 2000, produced by Caliber, New York, USA and distributed in Europe by MAVIG GmbH, Munich, Germany). The ex vivo confocal microscope works in reflectance mode (laser wavelength of 830 nm) or fluorescence mode (laser wavelength of 488 nm or 658 nm).

Processing
In reflectance mode no staining is required. However, aluminum chloride, acetic acid, or citric acid can be used to enhance the reflectance of nuclei [2][3][4]. When using the fluorescence mode, the entire surgical specimen is dipped in a solution of a fluorescent agent (e.g., acridine orange, methylene blue, fluorescein, Nile blue, or Patent Blue V) for 10-20 seconds and rinsed in physiological saline, acetic acid or phosphate buffered saline in order to remove the excess of the fluorescent agent. The most used fluorescent agent is acridine orange that targets nuclear DNA [2]. Several studies showed that acridine orange immersion does not affect subsequent frozen sections and formalin-fixed histopathology quality [5][6][7].
Cutting in thin slices is not necessary. The skin specimen can be observed entirely, or the tumor's margins can be vertically cut as for conventional histology [8].
In order to prevent sample deformation and movement during the examination, the skin sample should be mounted between two thick microscopy glass slides attached together by a small amount of silicon glue or modeling clay [9]. To facilitate the contact of the outer surface of the specimen with the glass slide facing the microscope objective, aqueous gel (e.g., Carbomer gel, Gel Larmes Thea Laboratories, Clermont-Ferrand, France) or silicone can be applied [9]. A "tissue press" can be used to better flatten the sample [9] and avoid to artifacts that result from the inhomogeneous contact between the sample's surface and the glass slide in the case of a specimen with a inhomogeneous thickness. The skin sample is then settled on the microscope stage.

Ex vivo Confocal Microscopy Images
The microscope produces horizontal images of 750 x 750 µm of the different layers of the skin up to a thickness of 200 µm. Single images are automatically stitched together into a reconstructed mosaic image to a maximum size of 20 x 20 mm (12 x12 mm for the previous generation of the device, the one that is most utilized). The depth of observation is manually adjusted. The acquisition time for a single image of 750 x 750 µm is 0.68 seconds (2 minutes/cm 2 ) for the third generation of the device [8]. According to the manufacturer's data, the lateral and axial spatial resolutions are 1 µm and 3 µm respectively.    The overall sensitivity and specificity of fluorescencemode EVCM for detecting BCC with narrow or incomplete margins were 88-96.6% and 89.2-99% respectively [4,7,10,27,30]. The largest study performed by Bennàssar and colleagues on 80 carcinomas demonstrated an overall sensitivity and specificity for residual BCC of 88% and 99% respectively [10].
Possible false positive results can be caused by the presence of hair follicles and eccrine glands or sebaceous glands that may be confused with BCC islands. However, the former reveals no palisading, less fluorescence, and smaller nuclei.
Furthermore, it could be difficult to distinguish the infiltrative cords of BCC from the surrounding peritumoral stroma, although the latter show no tendency to cluster.
A few case series reported the use of EVCM for controlling the surgical margins of SCC. Concerning the reflectance mode, an initial study found a positive correlation with histopathology in 13 out of 23 SCCs [11], whereas a sensitivity of 95% and a specificity of 96% for the identification of SCC were achieved in a second study on 10 lesions [37]. Only one study has been performed in the fluorescence mode, and eosin for cellular cytoplasm and dermis) and appears purple and pink [1]. In the future, digital staining could be used in order to change the grayscale into a color scale that mimics the appearance in histology [1].

Ex vivo Confocal Microscopy for the Control of Surgical Margins of Cutaneous Tumors
Different from MMS, EVCM enables the evaluation of the resection margins immediately after excision without freezing the specimen, thereby reducing the investment of time by two-thirds [10]. To date the analysis of surgical margins has been carried out by cutting the deep and lateral margins ("Tübingen torte" [26]) and analyzing them in vertical plane, perpendicular to the surface of the skin, as is done in conventional histology. Our group proposed an analysis of the entire specimen in horizontal plane, parallel to the surface of the skin ("en face" technique), without cutting the lateral and deep margins [8]. The "en face" technique provides images similar to in vivo reflectance confocal microscopy and is particularly suitable for horizontal spreading tumors [8]. (Figure 6).
Most of the EVCM studies were conducted on BCCs  BCC subtypes can also be identified [7]. In superficial BCC tumor islands are connected to the epidermis and separated from the dermis by some clefts (Figure 3

Basal Cell Carcinoma
In reflectance mode, large BCCs are easily detected [8].
However, tiny strands of micronodular infiltrating BCCs could remain hidden because of the hyper-reflectivity of the surrounding normal dermis [10].
In fluorescence mode with a contrast agent that stains nuclei, such as acridine orange, BCC islands are highly fluoresecent and are distinguishable from the surrounding tissue packed and irregularly distributed nuclei [11,37]. In situ SCC is more difficult to identify because of the difficulty of accurately detecting dyskeratotic cells and nuclear atypia [11,37].
Improved quality of images has been obtained by using acridine orange as fluorochrome [14,32]

Squamous Cell Carcinoma
Using the reflectance mode, squamous cell carcinomas (SCC) can be identified by the presence of keratinocytes with densely

Ex vivo Confocal Microscopy and Nails
EVCM could be particularly useful for nail tumors in order to confirm intraoperatively the diagnosis of the biopsy specimen and to proceed to the final excision without waiting for the histological examination. Notably, EVCM is more suitable than the optical examination of frozen sections of this site because nail tumors are often very small in size, and it is desirable that no material be wasted in producing sections that subsequently cannot be submitted to conventional pathology.
In a pilot study on six malignant epithelial tumors of the nail apparatus, Debarbieux et al [16] showed that EVCM in fluorescence mode could be a useful tool for the diagnosis
Proliferation of atypical melanocytes in the epidermis and consumption of the epidermis and nests of atypical melanocytes in the dermis were observed under EVCM in a study Bowen's disease showing marked nuclear and cytological atypia and the presence of numerous dyskeratotic cells. In particular, well-demarcated epithelial nests deeply invading the dermis, nuclear pleomorphism (variable size and shape of the nucleus), and densely packed and irregularly organized nuclei have been observed in invasive SCC and onycholemmal carcinoma [14,39,40]. Therefore, in these cases it would be possible to perform wide excision of the tumors just after the observation of a biopsy specimen under EVCM, shortening the management. However, in situ SCC and minimally invasive well-differentiated SCC were more difficult to diagnose, showing only focal epithelial acanthosis and not cytological atypia [14].
The same study of Debarbieux et al [14] included three benign epithelial tumors (two onychomatricomas and one onychopapilloma) that were differentiated from SCC because their cellular density was similar to the adjacent nail bed and because the cells had small monomorphic nuclei. Onychomatricomas showed acanthotic and papillomatous epithelium with no identified atypia under EVCM [14]. Interestingly, the dermal papillae embedded within the epithelial proliferation contained numerous spindle cells corresponding to fibroblastas [14]. Onychopapilloma is a very circumscribed epithelial tumor of the nail bed, characterized by thin digitiform epithelial projections within the upper dermis that were visible under EVCM [14]. In our experience, not only can nonpigmented epithelial tumors be recognized under EVCM, but also other amelanotic non-epithelial tumors such as glomus tumor and neurinoma [40].
EVCM seems to be less useful for subungual melanoma because the in vivo device that is used directly on the nail matrix performs well in this special site [15]. Moreover, with EVCM amelanotic melanocytes cannot be distinguished from epithelial cells or dermal inflammatory cells in fluorescence, and their reflectance can be low [14]. Debarbieux et al [41] also used the in vivo device for the ex vivo examination of nail biopsies of pigmented subungual melanoma in a series of eight cases. However, this procedure has limitations in that the specimen is not fixed, tends to move, and is difficult to orient, unlike when using the ex vivo device.

Ex vivo Confocal Microscopy and Infections
EVCM in reflectance mode has been used to identify the hyphae and/or conidia in onychomycosis [40], tinea capitis [16], and tinea barbae [16]. In reflectance mode hyphae are highly reflective, thick linear structures, and conidia are hyper-reflective roundish bodies with the same aspect of in vivo reflectance confocal microscopy [42][43]. In fluorescence mode with acridine orange, hyphae are bright [22]. Compared to fungal culture, EVCM is faster and does not have the problem of false positive results due to environmental con-