The use of reflectance confocal microscopy for monitoring response to therapy of skin malignancies

Summary Reflectance confocal microscopy (RCM) is a new non-invasive imaging technique that enables visualizing cells and structures in living skin in real-time with resolution close to that of histological analysis. RCM has been successfully implemented in the assessment of benign and malignant lesions. Most importantly, it also enables monitoring dynamic changes in the skin over time and in response to different therapies, e.g., imiquimod, photodynamic therapy, and others. Given the often traumatic nature of skin cancer that affects both the physiology and the psychology of the patients, it is crucial to have methods that enable monitoring the response to treatment but that minimize the distress and discomfort associated with such process. This article provides a very brief overview of the fundamentals of RCM and then focuses on its recent employment as a monitoring tool in skin cancer and other pathologies that may require frequent follow-up.


Introduction
Non-invasive techniques are the necessary future of diagnostic procedures. In dermatology, biopsy collection may be very invasive, producing scarring and malformation of the skin and having a deleterious aesthetic and psychological effect.
As imaging techniques have evolved recently, new possibilities have opened for the dermatology practitioner to gather information on the status of skin and external mucosal tissue in a non-invasive manner. The advantages of these new techniques include an increased precision and delimitation of the margins of a lesion, reduced distress for the patient, and the capability to perform repetitive analysis of the affected area. These techniques include dermoscopy [1], optical coherence tomography [2], high-frequency ultrasound [3], magnetic resonance imaging (MRI) [4], fluorescence-mode confocal microscopy [5], and reflectance-mode confocal microscopy (RCM) (see below). The major advantage of conventional, biopsy-based histology is its diagnostic value and its resolution power. However, non-invasive imaging is painless; it causes no tissue damage and its resolution is rapidly improving. Also, processing-based artifacts (due to fixation, sectioning, and mounting) are not an issue, since non-invasive imaging does not disrupt the native structure of the tissue. In addition, non-invasive in vivo imaging is less time-consuming (e.g., it obviates sample processing) than routine histology. Finally, the specialist can sample the same location of the skin over and over to collect a series of images. This improves the quality of the data and offers an invaluable advantage when treating diseases. Furthermore, it allows sequential evaluation of therapies that alter the architecture of the skin, including pre-surgical tumor evolution, post-surgical wound healing, evolution in response to nonsurgical therapy and determination of outcome (e.g., remission vs. relapse) [6][7][8][9][10]. Two major caveats stand out: one is that the skin is only a semi-translucent tissue, hence, lightbased applications meet with physical barriers that stop photons and complicate obtaining quality data from the lower layers of the skin. The second problem is that histology is a well-established technique with over 100 years of accumulated data and evolution in data collection, staining and quantification. In comparison, these non-invasive techniques are relatively young and findings obtained through them still need to be correlated with histology to ensure the correct diagnosis is achieved. This latter caveat is being overcome by recent efforts to endow the literature with a consensus terminology [11] and to publish atlases detailing observations made using these novel techniques and correlating them with conventional histology, e.g., the handbook of the use of RCM in dermatology [12]. These efforts offer multiple beneficial effects for the dermatology community: 1) they provide informative tools for dermatologists and imaging technicians; 2) they homogenize the nomenclature, which is an essential, often overlooked step for these techniques to gain widespread acceptance; and 3) they often help in clarifying diagnosis by correlating the findings in the clinic with published case reports. This review will focus on the employment of RCM in monitoring response to non-surgical skin treatments.

Principles and technology of reflectance confocal microscopy
Confocal microscopy is a relatively common tool used in basic sciences (cell biology, immunology or neurobiology), but its in vivo application has required major improvement in optics, light sources and system stabilization for the images to have enough quality to become useful. In 1995, confocal scanning laser microscopy was first reported to image human skin in vivo [13]. This opened the gates to its employment in a varied array of skin disorders and diseases.
Reflectance confocal microscopy collects the light reflected from specific structures (normally cells) present in small areas of the skin scanned with a low power laser.
This approach generates images of dark (non-reflecting) and bright (reflecting) structures in the skin, and represents them as thin sections of horizontal tissue in vivo.
RCM microscopes are, in fact, in vivo adaptations of confocal microscopes used in basic sciences. RCM microscopes use lasers as sources of illumination. The actual range of RCM varies from 800 to 1064 nm; and the lasers used in RCM are not very powerful (<30 mW) to avoid damaging the tissue during examination; the maximum depth of imaging being between 200 and 250 μm [6,14]. Another important component of RCM is the skin contact device, which is used to minimize spherical aberrations and movement artifacts. This is done through the use of a water-gel interface between the lens and the skin in a vacuum type device that encases the lens. A typical example of lens used in RCM is a 30x objective lens of numeric aperture (NA) of 0.9, which, together with a physical filter provides a lateral resolution of approximately 1µm and an axial resolution (section thickness) of 3-5 µm [14]. Finally, RCM allows for the collection of time-lapse photography to visualize dynamic events in the skin, e.g., cell migration or blood flow [6,8,10,13].

Applications of reflectance confocal microscopy for dynamic monitoring over time
The best-characterized application of RCM in the dermatologist's office is diagnosis, and a number of reviews have covered the topic recently [15][16][17][18]. This article, on the other hand, will focus on the employment of RCM in monitoring response to surgical and non-surgical skin treatments.
The gold standard for the treatment of several types of skin cancer is surgical removal (excisional biopsy) of the affected area and the adjoining healthy tissue to prevent the escape of tumor cells. However, this is often traumatic, particularly if the affected area is a cosmetic hotspot, e.g., the face. Excisional biopsy is often preceded by one or several incisional biopsies aimed at characterizing the tumor, determining malignancy, and helping the medical practitioner make an informed decision on the most appropriate treatment. Furthermore, excisional surgery needs to be followed up as well and may require additional incisional surgeries. With these problems in mind, the development of non-invasive treatment protocols has been important. Major breakthroughs are the uses of photodynamic therapy (PDT) and imiquimod application.

RCM monitoring of the response of BCC to PDT treatment (Figure 1)
PDT treatment has been very successful in clinical trials (reviewed in [19]). Different histological subtypes of BCC have been treated with PDT. For example, treatment of nodular BCC entails a higher risk of relapse than conventional sur-gical excision [20]. Conversely, PDT of superficial BCC has an efficacy comparable to that of surgery, but with a cosmetic outcome that is vastly improved [21]. Several studies have addressed the potential of RCM to follow up the response of BCC lesions to PDT using MAL [22]. Gorlin syndrome

Treatment of actinic keratosis with PDT in solid-organ transplanted patients: RCM follow-up
5-ALA PDT is rapidly becoming a reference treatment in the management of actinic keratosis (AK) [23]. In addition to close to 100% efficacy, the cosmetic outcome of the procedure est is the management of skin cancer, mainly SCC, of which AK is considered a direct precursor [24] in immunosuppressed individuals, e.g., those undergoing solid organ transplantation.
In this context, a recent study has evaluated, using RCM, the response to 5-ALA PDT in organ transplant recipients (OTR) that display AK. In this study, RCM proved to be an excellent tool to monitor the response to PDT [25]. This application of RCM is particularly important for immunocompromised individuals who have a high prevalence of AK. trideca-1(9),2(6),4,7,10,12-hexaen-7-amine) is an immunomodulator that promotes antitumor immunity driven by dendritic cell (DC) and macrophage recruitment to the tumor area and production of cytokines [26]. These cells then take up apoptotic and necrotic tumor cell bodies. Tumor antigenloaded DCs then migrate into the lymph nodes and promote T-cell mediated tumor immunity [27]. Imiquimod has received FDA-approval for the treatment of actinic keratosis, superficial basal cell carcinoma, and external genital warts, and is commercially available under various trade names, e.g., Aldara ® , Zyclara ® , Beselna ® , or R-837.

RCM monitoring of the response of BCC to imiquimod (Figures 2-4)
In 2003, our group reported the use of RCM for the diagnosis of BCC and for monitoring BCC clearance following the application of imiquimod [28]. The patient underwent a three-week course of daily application of a thin layer of 5% imiquimod cream to the clinically involved area. RCM demonstrated in exquisite detail the week-to-week response of the affected area to the treatment. RCM revealed intense inflammation, but after three weeks, there were no clinical signs of BCC. In 2004, another study used RCM to determine the effect of imiquimod treatment on BCC prior to Mohs microsurgery [29]. The area of the lesion was treated five times/week for two, four or six weeks, then Mohs microsurgery was performed two to four weeks after completion of imiquimod treatment. RCM was used to assess pre-surgical tumor clearance with an efficacy comparable to that of histology. This study suggests that RCM can be used to monitor the response of a lesion to imiquimod, aiding in the patient's and dermatologist's decision on whether surgery is further needed.

RCM monitoring of the response of actinic keratosis to imiquimod
The applicability of RCM to evaluate the response of AK to imiquimod has been assessed in a small group of patients with subclinical features of AK [30]. In these patients, RCM was used successfully to characterize the subclinical phenotypes of AK before treatment, as well as to assess the effect of imiquimod throughout a four-week course of treatment (which included three applications/week). Another study compared vis-a-vis the effect of imiquimod (three times/week for four weeks) and 5-ALA PDT in AK in organ transplantation patients [25]. This study, which involved imiquimod treatment of AK on the balding heads of patients, reports that RCM allowed excellent visualization of the evolution of response. This study was used to generate a table of RCMvisualized parameters and their usefulness for diagnosis of AK and for monitoring its response to imiquimod or PDT.
However, the sample size was small, and larger numbers are required to make these into general guidelines. An additional case of AK with SCC features was treated with a combination of imiquimod, tretinoin, and 5-fluorouracil [31]. The combined regimen was applied once daily for three weeks, followed by a two-week rest period, and then by another three-week treatment cycle. RCM evaluation 8 weeks after completing the treatment revealed improvement, although complete clearance was not achieved.

Treatment of melanoma with imiquimod and monitoring by RCM
Although imiquimod is not the reference treatment for melanoma, it has been assayed for the treatment of amelanotic or poorly melanotic melanoma and lentigo maligna melanoma (LMM). This is particularly important for LMM due to the fact that these tumors are heterogeneous, display poorly defined margins and appear in cosmetically sensitive areas, where surgery has an elevated risk of disfigurement [32]. In general, the response of melanoma to imiquimod has been often positive [33], but some foci seem to be resistant, casting severe doubt on this approach for long-term melanoma management [34]. RCM has been used to monitor the response of melanoma to imiquimod therapy in several studies. In one example, imiquimod (16-week course, one application/day) was used to treat one clinical case of melanoma where advanced age and other clinical complications prevented alternative treatments [31]. In this case, RCM revealed complete clearance after one year. In another case of amelanotic melanoma, the patient used 5% imiquimod cream twice a day for one week until erythema and irritation occurred, and then once a day for four additional weeks.
Mid-treatment RCM evaluation revealed inflammation, with clearance of atypical melanocytes. Imiquimod therapy was restarted, with application every three days for six months to minimize significant erythema and irritant response. At the 1-year follow-up visit, RCM examination revealed no evidence of relapse [35]. A recent report describes one case of in situ melanoma (lentigo maligna type) treated over 27 weeks with daily imiquimod, but with significant treatment breaks due to inflammatory complications. In this case, complete clearance as revealed by RCM was not achieved and surgical excision was necessary [32]. The same study reports another case of in situ melanoma, lentigo maligna type. The patient turned down surgery due to cosmetic complications and, instead, underwent daily treatment with imiquimod for three months. Before and after RCM monitoring revealed complete clearance up to one year after completion of the treatment [32]. A third case of melanoma on sun-damaged skin, treated daily with imiquimod for three months, was reported. RCM examination three and six months after treatment did not reveal any sign of relapse and instead showed recuperation of the normal morphology of the different epidermal layers [32].

Cryotherapy and shave biopsy
Recent studies have reported the use of RCM to monitor the response to two other types of therapy procedures: shave biopsy for AK [36] and cryotherapy for BCC [37].
In the AK study, the authors performed shave surgery in

Anti-inflammatory treatments (Figures 5 and 6)
RCM has also been used to follow up the response to other types of therapy, e.g., NSAIDs. One example is the response of actinic cheilitis (a form of actinic keratosis that affects the  Figure 2) with 5% imiquimod cream. Figure 3(a) shows the presence of spongiosis at the granular/spinous layer, which is characterized by accentuation and increased brightness of intercellular borders. Figure   3 lips) to anti-inflammatory (COX-1/COX-2 inhibitor) treatment [38]. The patients were treated with 3% diclofenac in 2.5% hyaluronic acid, twice daily for 90 days. In this study, the authors were able to diagnose actinic cheilitis with RCM in 6/7 (86%) patients. In addition, RCM offered good sensitivity in monitoring the response to diclofenac, revealing regression of the epidermal dysplasia upon consecutive RCM observations.

Laser-based ablation of non-malignant growths
Laser treatment is a viable treatment option for several benign tumors, e.g., angiomas or sebaceous hyperplasias.
RCM has been used to monitor the effect and evolution of laser treatment in these two types of lesions. One clinical case of cherry angioma was treated with a 585 nm flash lamppumped pulsed-dye laser (585 nm, with one pulse of 5 J/cm 2 using a 5 mm diameter spot) or a 568 nm continuous-wave krypton laser (power=0.75W, 1 mm diameter spot, exposure time = 1 second); the clinical response (from 1 hour up to 4 weeks) was evaluated using RCM [10]. RCM revealed early inflammation followed by resolution of the inflammation and disappearance of the lesion. In another study, the effect of PDL on sebaceous hyperplasia was monitored using RCM. The patients received three stacked, 5 mm wide pulses of a 585 nm pulsed-dye laser equivalent to 7 or 7.5 J/cm 2 [9].
RCM revealed that, whereas most lesions undergo significant involution over time, recurrence is frequent.

Determination of surgical success and presence of remaining malignant cells
A very important application of RCM as a monitoring tool is the identification of residual neoplastic cells that were left behind at the margins of surgery. These residual cells are the most frequent cause of BCC relapse. A few years ago, a study demonstrated that RCM could be used to identify residual cancer cells after Mohs micrographic surgery of BCC [39].
Although the study included a small group of patients, a strong correlation was observed between the RCM findings and those of conventional histology; thus, the study strongly supports the utility of RCM in scanning surgical margins for residual cancer cells.

Conclusion and future perspectives
Although its main application remains diagnostic, the potential of RCM to follow the response to therapy in a completely non-invasive manner is untapped. The major advantages are: RCM is not invasive, allowing repetitive optical tissue sampling; RCM is also independent of contrast agents, yet still provides good resolution, approaching that of histology; the usefulness of RCM has been improved by the existence of image atlases with histological correlation; and finally, to the trained practitioner, RCM offers ease of use. We identify three major problems, which are being overcome by a combination of ingenuity, hard work and interdisciplinary collaboration. One, technical, is the lack of penetration of RCM light into the deeper layers of the skin; however, technical and optical advances may increase our working depth in the near future. A related problem is resolution. However, the future expansion in clinical use of the new superresolution confocal microscopes (which rely on statistics to obtain resolutions under 50 nm using light microscopy [40]), promises to be of help. Finally, the data reported so far has been generated based on small-sized patient groups. This is related to the fact that the technique is relatively uncommon in dermatology practice. A combination of good formative opportunities for skin professionals and subsequent wordof-mouth regarding the wealth of high-resolution information this technique offers will increase manifold the patient samples in the near future.
In 2008, a group of basic researchers, clinicians and experts from related fields established an international RCM group (www.confocal-icwg.com/). Together, we have under-