In vivo reflectance confocal microscopy of mycosis fungoides: A preliminary study

Non-invasive Skin Imaging in Cutaneous Lymphomas

The diagnosis of cutaneous lymphomas is challenging and requires skin tissue for histology and immunophenotyping using immunohistochemistry and molecular studies. In recent years, the role of non-invasive imaging techniques has been described as part of the clinical assessment of cutaneous lymphoma lesions. Imaging modalities such as dermoscopy, reflectance confocal microscopy (RCM), and high frequency ultrasound (HFUS) have been shown to be very valuable in raising the clinical suspicion for lymphomas of the skin, and in distinguishing cutaneous lymphomas from inflammatory dermatoses such as lupus, psoriasis, or eczema. These non-invasive methods can be used to direct the clinician to the optimal biopsy site to maximize the histopathological results and minimize false negatives. These methods also have a potential place in monitoring treatment response. In this review we present a concise summary of the dermoscopic imaging, RCM, and HFUS features seen in cutaneous T-cell lymphomas (CTCL) and B-cell lymphomas (CBCL).

Reflectance Confocal Microscopy, Optical Coherence Tomography, and Multiphoton Microscopy in Inflammatory Skin Disease Diagnosis

BACKGROUND AND OBJECTIVES

Technological advances in medicine have brought about many novel skin imaging devices. This review aims to evaluate the scientific evidence supporting the use of noninvasive optical imaging techniques to aid in the diagnosis and prognosis of inflammatory skin diseases.

STUDY DESIGN/MATERIALS AND METHODS

PubMed and Scopus were searched in September 2020 according to PRISMA guidelines for articles using reflectance confocal microscopy (RCM), optical coherence tomography (OCT), and multiphoton microscopy (MPM) in inflammatory skin diseases, excluding studies monitoring treatment efficacy.

RESULTS

At the time of the study, there were 66 articles that addressed the utilization of noninvasive imaging in interface, spongiotic, psoriasiform, vesiculobullous, and fibrosing/sclerosing inflammatory skin dermatoses: RCM was utilized in 46, OCT in 16, and MPM in 5 articles. RCM was most investigated in psoriasiform dermatoses, whereas OCT and MPM were both most investigated in spongiotic dermatoses, including atopic dermatitis and allergic contact dermatitis.

CONCLUSIONS

There is preliminary evidence to support the diagnostic potential of noninvasive optical imaging techniques in inflammatory skin diseases. Improvements in the devices and further correlation with histology will help broaden their utility. Additional studies are needed to determine the parameters for diagnostic features, disease differentiation, and staging of inflammatory skin conditions. Lasers Surg. Med. © 2021 Wiley Periodicals LLC.

The Differential Diagnosis of Hypopigmented Mycosis Fungoides and Vitiligo With Reflectance Confocal Microscopy: A Preliminary Study

Objective: To investigate the role of reflectance confocal microscopy (RCM) in the differential diagnosis of hypopigmented mycosis fungoides (HMF) and vitiligo.

Methods: Cases with persistent hypopigmented patches, suspicious of early stage vitiligo, or HMF were imaged with RCM. The melanin contents and inflammatory conditions of the epidermis and superficial dermis of the lesions were compared with the same layers of the adjacent skin, and then, the imaged lesions were biopsied and analyzed by histology.

Results: 15 cases were enrolled in this study, and based on the RCM findings, there was just slight or moderate reduction of melanin but no melanin absence in the basal cell layer of HMF lesions. The finding of monomorphous weakly refractile, oval to round cells on the basis of vesicle-like dark space was clearly elucidated in the epidermis of the lesions by RCM, which indicates the Pautrier's microabscesses on histopathology. Among those 15 cases, 13 cases were identified as HMF, and the other two cases were vitiligo, based on RCM findings, which were confirmed by histology analysis.

Conclusions: The RCM findings correlated well with histology results in the screening of HMF, which indicates the RCM is an important tool in the early detection and differential diagnosis of HMF.

Reflectance confocal microscopy role in mycosis fungoides follow-up

BACKGROUND

Reflectance confocal microscopy (RCM) is a useful tool for many skin cancers, allowing non-invasive evaluation over time and identifying areas of active disease. Its role to follow-up mycosis fungoides (MF) patients has not yet been evaluated.

OBJECTIVE

To assess the level of agreement between RCM and histopathology and to develop a RCM checklist that could help monitoring MF patients.

METHOD

Prospective study in a cutaneous lymphoma clinic of a tertiary hospital in Australia. RCM and biopsies were performed on the same area at baseline, before commencing or changing treatment, and at 6 months after starting treatment. Normal skin sites were also analysed and acted as controls. RCM features and histopathological findings were blindly evaluated by the confocalist and pathologist. Correlation between RCM and histology was measured by overall per cent of agreement (OPA), kappa and ROC curves. Additionally, RCM images before and after treatment were assessed blinded from clinical information and correlated to clinical assessment.

RESULTS

Thirty-eight MF lesions were included. Nineteen of these 38 were re-assessed by RCM 6 months later. Fifty biopsies were performed (38 at baseline and 12 after 6 months). The combination of four RCM features corresponding to Pautrier's microabscess, epidermal and junctional lymphocytes and interface dermatitis formed the RCM checklist for MF that predicted the severity of disease with AUC of 0.95 (P = .003).

CONCLUSION

Reflectance confocal microscopy can assess activity within a lesion and over time and assist in the clinical management of patients with MF.

Reflectance confocal microscopy may be included as part of the diagnostic algorithm of early-stage mycosis fungoides

INTRODUCTION

Reflectance confocal microscopy (RCM), a noninvasive, real-time technique of computed tomography, has been widely used for pigmentary, inflammatory, and tumor diseases of the skin.

AIM

Our main purpose was to analyze the consistency between pathological and RCM characteristics of early-stage mycosis fungoides (MF) and the utility of RCM in the diagnosis of early-stage MF.

METHODS

According to the RCM features of MF in the early stage, the biopsy sites of 40 cases of suspected MF and 20 cases of chronic inflammatory skin diseases clinically were preliminarily located. Histopathologic and immunohistochemical examinations were performed to make a diagnosis based on the diagnostic algorithm proposed by the International Society for Cutaneous Lymphomas.

RESULTS

Among the 60 patients observed, there were 12 confirming cases of MF, 14 suspecting cases, 6 not completely excluding cases, and 28 eliminating cases according to the diagnostic algorithm, as well as characteristics of RCM were typical in 8 cases, suspected in 16 cases, not excluded in 3 cases, and excluded in 33 cases. The kappa value was 0.769 (P < .01), which means there is a strong consistency between the classification by RCM and the diagnosis algorithm. MF in patch stage and plaque stage (IA to IIB) has typical characteristics of RCM, respectively.

CONCLUSIONS

RCM can be used as an objective and convenient auxiliary means to diagnose early-stage MF and may be included as part of the diagnostic algorithm of early-stage MF.

Systematic review and proposal of an in vivo reflectance confocal microscopy assessment tool for cutaneous lymphoma

BACKGROUND

Reflectance confocal microscopy (RCM) is a non-invasive imaging technique that provides dynamic information and allows in vivo monitoring, with excellent histologic correlation. In the last decade, the use of RCM for cutaneous T-cell lymphomas (CTCL) has been reported. CTCL may require multiple biopsies for diagnosis due to its equivocal clinical presentation. RCM was described as a possible tool to help determine the best site for skin biopsy. This study aims to systematically review all RCM features reported in literature for CTCL.

METHOD

A systematic literature search concerning CTCL evaluated by RCM was performed in eight electronic databases until May 2019 following PRISMA-DTA quality assessment.

RESULTS

Eighteen RCM features were described in patients with CTCL. The most frequent were: interface dermatitis (89%), epidermal lymphocytes (82%), epidermal architectural disarray (81%), and vesicle-like structure (Pautrier microabscess) (51%).

CONCLUSION

In order to establish comparable parameters among the studies identified, we proposed descriptors for CTCL features and a grading system to quantify them. This will facilitate to define the role of RCM in the diagnosis and monitoring of CTCL patients.

High-Contrast, Low-Cost, 3-D Visualization of Skin Cancer Using Ultra-High-Resolution Millimeter-Wave Imaging

The goal of this paper is to develop a new skin imaging modality which addresses the current clinical need for a non-invasive imaging tool that images the skin over its depth with high resolutions while offering large histopathological-like contrasts between malignant and normal tissues. We demonstrate that by taking advantage of the intrinsic millimeter-wave dielectric contrasts between normal and malignant skin tissues, ultra-high-resolution millimeter-wave imaging (MMWI) can achieve 3-D, high-contrast images of the skin. In this paper, an imaging system with a record-wide bandwidth of 98 GHz is developed using the synthetic ultra-wideband millimeter-wave imaging approach, a new ultra-high-resolution imaging technique recently developed by the authors. The 21 non-melanoma skin cancer (NMSC) specimens are imaged and compared with histopathology for evaluation. A programmable measurement platform is designed to automatically scan the tissues across a rectangular aperture plane. Furthermore, a novel frequency-domain imaging algorithm is developed to process the recorded signals and generate an image of the cancerous tissue. The high correlations achieved between MMWI images and histological images allow for rapid and accurate delineation of NMSC tissues. The millimeter-wave reflectivity values are also found to be statistically significant higher for cancerous areas with respect to normal areas. Since MMWI does not require tissue processing or staining, it can be performed promptly, enabling diagnosis of tumors at an early stage as well as simplify the tumor removal surgery to a single-layer excision procedure.

Current and future applications of confocal laser scanning microscopy imaging in skin oncology

Confocal laser scanning microscopy (CLSM) is a modern imaging technique that enables the in vivo or ex vivo characterization of skin lesions located in the epidermis and superficial dermis with a high quasi-microscopic resolution. Currently, it is considered to be the most promising imaging tool for the evaluation of superficial skin tumors. The in vivo mode adds the advantage of noninvasive, dynamic, in real-time assessment of the tumor associated vasculature and inflammation. It offers the possibility to repeatedly examine the same skin area without causing any damage and to monitor disease progression and treatment outcome. Furthermore, this novel technology allows the evaluation of the entire lesion and can be used to guide biopsies and to define tumor margins before surgical excision or other invasive therapies. CLSM diagnostic features may differentiate between the various histologic subtypes of skin tumors and therefore helps in choosing the best therapeutic approach. In this study, we present the CLSM characteristic features of the most common melanocytic and non-melanocytic skin tumors, as well as future possible CLSM applications in the study of experimental skin tumorigenesis on animal models.

Cutaneous T-cell Lymphoma

Cutaneous T-cell lymphomas comprise a heterogeneous group of diseases characterized by monoclonal proliferations of T lymphocytes primarily involving skin, modified skin appendages, and some mucosal sites. This article addresses the basic clinical, histologic, and immunohistochemical characteristics of this group of diseases, with additional attention to evolving literature on dermoscopy, reflectance confocal microscopy, flow cytometry, and molecular data that may increasingly be applied to diagnostic and therapeutic algorithms in these diseases. Select unusual phenotypes or diagnostic examples of classic phenotypes are demonstrated, and flags for consideration while making a pathologic diagnosis of cutaneous T-cell lymphoma are suggested.

Reflectance confocal microscopy features of mycosis fungoides and Sézary syndrome: correlation with histopathologic and T-cell receptor rearrangement studies

BACKGROUND

Mycosis fungoides/Sézary syndrome (MF/SS) often requires multiple skin biopsies for definitive diagnosis. In vivo reflectance confocal microscopy (RCM) visualizes high-resolution cellular detail of the skin. The objective of this study is to evaluate the morphologic features of MF/SS using RCM and to correlate RCM features with histopathology and T-cell receptor (TCR) gene rearrangement studies.

METHODS

A cohort of patients with active/recurrent or suspicious MF/SS disease was prospectively recruited for RCM imaging and histopathologic/RCM images were evaluated. Statistical analyses were performed to identify unique RCM features and to correlate RCM features with histopathologic findings and TCR rearrangement studies.

RESULTS

Eighty-three lesions were evaluated. Correlation between RCM and histopathology was moderate for all relatable features (κ = 0.41, p<0.001), almost perfect for intraepidermal atypical lymphocytes [prevalence and bias-adjusted kappa (PABAK) = 0.90], and fair for Pautrier collections (κ = 0.32, p = 0.03). Lesions with Pautrier collections identified by RCM were significantly more likely to show TCR clonality (p = 0.04) and diagnostic features of MF/SS on histopathology (p = 0.03).

CONCLUSIONS

Our study captures morphologic RCM criteria for a variety of skin lesions. Pautrier collections visualized by RCM are associated with improved histopathologic diagnosis and detection of TCR gene clonality. Although further studies are needed to validate the diagnostic implications of RCM for MF/SS, our study highlights the potential utility of RCM.

Reflectance confocal microscopy for monitoring the density of Demodex mites in patients with rosacea before and after treatment

BACKGROUND

Demodex mites seem to serve as a pathogenic trigger in many Demodex-associated diseases such as rosacea. In facial skin of patients with rosacea significantly higher numbers of Demodex mites have been shown compared with healthy controls. Reflectance confocal microscopy (RCM) allows the detection and quantification of Demodex mites in vivo noninvasively. It is hypothesized that a reduction of Demodex mites under rosacea therapy can be monitored by RCM.

OBJECTIVES

To use RCM to monitor the density of Demodex mites in patients with rosacea before and after treatment.

METHODS

In 25 patients with facial rosacea RCM was performed before and after therapy. Mosaics of 5 × 5 mm(2) and 8 × 8 mm(2) were scanned, and the total numbers of mites per follicle and per area were counted, along with the number of follicles per area.

RESULTS

In all patients Demodex folliculorum could be detected and quantified using RCM. RCM showed significant differences pre- and post-treatment (P = 0.0053 for 5 × 5 mm(2) and P < 0.001 for 8 × 8 mm(2)). The mean numbers of mites per follicle were 0.63 (range 0.16-2.28) per 8 × 8 mm(2) area and 0.70 (range 0.11-2.20) per 5 × 5 mm(2) area before treatment, and 0.41 (range 0.074-1.75) and 0.51 (range 0.094-1.70), respectively, after treatment. The corresponding mean numbers of mites were 155 (range 45-446) and 86.2 (range 12-286), respectively, before treatment and 96.2 (range 18-363) and 58.5 (range 12-230), respectively, after treatment.

CONCLUSIONS

By RCM, a reduction in the density of Demodex mites in facial skin of patients with rosacea under therapy, correlating to clinical improvement, can be quantified and monitored noninvasively. Possible reasons for this therapeutic effect are discussed.

Malignant T cells secrete galectins and induce epidermal hyperproliferation and disorganized stratification in a skin model of cutaneous T-cell lymphoma

Cutaneous T-cell lymphomas (CTCLs) are the most common primary skin lymphomas, which are characterized by an accumulation of malignant T cells in the skin. The early lesion resembles both clinically and histologically benign inflammatory disorders and also presents with hyperproliferative epidermis and T-cell infiltration. Despite considerable progress in understanding the molecular mechanisms involved in the malignant transformation of T cells, the causes of the morphological and histopathological features of the disease are largely unknown. We used an organotypic model of CTCL to show that malignant T cells through the secretion of galectin-1 and -3 stimulate vigorous growth of keratinocytes. In parallel, malignant T cells induce disorganized keratinocyte stratification, resembling the early hyperproliferative stage of CTCL. We also observed a loss of attachment between the epithelial and mesenchymal compartments. In addition, hyperproliferation was followed by a downregulation of differentiation markers, such as keratin 10 and involucrin, and a decrease in barrier formation. In conclusion, we provide evidence that malignant T cells orchestrate the histopathological epidermal changes seen in CTCL.

Application of Confocal Laser Scanning Microscopy in Biology and Medicine

Fluorescence confocal laser scanning microscopy and reflectance confocal laser scanning microscopy are up-to-date highend study methods. Confocal microscopy is used in cell biology and medicine. By using confocal microscopy, it is possible to study bioplasts and localization of protein molecules and other compounds relative to cell or tissue structures, and to monitor dynamic cell processes. Confocal microscopes enable layer-by-layer scanning of test items to create demonstrable 3D models. As compared to usual fluorescent microscopes, confocal microscopes are characterized by a higher contrast ratio and image definition.

Non-invasive in vivo dermatopathology: identification of reflectance confocal microscopic correlates to specific histological features seen in melanocytic neoplasms

BACKGROUND

Reflectance confocal microscopy (RCM) allows for non-invasive, in vivo evaluation of skin lesions and it has been extensively applied in skin oncology although systematic studies on nevi characterization are still lacking.

OBJECTIVE

The aim of this study was to determine whether reliable RCM correlates to histological features used to diagnose melanocytic neoplasms exist.

METHODS

We blindly evaluated the RCM and histological features of 64 melanocytic neoplasms (19 non-dysplastic nevi, 27 dysplastic nevi, 14 melanomas) and analysed the data using Spearman's rho calculation.

RESULTS

Many histological features can be identified using RCM. Elongated rete ridges corresponded on RCM to edge papillae, whereas flattened rete ridges to several features which involve dermal-epidermal junction disruption. Bridging of junctional nesting (JN) corresponded on RCM to both JN with irregular size/shape and JN with short interconnections. While we could reliably identify dermal melanocytes, the RCM features did not reliably distinguish between benign and concerning dermal melanocytic arrangements, suggesting further refinement of dermal melanocytic RCM features is needed.

CONCLUSION

Reliable correlates for epidermal and junctional histological features used to diagnose melanocytic neoplasms are identifiable on RCM, suggesting harnessing histological criteria may be a reasonable method to move beyond the algorithmic approach.

Confocal microscopy patterns in nonmelanoma skin cancer and clinical applications

Reflectance confocal microscopy is currently the most promising noninvasive diagnostic tool for studying cutaneous structures between the stratum corneum and the superficial reticular dermis. This tool gives real-time images parallel to the skin surface; the microscopic resolution is similar to that of conventional histology. Numerous studies have identified the main confocal features of various inflammatory skin diseases and tumors, demonstrating the good correlation of these features with certain dermatoscopic patterns and histologic findings. Confocal patterns and diagnostic algorithms have been shown to have high sensitivity and specificity in melanoma and nonmelanoma skin cancer. Possible present and future applications of this noninvasive technology are wide ranging and reach beyond its use in noninvasive diagnosis. This tool can also be used, for example, to evaluate dynamic skin processes that occur after UV exposure or to assess tumor response to noninvasive treatments such as photodynamic therapy. We explain the characteristic confocal features found in the main nonmelanoma skin tumors and discuss possible applications for this novel diagnostic technique in routine dermatology practice.

Reflectance confocal microscopy for the characterization of mycosis fungoides and correlation with histology: a pilot study

BACKGROUND

Mycosis fungoides (MF) is cutaneous lymphoma of the T-cell lineage. MF is a diagnostic challenge. In vivo reflectance confocal microscopy (RCM) reproducible imaging technique already reported to be useful in the diagnosis of skin diseases. The aims of our study were to define RCM features of MF and to evaluate its feasibility in biopsy site selection.

METHODS

Each lesion was selected for RCM evaluation from 10 patients with an established diagnosis of MF. Subsequently, a biopsy of the same areas evaluated with RCM was rendered for histopathological examination.

RESULTS

A series of RCM features of MF was identified and shown to correlate well with histopathological evaluation. We could find hyperkeratosis in five patients (10 : 50%); disarray of honeycomb of stratum spinosum in three patients (10 : 30%). In 10 patients (10 : 100%) of the MF, we could find that dermal papillary rings were weak reflected light; round to oval cells diffusely widespread throughout the epidermis and in the papillary dermis, infiltration of inflammatory cells in superficial dermis. Two (10 : 20%) of them can find vesicle area opaca in plaque stage MF, filled with monomorphous weakly refractile oval to round cells.

CONCLUSION

The utility of RCM as a diagnostic tool for MF awaits further evaluation, although it appears to be promising for biopsy site selection.

Noninvasive in vivo detection and quantification of Demodex mites by confocal laser scanning microscopy

BACKGROUND

In many Demodex-associated skin diseases Demodex mites are present in abundance and seem to be at least partially pathogenic. So far all diagnostic approaches such as scraping or standardized superficial skin biopsy are (semi-)invasive and may cause discomfort to the patient.

OBJECTIVES

To see whether confocal laser scanning microscopy (CLSM) - a noninvasive method for the visualization of superficial skin layers - is able to detect and quantify D. folliculorum in facial skin of patients with rosacea.

METHODS

Twenty-five patients (34-72 years of age) with facial rosacea and 25 age- and sex-matched normal controls were examined by CLSM. Mosaics of 8 × 8 mm and 5 × 5 mm were created by scanning horizontal layers of lesional skin and quantification of mites per follicle and per area as well as follicles per area was performed.

RESULTS

In all patients D. folliculorum could be detected by CLSM and presented as roundish or lengthy cone-shaped structures. CLSM allowed the quantification of Demodex mites and revealed significant differences (P < 0·0001): the mean number of mites was 165·4 per 8 × 8 mm area and 94·2 per 5 × 5 mm area in the patients compared with 34·7 and 22·4, respectively, in the controls. The corresponding mean number of mites per follicle was 0·7 and 0·8, respectively, in the patients and 0·1 and 0·2, respectively, in the controls.

CONCLUSIONS

With the help of CLSM it is possible to detect, image and quantify Demodex mites noninvasively in facial skin of patients with rosacea.

Concordance between in vivo reflectance confocal microscopy and optical histology of lymphomatoid papulosis

BACKGROUND

Lymphomatoid papulosis is a primary cutaneous CD30+ T-cell lymphoma clinically characterized by a benign, chronic, recurrent course with self-limited papulo-nodular skin lesion eruption. In vivo Reflectance Confocal Microscopy is a non-invasive technique for real-time imaging of the superficial layers of the skin down to the superficial dermis with cellular-level resolution close to conventional histopathology. RCM has been previously reported to be useful in the in vivo evaluation of inflammatory diseases, skin tumours and also cutaneous lymphomas. Only two articles have been published on cutaneous lymphomas and none detailing confocal features of LyP. The aim of this manuscript was to describe the confocal features of LyP and their histological correlation to evaluate the possible application of this non-invasive tool in this T-cell lymphoma subtype clinical management.

METHODS

Five patients with histological diagnosis of LyP were imaged with RCM, followed by a skin biopsy on a clinically selected lesion.

RESULTS

High grade of correspondence between RCM and histopathology of LyP was observed, disclosing the potential rule of RCM at least for biopsy site selection.

CONCLUSION

Future studies on RCM for LyP vs. other cutaneous T-cells lymphomas and inflammatory skin diseases are needed to assess specificity and sensibility of our preliminary data.

Consistency and distribution of reflectance confocal microscopy features for diagnosis of cutaneous T cell lymphoma

Reflectance confocal microscopy (RCM) represents a noninvasive imaging technique that has previously been used for characterization of mycosis fungoides (MF) in a pilot study. We aimed to test the applicability of RCM for diagnosis and differential diagnosis of MF in a clinical study. A total of 39 test sites of 15 patients with a biopsy-proven diagnosis of either MF, parapsoriasis, Sézary syndrome, or lymphomatoid papulosis were analyzed for presence and absence of RCM features of MF. Cochran and Chi(2) analysis were applied to test the concordance between investigators and the distribution of RCM features, respectively. For selected parameters, the Cochran analysis showed good concordance between investigators. Inter-observer reproducibility was highest for junctional atypical lymphocytes, architectural disarray, and spongiosis. Similarly, Chi(2) analysis demonstrated that selected features were present at particularly high frequency in individual skin diseases, with values ranging from 73% to 100% of all examined cases.

In vivo reflectance confocal microscopy: usefulness for diagnosing hair diseases

BACKGROUND

Reflectance confocal laser scanning microscopy (R-CSLM) is a new diagnostic technique which allows visualization of "optical intersections" within the epidermis and superficial layers of the dermis. Outlines of cells and their architecture are imaged and may be analyzed both horizontally and vertically to the skin surface. The method proved useful in early melanoma detection. We evaluated the potential usefulness of this method in a short series of patients with hair diseases.

MAIN OBSERVATIONS

Two healthy persons and 6 patients with hair diseases (1 with alopecia areata, 1 with androgenic alopecia and 4 with genetic hair shaft abnormalities) were examined with the use of Vivascope 1500. In all patients one scalp location and one location in the mid forearm were evaluated. R-CSLM examination gave in all cases high quality images of the hair shaft intersections, at 1µm intervals, which allowed detailed analysis of the hair structure. Hair follicles could be partly visualized at a depth of up to 200µm, which allowed analysis of only superficial parts of the hair follicles. An additional hurdle was bright reflection within the follicular ostia, which decreased the perception of details in these images. Hair could be best visualized, when analyzed on flat surfaces. Receiving good quality images from convex surfaces on the scalp required additional effort from the patient (to not move) and from the physician (to obtain best possible fit of the "optic window" to the scalp).

CONCLUSIONS

These preliminary data show that R-CSLM may develop into a valuable tool in evaluation of hair shaft diseases. Further development is needed to apply this technique in abnormalities of the hair follicle and the perifollicular area.

In vivo reflectance confocal microscopy: automated diagnostic image analysis of melanocytic skin tumours

BACKGROUND

In vivo reflectance confocal microscopy (RCM) has been shown to be a valuable imaging tool in the diagnosis of melanocytic skin tumours. However, diagnostic image analysis performed by automated systems is to date quite rare.

OBJECTIVES

In this study, we investigated the applicability of an automated image analysis system using a machine learning algorithm on diagnostic discrimination of benign and malignant melanocytic skin tumours in RCM.

METHODS

Overall, 16,269 RCM tumour images were evaluated. Image analysis was based on features of the wavelet transform. A learning set of 6147 images was used to establish a classification tree algorithm and an independent test set of 10, 122 images was applied to validate the tree model (grouping method 1). Additionally, randomly generated 'new' learning and test sets, tumour images only and different skin layers were evaluated (grouping method 2, 3 and 4).

RESULTS

  The classification tree analysis correctly classified 93.60% of the melanoma and 90.40% of the nevi images of the learning set. When the classification tree was applied to the independent test set 46.71 ± 19.97% (range 7.81-83.87%) of the tumour images in benign melanocytic skin lesions were classified as 'malignant', in contrast to 55.68 ± 14.58% (range 30.65-83.59%; t-test: P < 0.036) in malignant melanocytic skin lesions (grouping method 1). Further investigations could not improve the results significantly (grouping method 2, 3 and 4).

CONCLUSIONS

The automated RCM image analysis procedure holds promise for further investigations. However, to date our system cannot be applied to routine skin tumour screening.