The differential diagnosis of lichenoid keratosis and seborrheic keratosis with reflectance confocal microscopy: a preliminary study

Lichenoid keratosis (LK, or lichen planus-like keratosis, LPLK) and seborrheic keratosis (SK) present as similar benign keratotic lesions on cosmetically sensitive area, but require different therapies. Both lesions can be easily differentiated based on histological evaluation of biopsy materials. However, the biopsies may cause scarring and result in hyper-pigmentation, which reduces the compliance of the patients to be treated. In this study, we investigated the role of reflectance confocal microscopy (RCM) in the non-invasive differential diagnosis of LK and SK. Cases with facial brown patches or plaques suspicious of SK were enrolled in the study. After written informed consent was obtained, the lesions were photographed, imaged by RCM, and then biopsied. The RCM findings were analyzed and correlated with histology results. Evaluation of the RCM pictures and confirmation with histological results were conducted by two independent dermatologists. In total, 10 cases were enrolled in the study. The main characteristics of LK lesions observed by RCM were the disarray of the dermal–epidermal junction (DEJ), and marked inflammatory infiltrates in the superficial dermis; while prominent cerebriform pattern, or elongated cords with bulbous projections without significant inflammation reaction, were the features of SK. Among the 10 cases, clinically suspicious of facial SK, 4 were determined as LK, 6 as SK by RCM imaging, and all the RCM findings were confirmed by histological results. The RCM features of LK and SK have significant difference, highlighting the important role of RCM in the differential diagnosis of LK and SK, avoiding biopsies and allowing safe treatments.


Introduction
Lichenoid keratosis (LK), also known as benign lichenoid keratosis (BLK), or lichen planus-like keratosis (LPLK), has similar clinical presentation with similar lesions and predilected locations with Seborrheic keratosis (SKs) [1][2][3]. Therefore, LK is easy to be misdiagnosed as SK, but it usually requires different treatment modality than SK. A definitive differential diagnosis can be difficult to the naked eye because of the variability in clinical appearance. Histologic analysis could easily differentiate those two entities, as they have different microscopic findings among LK and SK [2,3]. However, when these lesions are presented on a cosmetically sensitive area, avoiding biopsy, which may cause scarring and result in hyper-pigmentation, is of particular benefit.
In the past two decades, the non-invasive, cellular resolution imaging with reflectance confocal microscopy(RCM) had been studied and increasingly accepted by clinical physicians and patients [4,5]. The recently reported RCM features of LK [6,7] found in DEJ are those with edged papillae, elongated cord-like structures, bulbous projections, milialike cysts, and some bright round cells were also seen within the inter-papillary spaces, while for SK, a regular honeycomb pattern at epidermal levels, well-circumscribed dermal papillae at the DEJ, and inconspicuous dermal structures were observed [8].
In this study, we performed the skin imaging with RCM to investigate the role of RCM in the differential diagnosis of LK and SK.

Patients and methods
The study had been approved by the Ethics Committee of Henan Provincial People's Hospital. Patients with facial brown patches or plaques suspicious of SK were enrolled in the study. The inclusion criteria was: (1) Single facial lesion of brown patches or plaques, not papule lesions; (2) 5-20 mm in diameter; (3) Only suspicious for SK, no other skin tumors were included.
After written informed consent was obtained, the lesions were photographed, then imaged by a commercially available, reflectance mode confocal microscope (Vivascope 1500; Caliber Imaging & Diagnostics, Inc. formerly Lucid, Inc., Rochester, NY, USA). The captured horizontal images in a 500 × 500 μm field, and viva-block image of 3 mm × 3 mm at stratum spinosum, basal cell layer, dermal-epidermal junction and superficial dermis layer were obtained and compared with the adjacent normal skin. A detailed description of the technique and the device has been published previously [6]. After RCM image obtention, the same lesions were biopsied with surgery excision, and fixed in phosphatebuffered neutral formalin, embedded in paraffin, and stained with hematoxylin-eosin (HE), then analyzed using an optical microscope to investigate the accuracy of RCM imaging.
The evaluation of the RCM pictures and confirmation with histological results were conducted by two independent dermatologists who did not participate in the patient's enrollment.
The diagnostic criteria based on RCM images were: (1) Exclude a melanocytic lesion (absence of nests, pagetoid cells, and confluence of melanocytes) and an atypical keratinocytic lesion (no keratinocytic pleomorphism, not much parakeratosis) [8]. (2) Investigate the hyperplastic pattern of the epidermis. (3) Investigate the integrity of the basal cell layer and (4) the inflammation of the papillary dermis to differentiate LK and SK. Those features of the disarray of the dermal-epidermal junction (DEJ), and the marked inflammatory infiltrates in the superficial dermis were considered as LK; while those with cerebriform pattern, or elongated cords with bulbous projections without significant inflammation reaction, were the features of SK.

Results
In total, 10 cases (8 female and 2 males) were enrolled in the study, detailed information is presented in Table 1. The average age of the patients was 46.2 years old, and the average duration of the lesion was 5.7 months. Among the 10 cases, clinically suspicious of facial SK, 4 were determined as LK, 6 as SK by RCM imaging, then confirmed by histological results.
On clinical examination, single light brown or dark brown patch or plaque lesions, with 5-20 mm in diameter, was observed on facial area (Fig. 1a, 2a). Based on the RCM images, and compared with the perilesional skin, the hyperkeratosis, thickening of spinous layer, loss of the high refractive annular structure(disarray of DEJ architecture) and evenly dispersed bright inflammatory cells and melanophages in superficial dermis were defined as LK (Figs. 1b, d, 3a),, correlated with the vacuolar degeneration of basal layer and superficial dermal infiltration composed of lymphocytes and other cells in histology results, while those with cerebriform pattern (Figs. 2c,  3b), which correlated with the epidermal papillomatous hyperplasia in histological images (Fig. 2d), or elongated cords with bulbous projections (Fig. 2b), indicating the adenoid pattern of SK, without significant inflammatory reaction, were observed and defined as features of SK for those patches or plaque lesions under RCM images.
The most characteristic changes were the presence of DEJ disarray and the inflammatory infiltrates in the superficial dermis (Figs. 1b, 3a), or not (Figs. 2b, c, 3b), in the differential diagnosis of LK and SK. The definitive histological diagnoses were all confirmed with biopsy finding (Figs. 1d, 2d).

Discussion
The non-invasive skin imaging in dermatology clinic to provide clues for the diagnosis and differential diagnosis of skin diseases is world-widely accepted in the past two decades, among which, the RCM could reveal the epidermis and the upper dermis in cellular resolution. And the cellular changes of the lesion in different layers could be imaged and compared with that of the adjacent normal skin. RCM features of benign and malignant skin tumors were investigated with high sensitivity and specificity [9,10], compared with the histology results. In latest research, the determination of skin tumor boundary before surgery, the detection and follow-up of cutaneous tumor metastasis were explored [11,12]. Imaging studies Fig. 1 The clinical, RCM and histology images of LK. a The single brown patch of the right cheek; b (80 μm depth from the zero surface) showed the disarray of dermal-epidermal junction and the infiltration of pigment phagocytes (red arrow) and inflammatory cells (green arrow) in the superficial dermis (yellow circles), c showed the normal dermal-epidermal junction of the adjacent skin (80 μm depth from the zero surface); The histology of LK exhibited inflammatory infiltration obscuring the dermal-epidermal junction (d) of the skin conditions with RCM to avoid or reduce the invasive biopsy on cosmetically sensitive exposed areas is of increasingly interest. RCM may be considered a promising tool for differential diagnosis of syringoma and milia [13], and a non-invasive arsenal for the assessment of hair and scalp disorders, benefiting patients [14]. Even compared with the novel imaging device, such as the optical coherence tomography, and multi-photon microscopy, the RCM still has its unique advantages [15,16].
The popular lesion of SK is easily misdiagnosed as verruca plana (VP). Previously, we performed a study [17] to differentiate SK and verruca plana based on the RCM imaging,and we found that SK shows a cerebriform shape while VP has petal-like structures, which correlates well Fig. 2 The clinical, RCM and histology pictures of SK. a Showed a single light brown patch on the left cheek. b (80 μm depth from the zero surface) The pattern of the elongated cords with bulbous projections seen on RCM (arrow), c (80 μm depth from the zero surface) showed the typical cerebriform pattern of SK, d showed the typical histology image of SK Fig. 3 The mosaic picture of RCM. The presence of DEJ disarray and the inflammatory infiltration in the superficial dermis (noted the red circle and star area) is the character of LK (a) while the cerebriform pattern is SK (b). Noted the rectangular area, the red star (gyrus like structure) and red arrow (sulcus like structure) with the papillomatous hyperplasia of SK and vacuolized keratinocytes of VP in histology images.
The patch or plaque lesion of SK can be difficult to distinguish from LK by naked eye examination. The previous study using dermoscopy to image the LK lesions showed features of SK-like fingerprinting along with focal granularity [2,18], indicating dermoscopy may be less conspicuous in the differential diagnosis of LK and SK.
The main RCM features of LK were investigated and correlated with histology results in several recent studies [19][20][21][22]. The typical RCM feature of LK [19,20] was found in DEJ with elongated cords, bulbous projections, milia-like cysts, and some bright round cells were also seen within the inter-papillary spaces.
Based on their findings, we investigated 10 cases, clinically suspicious of SK, with single light brown or dark brown patch or plaque lesions on face. For SK lesions, the typical characteristics are the cerebriform pattern of the epidermis, or the pattern of the elongated cords with bulbous projections, without significant infiltrate of inflammation in dermal-epidermal junction, based on the RCM imaging. While for LK lesions, the typical features are the DEJ disarray and inflammatory infiltration in the superficial dermis. The most important image feature to differentiate LK and SK is the presence of DEJ disarray and inflammatory infiltration in the superficial dermis, or not, in RCM images.
After careful analysis of the RCM images, 4 of the 10 cases were determined as LK, 6 as SK, and the RCM results were well correlated with histological analysis. These results are promising and suggest a potential important role of RCM in differential diagnosis of LK and SK.
In this study, we focused on LK, which is easily confused with SK by the naked eye. And we found the RCM is beneficial in the non-invasive differential diagnosis of SK and LK, avoiding the biopsy, which may cause scarring and result in hyper-pigmentation and the differential diagnosis of LK and SK is important with respect to difference in the clinical management. The lesions of SK could subside completely after several sessions of laser treatments without significant side effects, while the treatment of LK with lasers is more difficult, which needs more sessions and lead to great probability of post inflammation hyper-pigmentation, due to the lichenoid inflammation reaction of LK. The most common treatment of LK is calcineurin inhibitor or imiquimod cream [22]. Clinically, there may be some SK, that should be differentiated with actinic keratosis or basal cell carcinoma [23], which indicates the important role of RCM in screening and differential diagnosis of LK.
In summary, we investigated the role of RCM in the differential diagnosis of LK and SK, and we found the presence of DEJ disarray and inflammatory infiltration in the superficial dermis are the main point of LK, which is different from SK, The RCM plays an important role in the non-invasive differential diagnosis of LK and SK, avoiding biopsy and allowing safe treatments.
The main limitation of our study is the small sample size, for the consents to do biopsy on the facial skin is difficult to be obtained. Clearly, we need more cases to confirm these findings in future study, and compared those results with dermoscopy findings.
Author contributions Jing Chen and Huaxu Liu designed the study, Zhaopeng Zhang and Min Gao wrote the main manuscript text, Yan Lin and Huaxu Liu collected the clinical and RCM data, Shengli Chen collected the histology data. All authors reviewed the manuscript.
Funding The study was supported by Medical Science and Technology Research Project of Henan Provincial Health Commission(SBGJ202002101), and Fund of Henan Provincial People's Hospital (SBGJ202002101).