1. Introduction
Lips constitute a special location for the development of numerous skin lesions due to their frequent exposure to exogenous factors, such as ultraviolet light, chemical, and biological agents.
Actinic cheilitis (AC) is one of the most often occurring pathologies that affects the lips [
1]. Occupational activities not considered, studies report an AC prevalence between 0.2% and 0.45% [
2,
3].
Upon clinical examination, AC has a broad spectrum of presentation, comprised of pale, dry, scaly lips, chronic ulcerations and erosions [
4], blurring of the vermillion-skin border [
2,
5,
6,
7,
8,
9,
10], white [
11,
12] and red [
2,
5,
9,
10] areas, and vermillion atrophy [
5,
6,
7,
8]. Palpation reveals a fine sandpaper-like feeling [
4], often accompanied by a stinging or burning and an inelastic or tight sensation of the lip [
4,
8].
AC differential diagnosis includes inflammatory disorders such as eczema, benign leukoplakia, lichen planus, granulomatous cheilitis, and xerosis with chronic irritation [
13,
14].
Some studies show that squamous cell carcinoma (SCC) represents approximately 90% of all oral malignancy cases [
15] while others estimate that 95% of SCCs of the lips emerge from ACs [
12,
16].
Changes suggestive of AC progression to SCC of the lip include thickening and induration of keratotic AC patches, the appearance of nodules with rapid growth and/or ulceration associating bleeding and pain [
4,
8,
17,
18,
19,
20]. Lip squamous cell carcinoma is also much more prone to metastasis than cutaneous SCC (0.5–3% vs. 3–20%) [
21,
22,
23].
A number of techniques hold promise for the early detection, aggressiveness profile and monitoring of keratinocyte carcinomas. The observed differences between normal, inflammatory and malignant keratinocyte proteomic profiles are likely to unearth novel markers for SCC, in terms of diagnosis and monitoring, and could maybe even come to the aid of targeted therapies [
24,
25,
26,
27].
Although the gold standard diagnostic technique for AC and lip SCC is the histopathological examination of a biopsy specimen, the anatomic characteristics of the lips increase the risk of postoperative bleeding and infection. Additionally, considering the cosmetic importance of this area, noninvasive diagnostic techniques are useful for selecting the biopsy site, thus avoiding repeated biopsies, and in some cases even acting as a surrogate for histopathology.
Imaging techniques, such as dermoscopy and
in vivo reflectance confocal microscopy (RCM), continue to highlight diagnostic and prognostic criteria for AC and SCC [
28,
29]. The lips are an ideal site for RCM examination, due to a thinner epidermis when compared to other body sites. Because early detection and swift therapy remain the two most important factors influencing the long-term survival of these patients [
30], we designed a retrospective study with the aim of consolidating previous observations regarding the RCM diagnostic criteria for AC and lip SCC.
2. Materials and Methods
2.1. Subjects
Patients with histopathologically confirmed lesions of actinic cheilitis or lip SCC were included in the study. Patients’ records were retrieved from the electronic database of the Dermatology Research Laboratory, “Carol Davila” University of Medicine and Pharmacy, in Bucharest. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Local Ethics Committee (No. 25/27.11.2017). All participants gave written informed consent as part of their investigation and treatment procedures, at the time of their registration.
2.2. RCM Imaging and Analysis
Despite it being a retrospective study, the RCM imaging protocol was the same for all lesions, as it is a well-established protocol in this clinic when examining non-melanocytic lesions.
Confocal imaging was done with a commercially available confocal microscope (Vivascope
® 1500, Caliber ID, Rochester, NY, USA) which uses a near-infrared laser diode with a maximum power output of 20 mW. The device and image acquisition protocol have been described elsewhere [
31].
Vertical imaging via Vivastack® was performed by capturing a series of images of 0.5 × 0.5 mm with 3 µm increments, in depth. Horizontal mosaics (via Vivablock®) of 4 × 4 mm were captured at different depths of the lesions. Mapping started at the stratum corneum and was continued to the papillary dermis.
Due to the problematic separation of epidermal layers on RCM, we adopted the methodology of Peppelman et al. [
32], so that the first appearance of nucleated cells, regardless of cell size and shape, was considered to be the stratum granulosum (SG). Since the SG is only a few cell layers thick, three steps in depth below this point was considered as the stratum spinosum (SS).
RCM images were then evaluated for these features by an RCM user (LM) with 4 years of experience with this technique. The observer systematically evaluated the lesions for the presence or absence of individual RCM criteria, but was not blinded to the histopathological diagnosis. Furthermore, the mean vessel diameter and blood vessel density per individual confocal image (500 × 500 µm) were determined for both AC and lip SCC lesions. Based on previously published data [
32], an increased vascular diameter was defined as a diameter greater than 5 µm and an increased blood vessel density as more than 5 vessels per single confocal image (500 × 500 µm). Blood vessel diameter measurements were taken on scale calibrated images using the open source software package ImageJ. A line, perpendicular to the axis of the vessel, was drawn from side to side in the widest visible part of the blood vessel and the result was recorded. This measurement was repeated for every visible vessel in the chosen RCM image. For every case, these two parameters were measured in 3 different single confocal images at approximately the same depth for every lesion, and the highest value for each of these two parameters was recorded for the case.
2.3. Histopathology
All lesions included in this study were surgically excised after RCM investigation and histologically confirmed as either AC or lip SCC by an experienced pathologist on haematoxylin–eosin (H&E) stained paraffin sections. SCC was defined based on the presence of an invasive component.
2.4. Statistical Analysis
The analysis was conducted in order to assess how various observed RCM criteria were associated with either AC or SCC. It comprised descriptive statistics and Fisher’s exact test to analyze the differences between subgroups. The differences in blood vessel diameter and vessel density per individual confocal image in AC and SCC lesions were measured using an independent t-test.
All data analyses were conducted using the statistical software package SPSS Inc. (v20, Chicago, IL, USA).
3. Results
Twelve subjects (10 males and 2 females) with a mean age of 66.82 ± 9.87 (range 43–80) years were included in the study.
In these patients, a total of 12 biopsy-proven lesions were evaluated with RCM, of which four ACs and eight invasive lip SCCs. The lesions were all located on the vermillion of the lower lip. In the actinic cheilitis subgroup, there were two smokers (one male and one female) and in the SCC subgroup there were three smokers (all male). None of the subjects included in this study were immunosuppressed.
The degree of differentiation for the SCC lesions included in the study, along with gender, age, immune and smoking status, are illustrated in
Table 2.
3.1. RCM Features for Differentiating between AC and Lip SCC
Hyperkeratosis (4/4), parakeratosis (3/4), an atypical honeycomb pattern (4/4), and the presence of dyskeratotic, targetoid cells within the epidermis (4/4) were RCM features present in virtually all the examined AC lesions (
Figure 1). Total epidermal disarray, dendritic and atypical cells in the dermis, and tumor nests in the dermis were found in none of the AC lesions.
Lip SCCs were typified in confocal examination by the presence of the complete disruption of the epidermal architecture (8/8) and the presence of dermal inflammatory infiltrate (6/8). Atypical cells in the dermis and dermal tumor nests could be detected upon RCM examination in only half of the SCC lesions (
Figure 1). Dendritic cells, most probably corresponding to Langerhans cells, were seen in three out of eight SCCs. Although in SCCs with a pigmentary component, melanocytes can be seen in RCM, none of the tumors included in our study had a clinically or dermatoscopically visible pigmentary component. Even so, we cannot exclude the possibility that some of the observed dendritic cells could indeed be melanocytes.
Ulceration, hyperkeratosis/scale, and dermal solar elastosis were present in both AC and SCC with similar frequencies.
Table 3 shows the frequencies of the various RCM criteria in the studied AC and lip SCC lesions.
Table 4 contains the frequencies of the observed RCM criteria according to SCC degree of differentiation. None of the RCM criteria varied significantly between well and moderately differentiated SCC, as assessed by the Chi-square test.
The assessment of associations between RCM criteria and the diagnosis of AC or lip SCC revealed that an atypical honeycomb pattern (
p = 0.002, Fisher’s exact test) and the presence of target keratinocytes in the epidermis (
p = 0.01, Fisher’s exact test) were strongly associated with AC (
Figure 2), while complete epidermal disarray (
p = 0.002, Fisher’s exact test) was characteristic for lip SCC. The RCM features for AC/SCC discrimination in this study have been summarized in
Table 5.
3.2. Vascularization in AC and SCC Lesions
The vessel diameter and number of blood vessels per single RCM image were higher in SCC (
Figure 3). The mean blood vessel diameter in SCCs was 18.55 µm larger than that in AC lesions (
p = 0.006). There was no significant difference (
p = 0.64) in blood vessel density, as measured by RCM, between SCC and AC lesions in our sample (
Table 6).
4. Discussions
Aside from the particular functional and cosmetic significance of the lips, several conditions ranging from benign infections to dysplasias, and the potentially fatal SCC, may develop in this area. Hence, the examination of the vermillion and lip mucosa is an important part of the dermatological examination.
While there are numerous publications regarding RCM imaging of various skin conditions ranging from tumors to infections and inflammatory conditions [
33,
34,
35,
43,
44,
45,
46,
47], there are only a few reports in the literature related to the non-invasive diagnosis of lip lesions. Whilst RCM knowledge and experience in the field of non-melanoma skin cancer is constantly expanding [
48,
49,
50,
51], obtaining biopsies is still an invasive procedure with its own limitations, mainly due to cosmetic reasons and the risk of sampling errors.
An obvious limitation of our study is the small sample size of only 12 lesions. It remains difficult to have a large sample size of SCCs in non-invasive diagnostic studies. The hyperkeratotic scale makes the evaluation of SCC difficult with either dermoscopy or RCM, hence the limited advantage of these techniques in clinically evident SCCs. Additionally worth mentioning is the difficulty raised by the need for histopathological confirmation in this type of study. The most important struggle in this field remains in distinguishing between SCC in situ and invasive SCC in the case of clinically similar lesions.
Ulrich et al. [
41] defined the RCM criteria for AC as: disruption of stratum corneum, hyperkeratosis, parakeratosis, atypical honeycomb pattern at the SS and SG, dermal solar elastosis, dilated blood vessels, and the presence of inflammatory cells in the upper dermis. In our study, the most common RCM feature for both AC and SCC was the presence of keratinocyte pleomorphism resulting in either an atypical honeycomb pattern in the case of AC or a complete disruption of the epidermal architecture in SCC. Furthermore, testing for the association between RCM criteria and these two entities, we uncovered that target cells are significantly associated with AC. These findings are in accordance with previous studies which examined the RCM appearance of actinic keratoses and cutaneous SCCs [
32,
38,
52,
53].
As opposed to previous studies [
54], dermal dyskeratotic keratinocytes were observed in only 50% of SCC lesions included in our study, either as isolated, scattered cells or as tumor nests. However, similar to earlier research [
54], we found dermal dendritic cells in under half of the lip SCCs (3/8) and these results do not allow for a significant association between this element and lip SCC. On the other hand, we observed the presence of inflammatory cells in the dermis in half of the ACs and almost all (6/8) of the SCCs, which is in line with other studies [
54], and reveals traits of the tumor microenvironment [
55]. While not statistically significant, Hartmann et al. [
56] reported an increased frequency of RCM-observed dermal tumor nests and peritumoral inflammatory infiltrates in moderately differentiated SCCs compared to well differentiated SCCs. In our data, none of the RCM criteria were significantly different between well and moderately differentiated SCC. In our case, we attribute this to the small sample size.
When analyzing AC and lip SCC vasculature, we found a significantly increased mean vascular diameter and a larger blood vessel density for SCC compared to AC. Our results are in accordance with other studies, and can be explained by the high metabolic needs of a tumor, which determines vascular dilation and neovascularization [
32,
57,
58].
To conclude, this study, building upon previous research, confirms several RCM criteria which can be used to distinguish between AC and lip SCC in vivo. This warrants further prospective, large sample-size studies, which will form the basis for the development of protocols for the correct, efficient and expeditious diagnosis of AC and SCC of the lips.