1.

Introduction

The skin is one of the tissues most investigated using optical spectroscopy because of its easy access and the great diversity of skin alterations that give significant changes in optical properties.¹ Cutaneous squamous cell carcinoma (SCC) and basal cell carcinoma (BCC), commonly termed ‘nonmelanoma skin cancer’ (NMSC) or ‘epithelial skin cancers,’ are among the most common cancers in white-skinned populations. While malignant melanoma originates from melanocytes, NMSC arises from keratinocytes. Despite its low mortality rate, BCC may cause considerable morbidity as it may occasionally grow aggressively causing extensive local tissue destruction or recurrences after treatment.² Cutaneous SCC, is associated with a 0.5% to 5% risk of metastasis in the general population, and often displays a more aggressive and difficult-to-treat course.³

Existing treatments, however, often do not meet the challenge of an effective management,³ especially in patients with multiple lesions, advanced or widespread disease, highlighting the need for early diagnosis. Recent developments in optical imaging techniques may provide the possibility of real-time noninvasive, high-resolution analysis of skin and can potentially decrease morbidity and mortality in NMSC. Novel diagnostic methods, using the fluorescence and reflectance spectroscopic modes have been introduced to detect human tumors based on the spectral properties of tissue, avoiding the need for a skin biopsy. Several of these technologies have become commercially available, including light-induced fluorescence (LIF) and diffused reflectance spectroscopy (DRS), optical coherence tomography (OCT), near-infrared and Raman spectroscopy, multispectral fluorescence polarization imaging, and reflectance or fluorescence confocal microscopy.^4,5 In this review, we will address spectroscopic well-established as well as upcoming and bench techniques used for the photodiagnosis of NMSCs.

2.

Fluorescence

Light-induced fluorescence spectroscopy is a very attractive potential diagnostic technique for early diagnosis and demarcation of skin cancer due to its high sensitivity, easy-to-use methodology for real-time measurements, and noninvasiveness. Fluorescence emission from malignant skin tissue is usually excited using as an irradiation source a laser, light-emitting diode (LED), a xenon, or halogen lamp.

2.1.

Autofluorescence Spectroscopy

Laser-induced fluorescence (LIF) has been proposed for the diagnosis of NMSC.^6–11 Autofluorescence imaging is an attractive potential tool due to lack of need for contrast agents’ application on the tissue under investigation. It allows differentiation on the base of differences in biochemical content and metabolic state of the pathology.^1,12–14 After illumination of the tissue with an appropriate light source, the photons diffuse in the tissue and fluorescence is re-emitted, enabling noninvasive spectroscopic measurements and the ‘optical biopsy’ approach.¹

Skin components, like aromatic aminoacids: phenylalanine, tryptophan, and tyrosine fluoresce under excitation with deep ultraviolet (UV) (in the spectral region 260 to 300 nm). Structural proteins: collagen and elastin and their cross-links fluorescent under excitation with 320 to 400 nm, in addition with nicotinamide adenine dinucleotide (NADH) and flavins, which are indicators of metabolic alterations, in the cutaneous tissue.^15–19 As far as the source of autofluorescence excitation of skin components is concerned, various light sources in UV and the blue-green spectral region are used for excitation of autofluorescence in skin tissues.^1,16,20–23 A scheme of a typical setup for autofluorescence spectroscopy is shown in Fig. 1.

Fig. 1

Schematic diagram of the light-induced fluorescence.

Researchers in 2001, such as Brancaleon et al.,²⁰ observed higher fluorescence intensity in nonmelanoma skin tumors versus normal skin, using UV excitation for the tryptophan residues, which may be a result of epidermal thickening in the tumor site. In contrast, the fluorescence intensity associated with collagen cross-links was lower in tumors, because of the erosion and degradation of the connective tissue and the decrease in collagen and elastin crosslinks.²⁰

In 2002, Panjehpour et al. studied 49 patients [BCC, SCC, actinic keratoses (AK), and normal skin] and compared diagnostic accuracy in laser-induced fluorescence spectroscopy for skin types I to III to determine the skin colors’ effect on the results.²² Normal skin exhibited stronger fluorescence emission than BCC and SCC with 410 nm excitation. The accuracy of classifying NMSC was 93% in type I skin and 78% for type III, due to the higher absorption of melanin.^5,22

Different excitation wavelengths exhibited lower fluorescence signal in BCC tumors, in comparison with normal skin fluorescence, which is observed by the group of Na et al.,¹⁶ Zeng et al.,²⁴ Borisova et al.,^25,26 and Drakaki et al.¹ In addition, a red shift of the spectral profile of BCC skin biopsies ex vivo was revealed during the research of optical characterization of normal and malignant skin after excitation with a nitrogen laser by Drakaki et al.¹ An example of that spectroscopic process is shown in Fig. 2.

Fig. 2

(a) A representative photo of skin sample with BCC. (b) Histopathologic image in BCC area of the skin sample of part (a) (H&E). (c) Fluorescence spectra of healthy skin tissue and skin tissue with basal cell carcinoma after excitation with 337 nm of nitrogen laser.

In the wavelength spectrum of 350 to 450 nm, the higher fluorescence intensity shown in the normal tissues is due primarily to a conversion of collagen (emission around 390 to 405 nm) and elastin (emission around 400 to 410 nm) from normal to malignant tissues and a decrease in NADH (emission around 440 to 460 nm) levels in the malignant tissues.¹

However, skin is a complex multilayered and inhomogeneous organ, with spatially varying optical properties that complicate the quantitative analysis of the fluorescence spectra.²⁷ In addition, the intensity of the fluorescence signal may not be a consistent parameter for a decision as several factors, such as surface morphology of the distance between the tissue surface and probe, can affect the signal. Diagnostic algorithms are able to correlate the varied spectra and differentiate malignant tumors from normal tissues.^1,5,19,28

3.

Diffuse Reflectance Spectroscopy

The diffuse reflectance spectrum reflects the absorption and the scattering properties of the skin tissue, where the absorption coefficient is directly related to the concentration of physiologically relevant absorbers in the tissue, and the scattering coefficient reflects the size and the density of the scattering centers in tissue.¹ As the detected diffuse reflectance signal is superposition from diffuse scattering and absorption from tissues’ pigments, the resultant spectrum can reveal information about main absorbers in the skin tissues, like hemoglobin and melanin and its pathologies and can provide valuable information regarding both tissue morphology and function and identify benign and cancerous or cancer prone skin lesions for the purpose of reducing unwanted benign biopsies.

A number of diffuse reflectance-based techniques have been developed for the evaluation of nonmelanoma.^45–47 The advantage of a DR spectroscopic device lies in its binary diagnosis (cancer/no cancer) making it an attractive monitoring method, that can be operated with minimal additional training.

Most of the researchers found lower spectral intensities in BCC skin tissues compared to the corresponding normal skin.^28,48 Thomson et al.⁴⁸ were able to diagnose BCCs and neighboring healthy skin, using linear discriminant analysis in a leave one-out protocol, with a sensitivity and specificity of 100% and 71% respectively. Rajaram et al.²⁸ found statistically significant differences ($p<0.05$) in the optical properties between normal skin and the different histologic groups, like BCC, SCC, and AK. One explanation according to Rajaram et al.²⁸ is the possible breakdown in the collagen matrix present in the dermis. The contribution of collagen to the overall scattering at the source-detector separation exceeds that of cells and nuclei. Because of this, a small increase in the epidermal scattering due to large-scale malignant cell proliferation would still be masked by a larger reduction in dermal collagen scattering. This trend in scattering from normal to malignant cancer is consistent with studies of other researchers on skin, where the scattering coefficient was significantly lower for every cancerous group, when compared to normal skin.^45,49 DRS spectra from Rajaram et al ^28,50 exhibited a large flattening of the diffuse reflectance spectrum at approximately 420 nm corresponding to the Soret band of hemoglobin absorption, while change in line shape of the diffuse reflectance spectra of BCC showed large depressions in the Q-bands (540 to 575 nm) of hemoglobin absorption too. Similar trends were also noticed for AK and SCC, however, these were not found to be statistically significant. Based on statistic and mathematical analysis they determined the values of mean vessel diameter for both normal skin and BCC and they reported that the change in mean vessel diameter from normal skin ($15\pm 2\mu \mathrm{m}$) to BCC ($40\pm 6\mu \mathrm{m}$) is statistically significant.⁵⁰

When autofluorescence is detected in vivo, researchers need to take into account the influence of cutaneous absorbers and scatterers, which distort the spectral shape of the autofluorescence signal coming from the tissue. To remove distortions introduced in measured tissue fluorescence spectra by tissue scattering and absorption, the fluorescence spectra may be analyzed in combination with information from the corresponding reflectance spectra. This way, the intrinsic tissue fluorescence may be extracted.⁷ In addition, application of DRS could be combined with fluorescence also in highly pigmented, which exhibit too fluorescence signal.^1,19 Therefore, in the recent years, there has been growing interest in the common use of light-induced autofluorescence and reflectance spectroscopy to differentiate disease from normal surrounding tissue, mainly for detection and differentiation of cancerous and precancerous changes in human body.

Unfortunately there has been variability between and within the DR spectra groups and a classification difficulty due to the heterogeneity of lesions and anatomic differences, inter-patient differences in factors such as skin photo-type, level of sun exposure and age. There is a need for localization of the measurement volume and support for development of improved classifiers on groups of lesions, divided by specific lesion pathology.⁴⁸

4.

Raman Spectroscopy

Raman spectroscopy has strong potential for providing noninvasive dermatological diagnosis of NMSC. Raman scatters in tissues, including the cell cytoplasm, cell nucleus, fat, collagen, cholesterol-like lipid deposits, and water.¹⁵

Fendel et al.,⁵¹ Gniadecka et al.,⁵² Nunes et al.,⁵³ Schrader et al.,⁵⁴ Skrebova et al.,⁵⁵ and Barry et al.,⁵⁶ used Fourier transform (FT)–Raman spectroscopy to distinguish between BCC and surrounding normal tissue.

Gniadecka et al.⁵⁷ indicated alterations in protein and lipid structure in NMSC. Spectral changes were observed in protein bands, amide I (1640 to $1680{\mathrm{cm}}^{-1}$), amide III (1220 to $1300{\mathrm{cm}}^{-1}$), and n(CC) stretching (probably in amino acids proline and valine, 928 to $940{\mathrm{cm}}^{-1}$), and in bands characteristic of lipids, CH2 scissoring vibration (1420 to $1450{\mathrm{cm}}^{-1}$), and -(CH2)n- in-phase twist vibration around $1300{\mathrm{cm}}^{-1}$.

Nijssen et al.⁵⁸ demonstrated pseudo-color Raman maps based on FT–Raman spectra obtained from a frozen tissue section of BCC. In their work, they showed that the pseudo-color Raman map is closely related to the microscopic image of the hematoxylin and eosin (H&E)-stained sample.

Silveira et al.⁵⁹ used a near-infrared Raman spectrometer to obtain spectra of normal skin and BCC in vivo. There were notable differences in the Raman peak intensities at 800 to $1000{\mathrm{cm}}^{-1}$, 1200 to $1400{\mathrm{cm}}^{-1}$, and around $1600{\mathrm{cm}}^{-1}$. The authors implemented a diagnostic model to discriminate BCC from normal skin, using algorithms such as principal components analysis (PCA) and Mahalanobis distance. They reported 85% accuracy in differentiating normal skin from BCC.

R. de M. F. Pereira et al⁶⁰ used the FT-Raman spectroscopy to differentiate between normal skin and SCC in vitro, where spectra in the normal tissue showed the presence of vibrational bands in 860 and $939{\mathrm{cm}}^{-1}$ with higher intensity than in the carcinoma spectra. These modes were assigned to the vibration of proline and hydroxiproline. The shift region of 1555 to $1560{\mathrm{cm}}^{-1}$ showed a difference of intensity to the samples of SCC, which were attributed to the nucleic acid.

Several studies with FT–Raman spectroscopy has applied a long wavelength excitation laser (1064 nm $\mathrm{Nd}:\mathrm{YAG}$ laser or 850-nm titanium–sapphire laser) in order to minimize autofluorescence from skin tissue. However, the longer wavelength laser had poor Raman scattering intensities compared to the shorter excitation wavelength laser. As a result, previously reported Raman spectra on BCC tissues show poor signal-to-noise ratios and require statistical adjustment of spectroscopic data, such as PCA,⁵³ K-mean clustering analysis (KCA),⁵⁸ and neural networks⁵² to differentiate between Raman signals of cancerous and normal tissues.

Choi et al.⁶¹ showed that with confocal Raman microscopy, using a shorter wavelength argon ion laser at 514.5 nm, distinct Raman band differences up to 95% between normal and BCC tissues for the amide I mode and the ${\mathrm{PO}}_2^{-}$ symmetric stretching mode were found, without the need for statistical adjustment of spectral data. It was also possible to precisely differentiate BCC tissue from surrounding noncancerous tissue using the confocal Raman depth profiling technique.

However, none of these groups have explored the ability of Raman spectroscopy to provide diagnosis of NMSC (BCC and SCC) and inflamed scar tissue in vivo at scarred areas of previous biopsy or surgical resection, particularly important for the dermatologist. Lieber et al.⁶² predicted correctly with a diagnostic algorithm all of the BCC, SCC, and inflamed scar tissues and normal tissues with a 100% sensitivity, 91% specificity for abnormality, and with a 95% overall classification accuracy.

5.

Near-Infrared Spectroscopy

Infrared spectroscopy and Raman spectroscopy are two mutually complementary methods, which provide information on molecular composition and structure and interactions occurring in the sample.⁶³

In 1999, McIntosh et al. showed that biopsies from BCC, SCC, and melanocytic tumors have distinct mid-IR signatures compared to normal skin.^64,65 McIntosh et al.⁶⁴ used infrared spectroscopy to examine BCC to explore distinctive characteristics versus normal skin samples and other skin neoplasms. Spectra of epidermis, tumor, follicle sheath, and dermis were acquired from unstained frozen sections, and analyzed qualitatively, by t-tests and by linear discriminant analyses. Dermal spectra were significantly different from the other skin components mainly due to absorptions from collagen in dermis. Spectra of normal epidermis and BCC were significantly different by virtue of subtle differences in protein structure and nucleic acid content. Linear discriminant analysis characterized spectra as arising from BCC, epidermis, or follicle sheath with 98.7% accuracy. Use of linear discriminant analysis accurately classified spectra as arising from epidermis overlying BCC versus epidermis overlying nontumor-bearing skin in 98.0% of cases. Spectra of BCC, SCC, nevi, and malignant melanoma were qualitatively similar. Distinction of BCC, SCC, and melanocytic lesions by linear discriminant analyses was 93.5% accurate.⁶⁴

Complete absorption of mid-IR light results with samples greater than 10 to 15 mm in thickness, when in contrast, near-IR light is scattered to a much greater extent than it is absorbed, making tissues relatively transparent to near-IR light, thus allowing the examination of much larger volumes of in vivo tissue.⁶⁶ In vivo near-IR spectroscopic characterization of skin tumors presented by McIntosh et al., ⁶⁷ for differential diagnosis of BCC from other skin cancer types, where linear discriminant analysis (LDA) was used to determine whether spectra could be classified according to lesion type. This resulted in accuracy of 70% to 98% in differentiating benign from premalignant or malignant lesions.

In 2006, Salomatina et al.⁴⁹ determined and compared the in vitro NIR spectroscopic properties of human skin to those of NMSC. They observed a significant difference in the NIR scattering between cancerous tissues and control skin, and noted that the absorption was significantly lower for BCC than for control skin.

A combination of infrared spectroscopy with high pressure (pressure-tuning infrared spectroscopy) was reported by Wong et al.⁶⁸ who worked with cluster analysis in 1992 and applied it to the study of paired sections of BCC and normal skin from 10 patients. Atmospheric pressure IR spectra of BCC were dramatically different from those of the corresponding normal skin. Compared to their normal controls, BCCs displayed increased hydrogen bonding of the phosphodiester group of nucleic acids, decreased hydrogen bonding of the C-OH groups of proteins, increased intensity of the band at $972{\mathrm{cm}}^{-1}$, decreased intensity ratio between the ${\mathrm{CH}}_3$ stretching and ${\mathrm{CH}}_2$ stretching bands and accumulation of unidentified carbohydrates.

6.

Laser-Scanning Microscopy

Within the last years, laser spectroscopic methods have achieved substantial improvements in the imaging of dermal tissue in vivo. With the recent development of in vivo LSM systems, which can be directly applied on any human skin area, new promising potentials have opened up for LSM in medicine. Laser-scanning microscopy allows the distinction of all types and formations of cells, as well as skin appendages, such as sweat glands and hair follicles. The laser-scanning microscopes are fiber-based systems which allow the investigation of all body sites. They have proved a potentially valuable diagnostic aid in BCC and SCC diagnosis, both ex vivo and in vivo.^69,70

The basic confocal-principle involves focusing a laser light onto a small spot within the dermal tissue. A light signal out of the focus point is collected at the same time in order to create a confocal image. Mainly the returning light directly from the focal point is only detected, whilst prevention of any scattered and reflected light from out-of-focus planes increases the imaging contrast. A deep scan can be performed easily by moving the light focus deeper into the tissue. As a result, different cell layers can be observed, the skin architecture can be studied without taking a destructive biopsy.⁷¹

There are three types of confocal spectroscopic methods in use: the reflectance, fluorescence, and the Raman spectroscopy.^72–74 Depending on the laser source like an Argon laser 488 nm, or a NIR diode laser at 830 nm, the detection system and the usage or not of a contrast dye, different dermatologic in vivo laser-scanning microscopes are now commercially available. In the case of reflectance confocal microscopy (RCM), systems illumination and detection takes place at the same wavelength. in vivo laser-scanning microscopy in reflectance and fluorescence mode generate horizontal optical sections at a resolution that is comparable to routine histology, but at the present time, these techniques are limited by their penetration depth.⁴ Compared to routine histology, those images are oriented horizontally to skin surface (en face) and they are analyzed in gray scale. There are several advantages of in vivo laser-scanning microscopy (LSM) over conventional histology. The imaging is painless and noninvasive, causing no tissue damage or alteration by processing or staining. Another aspect is the rapidity with which the data can be collected, and the ability of LSM for repeated evaluations of dynamic processes in real time.

7.

Optical Coherence Tomography

Optical coherence tomography (OCT), another type of optical imaging modality, was first demonstrated in 1991.⁹³ In particular, OCT, an interferometry-based method, performs high resolution, cross-sectional tomographic imaging of the internal microstructure of skin by measuring backscattered and back reflected light and has been recognized as a promising tool for the diagnosis of pathological changes.⁹⁴ OCT has greater image resolution than ultrasound, typically 5 to 15 μm. Compared to confocal microscopy and multiphoton microscopy, OCT has a larger field of view and penetration depth and offers cross-sectional images of the skin similar to histopathology sections.⁹⁵

However, in a highly scattering tissue like skin, OCT has a relatively limited penetration depth of up to 2 mm, since this method operates with ballistic photons,^5,96,97 while lateral resolution of 10 to 15 μm is more typical.^{94,95,97–101} New systems with ultra-high resolution of about 1 μm have recently been developed. OCT uses wavelengths within the band 600 to 2000 nm where the main constituents of the tissue, water, pigments, etc. exhibit low absorption.⁵ In order to obtain maximum penetration depth, matching index gel can be applied in the skin tissue or we could use longer wavelengths around 1300 nm, while for a better image contrast, central wavelengths around 800 nm are more appropriate.¹⁰² Broadband sources such as superluminescent diodes, semiconductor-based light sources, or femtosecond lasers (e.g., $\mathrm{Ti}:{\mathrm{Al}}_2{\mathrm{O}}_3$) have short coherence lengths and thus are good candidates for use in OCT.¹⁰³

Some OCT scanners for dermatological purposes are already available on a commercial basis.¹⁰⁴ OCT can reliably identify epidermis and the dermo-epidermal junction (DEJ),¹⁰⁵ while it is also able to identify the normal regional differences.¹⁰⁵ It has shown promise as a diagnostic tool for the detection and imaging of NMSC.^5,94,100 A breakup of the characteristic layering of normal skin^105,106 is found in OCT images of NMSC lesions.^{5,94,100,106–108} However, this disruption of layering is also seen in various benign lesions as sebhorreic keratosis¹⁰⁸ and benign melanocytic nevi.¹⁰⁹ Gambichler et al.⁹⁴ confirmed with their experiments that thinning of the epidermal layer corresponded to severe actinic changes including epidermal atrophy, as shown in histology. Moreover, faint lobular structures, often appearing in honeycomb-like aggregates, could be visualized in the upper dermis. These structures usually had a signal-poor core surrounded by a light backscattering ring.⁹⁴

Gambichler et al.,¹⁰⁴ in a more systematic study, found a loss of normal skin architecture and a disarrangement of the epidermis and upper dermis on OCT images of BCCs as compared with adjacent nonlesional skin. Furthermore, large plug-like signal-intense structures, honeycomb-like signal-free structures, and prominent signal-free cavities in the upper dermis were requently detected.

Several other NMSC features in OCT images have been described. The most important are: focal changes including thickening of epidermis (AK);^5,110 dark rounded areas, sometimes surrounded by a white area (BCC basaloid island cell clusters and surrounding stroma); and increased penetration depth in OCT images.^5,94

Mongensen et al.^97,111 found a sensitivity of 79% to 94% and specificity of 85% to 96% in differentiating normal skin from lesions for OCT. Discrimination of AK from BCC was not possible in their research.

Other studies also suggested a good match between OCT images and histopathology in BCC lesions,¹⁰⁷ although BCC subtypes could not be identified in OCT images.¹¹² In the previous year, Olmedo et al.¹⁰⁷ performed an interesting OCT study on BCC where they found lobules, islands, and infiltrating strands of BCC appeared similar in OCT images and histologic sections regardless of the type of BCC. Notably, Olmedo et al.¹⁰⁷ reported that multifocal superficial BCC can be differentiated from other BCC subtypes. In addition, the tumor tissue observed in the deeper dermis appeared as signal-poor round structures due to penetration depth of OCT. In that matter, when Korde et al.¹¹⁰ examined superficial BCC, a subepidermal highly reflecting band was found in the OCT image that corresponded to the tumor in histological sections.

SCC has mainly been studied on oral mucosal surfaces using OCT, but changes similar to BCC have been described.¹¹³ Coleman et al.¹¹⁴ confirmed previous reports of the limitation in the use of OCT in SCC,^97,101,110 where there is shadowing by surface hyperkeratosis.

An advanced version of OCT, the polarization-sensitive OCT (PS-OCT) has the ability to visualize and quantify the birefringence properties of skin and thus has been used as an efficient tool for tissue segmentation.¹⁰⁴ Many types of tissue are distinguishable in PS-OCT, in particular those with a significant collagen content such as the skin.^104,105 Several types of invasive lesions, including BCC, can cause a breakdown of collagen structure, resulting in a loss of birefringence which has been quantified using PS-OCT.^100,115

In most researchers work, no statistically significant difference was found with conventional OCT between nodular, multifocal superficial, and infiltrative BCC. This may be due to the relatively small sample size in studies of BCC subtypes such as multifocal superficial BCC and infiltrative BCC.⁹⁴ In this matter, nowadays a newer developed version of OCT with high-definition facilitates in vivo diagnosis of BCC and allows the distinction between different BCC subtypes for increased clinical utility.¹¹⁵

This is an innovative technique based on the principle of conventional OCT, with the ability to carry out optical imaging up to 570 μm deep within highly scattering media such as skin and with a μm resolution in both lateral and axial directions, giving it the potential to visualize individual cells.¹¹⁵ Instead of a single pin diode, it uses a two-dimensional, infrared-sensitive (1000 to 1700 nm) imaging array for light detection. This enables focus tracking: the focal plane is continuously moved through the sample,¹¹⁶ which in addition to the vertical OCT imaging mode offers a real-time, horizontal, so-called en-face imaging mode, which allows the immediate visualization of OCT pictures with high resolution in both dimensions.¹¹⁷

On the other hand, the higher resolution results in a reduced penetration depth making HD-OCT not the medium of choice for the measurement of thick tumors. Here, conventional OCT with a reported penetration depth of nearly 1 to 2 mm seems to be superior.¹¹⁷ Nevertheless, HD-OCT may serve as an important adjunct which may reduce the need for invasive biopsies in some of the cases, e.g., multiple tumor or patients who may request nonsurgical therapy. In such cases, HD-OCT makes it possible to conduct longitudinal studies of lesional therapy such as PDT or other nonsurgical treatment procedures.¹¹⁶

It may be advantageous to combine OCT with other imaging modalities, in the so called multimodality imaging such as Raman, Doppler, fluorescence, and ultrasound. This approach may further improve diagnostic capabilities in skin research as well as future dermatology practice.

8.

Conclusions

Novel and established diagnostic techniques, involving spectroscopic and imaging diagnosis for NMSC have been reviewed in this article. Some of these novel techniques have the potential to become clinically available for the noninvasive diagnosis of NMSC. Many of these diagnostic methods seem to offer an adequate diagnostic accuracy, especially as a supplement to clinical diagnosis. The optical imaging technologies, such as fluorescence diffuse reflectance, Raman and near-infrared spectroscopy, optical coherence tomography, and confocal laser-scanning microscopy have been discussed for their resolution and accuracy in dermatological diagnosis. While the skin biopsy and histopathological assessment are yet to be replaced, the evolution and combination of spectroscopic techniques may be a promising technology for the photodiagnosis of NMSC, in the future.