Raman Spectroscopy: Incorporating the Chemical Dimension into Dermatological Diagnosis.

Raman spectroscopy provides chemical analysis of tissue in vivo. By measuring the inelastic interactions of light with matter, Raman spectroscopy can determine the chemical composition of a sample. Diseases that are visually difficult to visually distinguish can be delineated based on differences in chemical composition of the affected tissue. Raman spectroscopy has successfully found spectroscopic signatures for skin cancers and differentiated those of benign skin growths. With current and on-going advances in optics and computing, inexpensive and effective Raman systems may soon be available for clinical use. Raman spectroscopy provides direct analyses of skin lesions, thereby improving both disease diagnosis and management.

What was known?

The Raman Effect describes an inelastic exchange of energy between light and matter

Modern Raman spectroscopy uses laser light source to non-invasively determine molecular composition

Raman spectroscopy has been used to probe and classify diseased and normal tissue in vivo.

Introduction

Raman spectroscopy is an innovative optical technology that provides in vivo biochemical and structural analyses of skin changes. It permits diagnosis at the chemical level, offering the clinician a noninvasive, real-time analysis of tissue without requiring tissue preparation or labeling.

Changes in diseased tissue are driven by changes in cellular biochemistry. Qualitatively and quantitatively analyzing the biochemical milieu, Raman spectroscopy can provide both a diagnosis and information on the extent of the disease process. Thus, entities that may look similar clinically may be noninvasively differentiated chemically. Raman spectral analysis may facilitate selecting therapeutic options for not only histoclinical variants of disease but also chemical variants of disease. Given that biochemical changes precede gross changes, Raman spectroscopy allows for early diagnosis and, thus, improved care.

History

Chandrasekhara Venkata Raman (November 7, 1888–November 21, 1970) [Figure 1] was born in Tiruvanikkayal, Chennai, India. Raman, a talented young scientist, matriculated into the Presidency College, Chennai, at the age of 12. He received his bachelor's degree at 16 and his masters in Physics at 19. On graduating, Raman received the esteemed position of Assistant Accountant General in the Indian Finance Department. A dedicated scientist, Raman spent his evenings and early mornings studying acoustics while maintaining his daytime job at the Finance Department. After 10 years of working in finance, Raman resigned his government position to occupy the Palit Professorship in Physics at the University of Calcutta. While returning from Oxford University, in 1921, Raman was fascinated by the deep blue color of the Mediterranean, questioning whether or not the water's hue was due only to the reflection of the sky. On arriving in India, Raman showed that the color of the sea resulted from scattering of sunlight by water molecules, independent of the sky's reflection. This was the turning point for Raman, marking his exploration of light. He performed these intricate but momentous experiments in the laboratories of the oldest research institute in India, the Indian Association for the Cultivation of Science (IACS). The IACS has a rich history, being founded and financed by enlightened nationalist Bengalis to pursue fundamental research in frontier areas in science. The IACS was a wholly indigenous endeavor, not receiving patronage from the then colonial rulers. Raman became the IACS Honorary Secretary in 1919 and carried on his experiments in its Bowbazar Laboratory in Calcutta.

Figure 1

Sir Chandrasekhara Venkata Raman, FRS (November 7, 1888–November 21, 1970) (Courtesy Raman Research Institute, Bengaluru)

In 1923, Arthur Compton demonstrated that X-rays lose energy when scattered inelastically by electrons. Raman and his co-investigator, K. S. Krishnan, hypothesized a similar energy transfer when normal light is scattered by molecules. To prove their hypothesis, the pair designed an intricate experiment in which they isolated violet light and passed it though a liquid sample. Most of the scattered light remained the same color; some of light, however, changed color, indicating a transfer of energy to or from the water molecules. This is referred to as the Raman effect. Raman's work was immediately recognized as it gave further proof for the quantum nature of light. He was bestowed many honors, including knighthood and, in 1930, the Nobel Prize in Physics for his work on the scattering of light and the discovery of the effect named after him. Raman was the first person of color, let alone the first Indian, to receive a Nobel Prize in the sciences.

The Raman Effect

The Raman effect describes an inelastic exchange of energy between light and matter.[1] Matter can either absorb or scatter light. A major fraction of the scattered light has the same energy, and thus the same frequency as that of the incident radiation. This is known as Rayleigh or elastic scattering.[2] There is a small fraction of light, however, that interacts with electron clouds around molecular bonds in a matter, which can be in the form of a gas, liquid, or a solid. Energy, in the form of a photon, excites vibrational states within a molecule in a gas or liquid or between molecules in a solid to cause displacement of the electron cloud and induce a change in bond polarizability. When the molecule relaxes, it emits a photon. Depending on the energy difference between the original and the relaxed states, there is a change in the frequency of the emitted light from that of the incident light. Molecules that relax to the preexcited state emit the same color light to give rise to elastic scattering. Molecules that relax to a different vibrational state, however, emit a photon of different energy and give rise to inelastic scattering. This generated photon is shifted in frequency by an amount corresponding to the energy of a particular transition. Such a change in energy gives rise to inelastic scattering of photons. The energy difference corresponds to a vibrational or rotational frequency of a molecule within gaseous, liquid, and solid states, or of vibrations between molecules in a solid, and is referred to as the Raman effect.

Modern Raman spectroscopy uses laser light source to generate elastic and inelastic scattering and results in vibrational spectroscopic information about molecules. The so-called spectral fingerprint contains information about the energies of molecular vibrations. The vibrational energies depend on the mass of particular, the nature of the chemical bonds connecting them, the symmetry of the molecular structure in a gas or liquid, and the local structure of the environment where they reside.

Raman Technology

Over the past decade-and-a-half, there have been multiple advances in producing systems for Raman spectroscopic investigations. The construction of durable lasers, efficient collection geometries, small diameter fiber optic probes and sensitive detectors have allowed for Raman spectroscopy's potential application to clinical science. Recent developments related to Raman spectroscopy have been widely recognized. Nobel Prizes in 1998, 1999, 2005, and 2009 were awarded for advancements in computational chemistry and density functional theory, laser femtosecond spectroscopy, laser-based precision spectroscopy and fiber optics and charge-coupled devices, respectively.

Optical fibers allow for remote sampling since both laser and scattered light travel through the same fiber optic probe. The micro-Raman spectrometer although not designed for remote sampling couples the spectrometer to an optical microscope, thus allowing Raman analysis with microscopic laser spot ideal for small sample features.

For in vivo dermatological studies, Schut et al. developed a Raman system to probe skin, nail, teeth and tongue.[3] Spectra acquisition took approximately 5 s. Lieber and Mahadevan-Jansen developed a hand-held Raman microspectrometer for clinical dermatological investigation.[4] Currently, there are instruments that can collect, in vivo, high-quality Raman spectra from human skin in less than a second and can penetrate to depths of several 100 μm.[567] To optimize data collection, Raman spectroscopy has been combined with other imaging technologies [Table 1].

Table 1

Instruments that integrate Raman spectroscopy with other imaging technologies

The Raman Spectrum

Since Raman scattering is relatively weak, high power excitation sources are used to induce a Raman signal from the sample. On excitation, the probed region produces a beam containing different wavelengths of light; some are from the Raman effects while others are from the elastic Rayleigh scattering. After the Rayleigh component is filtered out from the beam, the Raman scattered radiation is dispersed and then collected by the detector which converts the incoming photons into an electric signal for the entire frequency range.

The data are then displayed as a spectrum with the intensity (in counts per second or arbitrary units) of the scattered light expressed as a function of the Raman shift (in cm−1). Each molecule has a unique set of vibrational states and, thus, a Raman spectrum with a characteristic set of peaks [Figure 2]. Thus, the Raman spectrum of a substance provides an optical fingerprint from which its biochemical diversity can be deciphered.[13] Moreover, a molecule's influence on the spectrum is relative to its abundance in the sample. Thus, the Raman spectrum provides information on not only molecular composition but also relative composition. Statistical, chemical, and morphological analytical chemometric methods are then applied to extract quantitative information from a Raman spectrum.

Figure 2

Typical Raman spectra for human ventral forearm skin, measured in vivo. Spectrum shows sharp Raman peaks characteristic for carotenoid molecules. Permission provided

Raman Studies of Cutaneous Tissue

The study of skin tissue by Raman spectroscopy is relatively recent. In the 1990s, near-infrared Raman spectroscopy became available to overcome this obstacle. In 1992, Barry et al. recorded Raman spectra of the stratum corneum.[14] Further studies by this group identified Raman spectral biomarkers to characterize skin components.[1516] Skin pigmentation, hydration, and thickness were also evaluated with Raman spectroscopy.[171819] In 1995, Raman spectroscopy characterized the molecular state of the 5200-year-old skin from the “Iceman” (Similaun man or Ötzi).[20] The possibility of diagnosing hyperkeratotic skin conditions with Raman spectroscopy was also studied.[21]

Raman spectra have characterized the percutaneous absorption of drugs,[22] the effects of protective agents in skin creams[23] and the effect of aging on water and protein structures in the skin.[24] In 1997, Raman systems were developed to perform in vivo examination of the human skin.[2526] Caspers et al. pioneered the application of confocal Raman spectroscopy in studying the skin. In vitro spectra from 6 μm thick sections as well as in vivo spectra focused 85 μm below the skin surface were collected with this confocal system.[5] A confocal Raman instrument could inspect depth profiles of skin components such as natural moisturization factor (NMF) and the level of hydration.[5272829] There have also been in vivo Raman studies on drug penetration[3031323334] as well as the effect of moisturizers on the skin.[35]

Raman spectra, collected in vivo, of healthy human skin, have been analyzed.[636] There have also been in vivo studies measuring carotenoid concentration in the epidermis as a marker of the antioxidative potential of the skin[3738] or to reflect human nutrition status.[3940] Though much work has been done on normal skin and its functioning, there has been growing interest in applying Raman spectroscopy to the diagnosis of skin pathology. The following sections highlight studies that evidence the potential of Raman spectroscopy in helping make dermatological diagnoses.

Cutaneous Oncology and Raman Spectroscopy

Initial in vitro studies found that benign and malignant skin growths have differences in their Raman spectra. The changes are mostly in the spectral representations for lipids and proteins, some of which reflect known molecular changes related to skin pathology. For example, the increased intensity of the phenyl ring-related spectral band found in seborrheic keratosis (SK), actinic keratosis, squamous cell carcinoma (SCC), or basal cell carcinoma (BCC) is due to hyperkeratosis; the keratin molecule, which contains the phenyl ring structure, is more abundant in this process.[41] SKs have increased amounts of lipids and thus, have a higher C-H peak.[41] A keratoacanthoma's Raman spectra display a similar pattern, indicating an increased concentration of unsaturated lipids.[41] In studying carotenoids in the skin using Raman spectroscopy, Hata et al.[42] reported the carotenoid concentration correlates with presence or absence of skin cancer and precancerous lesions.

In an in vitro study applying neural network decision algorithms to recognize Raman spectra of BCC and malignant melanoma (MM), the correct classification rate close to 95% for BCC and 80% for MM.[43] A neural network is a statistical method, which emulates a biological neural system. It can resolve the paradigm that linear computation cannot by gathering representative data and invoking training algorithms to learn the structure of the data. Small distinctive bands, corresponding to specific proteins and lipids, were shown to hold the discriminating values used to diagnose skin lesions. Neural networks were also applied to differentiate MM from pigmented nevi, BCCs, and SKs based on spectral differences in the structures of proteins and lipids.[44] Another study distinguishing MM from benign pigmented melanocytic nevi emphasized the pivotal nature of preprocessing data to facilitate reproducibility and reliability of the spectral data analysis.[45] In vivo spectra of MM do not differ from in vitro spectra.[46] Pseudo-color maps have also been developed to help visualize the Raman data collected from the skin.[47] To do this, Raman spectra were collected on two-dimensional grids from frozen sections of a BCC specimen. Statistical methods assigned similar spectra the same color. A prediction model could correctly classify new tissue samples as BCC with 100% sensitivity and 93% specificity.[47]

More recently, real-time in vivo Raman spectroscopy, with an integration time of <1 s/lesion, could distinguish MM from SKs with a sensitivity of 95–99% and specificity ranging from 15% to 54%.[48] Moreover, the pigmentation of the skin does not lessen in vivo Raman spectroscopic evaluation. Yet this study only had 55 lesions and, as is the case for many Raman spectroscopy studies, a larger data set is required to establish significant power.[49]

Confocal Raman spectroscopy applied at various skin depths showed 95% separation between normal skin and BCC in vitro.[50] Polarized Raman microspectroscopy has been used to monitor peritumoral changes in the stroma.[5152] Though subtle, these changes can differentiate the peritumoral dermis in superficial and nodular BCC from normal, healthy dermis. In discriminating tumoral, nontumoral and peritumoral areas in skin sections, Raman spectroscopy could expedite Mohs excision of a tumor by providing a real-time intraoperative determination of tumor borders. Using Raman spectroscopy that probes in regions of high wave-numbers as to minimize interference from fiber optic connections, confocal Raman analysis has also been applied to identify tumor margins in vitro. In these experiments, prediction methods were applied that only classified spectra as BCC or non-BCC when they had a 70% chance of being correct. Under such restrictions, BCC could be detected with 100% certainty and noninvolved tissue with a certainty of 99%.[53] Another study validated that Raman spectra correlate well with histopathologic evaluation Mohs sections of BCCs.[54]

There is a dearth of studies on the collection and analysis of Raman spectra from squamous cell skin cancer. An in vitro experiment found significant differences in the Raman spectra among SCC, MM, BCC, and normal skin.[55] The Raman spectra were classified using data reduction algorithms into pathological states. Another study that looked specifically at SCC reported significant changes in spectral representations for collagen, keratin, lipids, nucleic acids and secondary structure conformations of proteins.[56] Lieber et al.[57] developed a portable confocal Raman system to measure Raman spectra for SCC, BCC, inflamed scar tissue and normal skin in vivo. All BCC, SCC and inflamed scar tissue could be correctly identified, and 19 of the 21 normal skin tissue were correctly classified. This conferred a 100% sensitivity and 91% specificity for abnormality with a 95% overall classification accuracy.

Diagnosis of Other Skin Diseases

Raman spectra of vitiliginous skin show accumulation of H2O2 and L-phenylalanine, supporting earlier observations concerning the pathophysiologic mechanism of this disease.[58] Raman spectroscopy has also followed the oxidation of methionine and cysteine in vitiliginous skin.[5960616263] In vivo Raman studies have found that epidermal L-tryptophan is oxidized via an H2O2 -mediated Fenton reaction.[64] Using Raman spectroscopy, Vafaee et al.[65] also found increased H2O2 -mediated oxidative stress in the patches of depigmentation caused by piebaldism.

Micro-Raman spectroscopy was used to detect chemical changes associated with bullous pemphigoid (BP). In comparing the dermoepidermal junction of normal skin to samples of diseased skin, the Raman spectra of the BP lesions had significant changes in protein content, reflecting the presence immunoglobulin and fibrin at the basement membrane.[66]

Raman spectroscopy has been applied to study the loss-of-function mutations of the filaggrin (FLG) gene. Such mutations contribute to the development of cornification and ichthyosis vulgaris and are strongly associated with atopic dermatitis. Spectra produced from confocal Raman microspectroscopy showed that individuals who were carriers of the FLG-null mutation had lower levels of NMF than those without the mutation.[67] A prospective study concluded that it is possible to determine infants susceptibility to atopic dermatitis by taking a Raman spectra of their skin at birth to analyze the content of the FLG protein.[68]

The biochemical profile of psoriatic stratum corneum demonstrates changes in lipid metabolism.[69] Raman spectra of psoriatic biopsies show significant changes in lipid and protein structure when compared to normal skin.[70] In vitro Raman spectroscopy also found significant differences in the stratum corneum of normal skin from that of symptom-free areas of skin in patients with psoriasis and atopic dermatitis.[71] Confocal Raman spectroscopy combined with optical coherence tomography could detect changes in psoriatic plaques associated with successful treatment in vivo.[72] As treatment progressed, Raman spectroscopy measured an increase in skin surface hydration, NMF, trans-urocanic acid and ceramide III along with a marked decrease in skin layer thickness of affected areas.[72]

Tophi and skin calcifications produce distinctive Raman spectra.[73] Though the diagnosis of tophi is simple in known cases of gout, the need for needle aspiration to demonstrate the presence of monosodium urate crystals is uncomfortable to the patient. Furthermore, when lesions are present in atypical regions (e.g., finger pads), diagnosis becomes difficult. Raman spectroscopy can noninvasively diagnose these tender lesions and examine many of them in a few seconds. Recently, Raman spectroscopy was used in vivo to detect a foreign body reaction from trauma induced by a paddle.[74]

Although it is thought that melasma is due to excessive melanin production, an alternate explanation for the excessive pigmentation is changes in melanin's molecular structure and concentration in the stratum corneum. Raman spectra of the stratum corneum showed no significant difference in the concentration of melanin in patients with melasma and those with normal skin. Thus, in patients with melasma, the melanin may be concentrated in deeper skin layers.[75] The spectra in some of the melasma patients also showed degraded melanin, which may explain the variability in therapeutic success for this pigmentary disorder.

Raman spectroscopy has accurately measured analytes of blood and urine samples.[76] Caspers et al.[11] have taken high-quality Raman spectra of dermal capillaries in vivo. This technique can monitor blood analytes such as glucose. Moreover, Raman examination of blood can identify pathogenic microorganisms. Raman spectroscopy has had success in surveillance for methicillin-resistant Staphylococcus aureus.[77] Combining Raman spectroscopy with optical tweezers and fluorescence in situ has advanced the detection and characterization of bacteria.[7879] More recently, dermatophytic and nondermatophytic agents of onychomycosis growing on ex vivo human nail could be distinguished by Raman spectroscopy.[80]

Conclusion

Capable of providing a detailed analysis of chemical composition, Raman spectroscopy is gaining recognition as a direct method in disease diagnosis. With further studies on the horizon, it is evident that Raman spectroscopy will play a role in helping define dermatologic conditions. A collective effort is, therefore, necessary to gather and analyze the optical fingerprints of various pathologies. Cost-effective, noninvasive and reliable, Raman spectroscopy meets the criteria for a model diagnostic instrument. Raman spectroscopy provides objective analyses of skin lesions; thus, improving both disease diagnosis and management.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

What is new?

Raman spectroscopy can classify melanoma, basal cell carcinoma and squamous cell carcinoma with good sensitivity and specificity

Raman spectroscopy has been applied to study changes in psoriasis, vitiligo, gout and melasma.