Melanoma Diagnosis by Raman Spectroscopy and Neural Networks: Structure Alterations in Proteins and Lipids in Intact Cancer Tissue
2004; Elsevier BV; Volume: 122; Issue: 2 Linguagem: Inglês
10.1046/j.0022-202x.2004.22208.x
ISSN1523-1747
AutoresMonika Gniadecka, Peter A. Philipsen, Sonja Wessel, Robert Gniadecki, Hans Christian Wulf, Sigurður Sigurðsson, O. Faurskov Nielsen, Daniel H. Christensen, Jana Hercogová, Kristian Rossen, Henrik Klem Thomsen, Lars Kai Hansen,
Tópico(s)Microbial Inactivation Methods
ResumoMelanoma is the most aggressive skin cancer. The specificity and sensitivity of clinical diagnosis varies from around 40% to 80%. Here, we investigated whether the chemical changes in the melanoma tissue detected by Raman spectroscopy and neural networks can be used for diagnostic purposes. Near-infrared Fourier transform Raman spectra were obtained from samples of melanoma (n=22) and other skin tumors that can be clinically confused with melanoma: pigmented nevi (n=41), basal cell carcinoma (n=48), seborrheic keratoses (n=23), and normal skin (n=89). A sensitivity analysis of spectral frequencies used by a neural network was performed to determine the importance of the individual components in the Raman spectra. Visual inspection of the Raman spectra suggested that melanoma could be differentiated from pigmented nevi, basal cell carcinoma, seborrheic keratoses, and normal skin due to the decrease in the intensity of the amide I protein band around 1660 cm-1. Moreover, melanoma and basal cell carcinoma showed an increase in the intensity of the lipid-specific band peaks around 1310 cm-1 and 1330 cm-1, respectively. Band alterations used in the visual inspection were also independently identified by a neural network for melanoma diagnosis. The sensitivity and specificity for diagnosis of melanoma achieved by neural network analysis of Raman spectra were 85% and 99%, respectively. We propose that neural network analysis of near-infrared Fourier transform Raman spectra could provide a novel method for rapid, automated skin cancer diagnosis on unstained skin samples. Melanoma is the most aggressive skin cancer. The specificity and sensitivity of clinical diagnosis varies from around 40% to 80%. Here, we investigated whether the chemical changes in the melanoma tissue detected by Raman spectroscopy and neural networks can be used for diagnostic purposes. Near-infrared Fourier transform Raman spectra were obtained from samples of melanoma (n=22) and other skin tumors that can be clinically confused with melanoma: pigmented nevi (n=41), basal cell carcinoma (n=48), seborrheic keratoses (n=23), and normal skin (n=89). A sensitivity analysis of spectral frequencies used by a neural network was performed to determine the importance of the individual components in the Raman spectra. Visual inspection of the Raman spectra suggested that melanoma could be differentiated from pigmented nevi, basal cell carcinoma, seborrheic keratoses, and normal skin due to the decrease in the intensity of the amide I protein band around 1660 cm-1. Moreover, melanoma and basal cell carcinoma showed an increase in the intensity of the lipid-specific band peaks around 1310 cm-1 and 1330 cm-1, respectively. Band alterations used in the visual inspection were also independently identified by a neural network for melanoma diagnosis. The sensitivity and specificity for diagnosis of melanoma achieved by neural network analysis of Raman spectra were 85% and 99%, respectively. We propose that neural network analysis of near-infrared Fourier transform Raman spectra could provide a novel method for rapid, automated skin cancer diagnosis on unstained skin samples. melanoma near-infrared Fourier transform pigmented nevi seborrheic keratosis Melanoma (MM) is the most aggressive of skin cancers and is invariably fatal if left untreated (MacKie, 2000MacKie R.M. Malignant melanoma: Clinical variants and prognostic indicators.Clin Exp Dermatol. 2000; 25: 471-475Crossref PubMed Scopus (47) Google Scholar;Marks, 2000Marks R. Epidemiology of malignant melanoma.Clin Exp Dermatol. 2000; 25: 459-463Crossref PubMed Scopus (241) Google Scholar;Martinez and Otley, 2001Martinez J.C. Otley C.C. The management of melanoma and nonmelanoma skin cancer: A review for the primary care physician.Mayo Clin Proc The. 2001; 76: 1253-1265Abstract Full Text Full Text PDF PubMed Scopus (98) Google Scholar). MM removal at early stages is almost always curative and therefore early detection is essential. Clinical diagnostic sensitivity differs greatly depending on the length of training of the clinician: 80% for trained dermatologists, 62% for senior registrars, 56% for registrars, and approximately 40% for nondermatologists (Cassileth et al., 1986Cassileth B.R. Clark Jr., W.H. Lusk E.J. Frederik B.E. Thomson C.J. Walsh W.P. How well do physicians recognize melanoma and other problem lesions?.J Am Acad Dermatol. 1986; 14: 555-560Abstract Full Text PDF PubMed Scopus (181) Google Scholar;Morton and MacKie, 1998Morton C.A. MacKie R.M. Clinical accuracy of the diagnosis of cutaneous malignant melanoma.Br J Dermatol. 1998; 138: 283-287Crossref PubMed Scopus (171) Google Scholar;Chen et al., 2001Chen S.C. Bravata D.M. Weil E. Olkin I. A comparison of dermatologists’ and primary care physician accuracy in diagnosing melanoma: A systemic review.Arch Dermatol. 2001; 137: 1627-1634Crossref PubMed Google Scholar). Difficulties in diagnosis of cutaneous MM arise because benign lesions like pigmented nevi (PN), seborrheic keratosis (SK), and other types of skin cancer such as basal cell carcinoma (BCC) may resemble melanoma (MacKie, 2000MacKie R.M. Malignant melanoma: Clinical variants and prognostic indicators.Clin Exp Dermatol. 2000; 25: 471-475Crossref PubMed Scopus (47) Google Scholar;Kanzler and Mraz-Genrnhard, 2001Kanzler M.H. Mraz-Genrnhard S. Primary cutaneous malignant melanoma and its precursor lesions: Diagnostic and therapeutic overview.J Am Acad Dermatol. 2001; 45: 260-276Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar). Histopathologic examination of the excised suspicious element is considered to be the golden standard (Slater, 2000Slater D.N. Doubt and uncertainty in the diagnosis of melanoma.Histopathology. 2000; 37: 469-472Crossref PubMed Google Scholar). Removal of every pigmented lesion is unacceptable for the patient, however, especially in the case of multiple skin lesions or lesions localized in cosmetically important parts of the body such as the face because of risk of scarring. Eighty percent of biopsies taken by nondermatologists of suspected malignant skin lesions have been reported to be benign (Jones et al., 1996Jones T.P. Boiko P.E. Piepkorn M.W. Skin biopsy indications in primary care practice: A population based study.J Am Board Fam Pract. 1996; 9: 397-404PubMed Google Scholar) and thus inappropriate surgery is frequent (Cox et al., 1992Cox N.H. Wagstaff R. Popple A.W. Using clinicopathological analysis of general practitioner skin surgery to determine educational requirements and guidelines.Br Med J. 1992; 304: 93-96Crossref PubMed Scopus (38) Google Scholar). Development of noninvasive diagnostic methods is therefore crucial for MM detection. Despite many attempts to implement various instrumental techniques (total body photography, dermatoscopy, high-frequency skin ultrasonography, fluorescence spectroscopy, positron emission tomography, etc.) a noninvasive, reliable method for MM diagnosis has not yet been established (Harland et al., 2000Harland C.C. Kale S.G. Jackson P. Mortimer P.S. Bamber J.C. Differentiation of common benign pigmented skin lesions from malignant melanoma by high-resolution ultrasound.Br J Dermatol. 2000; 143: 281-289Crossref PubMed Scopus (84) Google Scholar;Bafounta et al., 2001Bafounta M.L. Beauchet A. Aegerter P. Saiag P. Is dermatoscopy (epiluminescence microscopy) useful for the diagnosis of melanoma? Results of a meta-analysis using techniques adapted to the evaluation of diagnostic tests.Arch Dermatol. 2001; 137: 1343-1350Crossref PubMed Google Scholar;Kittler et al., 2002Kittler Pehamberger H. Wolff K. Binder M. Diagnostic accuracy of dermatoscopy.Lancet Oncol. 2002; 3: 159-165Abstract Full Text Full Text PDF PubMed Scopus (817) Google Scholar;Prichard et al., 2002Prichard R.S. Hill A.D. Skehan S.J. O'Higgins N.J. Positron emission tomography for staging and management of malignant melanoma.Br J Surg. 2002; 89: 389-396Crossref PubMed Scopus (73) Google Scholar). Raman spectroscopy is a technique that provides information about the molecular structure of the investigated sample and has been widely used for the past 70 y for nondestructive chemical analysis (Raman and Krishanan, 1928Raman C.V. Krishanan K.S. A new type of of secondary radiation.Nature. 1928; 121: 501-502Crossref Scopus (858) Google Scholar;Edwards et al., 1995Edwards H.G.M. Williams A.C. Barry B.W. Potential applications of FT-Raman Spectroscopy for dermatological diagnosis.J Mol Struct. 1995; 347: 379-388Crossref Scopus (67) Google Scholar;Hanlon et al., 2000Hanlon E.B. Manoharan R. Koo T.W. et al.Prospects for in vivo Raman spectroscopy.Phys Med Biol. 2000; 45: R1-R59Crossref PubMed Scopus (721) Google Scholar;Puppels, 2001Puppels G.J. Medical applications of Raman spectroscopy: From proof of principle to clinical implementation.Biopolymers. 2001; 67: 1-9Google Scholar). In Raman spectroscopy a sample is irradiated with laser light, which results in light scattering. The majority of scattered light has unchanged frequency (so-called Rayleigh line) whereas the rest is shifted in frequency (Raman effect). The frequency shifts can be analyzed and presented as a Raman spectrum. The Raman effect is caused by molecular vibrations in the irradiated sample and thus gives information about the structure of the molecules. Molecular vibrations can be divided into deformational vibrations where the angles between the bonds change, and stretching vibrations where the lengths of the bonds change but the angle between them remains constant. Previous studies of samples obtained from the genital tract, brain, breast, and larynx revealed that the transition from normal to cancer tissue is associated with significant differences in chemical structure, which are reflected in the Raman spectra (Liu et al., 1992Liu C.H. Das B.B. ShaGlassman W.L. et al.Raman, fluorescence, and time-resolved light scattering as optical diagnostic techniques to separate diseased and normal biomedical media.J Photochem Photobiol B - Biol. 1992; 16: 187-209Crossref PubMed Scopus (203) Google Scholar; Mizuno et al, 1994; Hanlon et al, 2000; Stone et al, 2000). Benign and malignant changes in the gastrointestinal tract have recently been studied intraoperatively in vivo with the aid of Raman spectroscopy (Shim et al., 2000Shim M.G. Song L.M. Marcon N.E. Wilson B.C. In vivo near-infrared Raman spectroscopy: Demonstration of feasibility during clinical gastrointestinal endoscopy.Photochem Photobiol. 2000; 72: 146-150PubMed Google Scholar;Dacosta et al., 2002Dacosta R.S. Wilson B.C. Marcon N.E. New optical technologies for earlier endoscopic diagnosis of premalignant gastrointestinal lesions.J Gastroenterol Hepatol. 2002; 17: S85-S104Crossref PubMed Scopus (139) Google Scholar). Our earlier studies have shown that samples of various benign skin lesions and nonmelanoma skin cancer have characteristic Raman spectra (Gniadecka et al., 1997aGniadecka M. Wulf H.C. Nymark Mortensen N. Faurskov Nielsen O. Christensen D.H. Diagnosis of basal cell carcinoma by Raman spectroscopy.J Raman Spectrosc. 1997; 28: 125-129Crossref Google Scholar;Gniadecka et al., 1997bGniadecka M. Wulf H.C. Nielsen O.F. Christensen D.H. Hercogova J. Distinctive molecular abnormalities in benign and malignant skin lesions: Studies by Raman spectroscopy.Photochem Photobiol. 1997; 66: 418-423Crossref PubMed Scopus (139) Google Scholar;Gniadecka et al., 1998aGniadecka M. Wulf H.C. Nielsen O.F. Christensen D.H. Hercogova J. Rossen K. Potential of Raman spectroscopy for in vitro and in vivo diagnosis of malignant melanoma.in: Heyns E.M. Proceedings of the XVI International Conference on Raman Spectroscopy. John Wiley and Sons, Chichester1998: 764-765Google Scholar) and that this technique is highly reproducible (Gniadecka et al., 1998bGniadecka M. Faurskov Nielsen O. Christensen D.H. Wulf H.C. Structure of water, protein, and lipids in intact human skin, hair, and nail.J Invest Dermatol. 1998; 110: 393-398Crossref PubMed Scopus (268) Google Scholar). Recent studies on BCC by Raman spectroscopy showed that it is feasible to differentiate malignant tissue from the healthy surrounding tissue (Nijssen et al., 2002Nijssen A. Bakker Schut T.C. Heule F. Caspers P.J. Hayes D.P. Neumann M.H.A. Puppels G.J. Discriminating basal cell carcinoma from its surrounding tissue by Raman spectroscopy.J Invest Dermatol. 2002; 119: 64-69Crossref PubMed Scopus (311) Google Scholar). Raman spectra are complex and the traditional interpretation of the spectra by their visual inspection is subjective and time consuming. Artificial neural networks are a pattern recognition tool for modeling nonlinear functional relationships in data. Neural networks may be applied to problems where the relationship is complex or unknown. Automated computer methods like principal component analysis and neural networks have been developed in the last decades for many different applications in biomedical research (Hanlon et al., 2000Hanlon E.B. Manoharan R. Koo T.W. et al.Prospects for in vivo Raman spectroscopy.Phys Med Biol. 2000; 45: R1-R59Crossref PubMed Scopus (721) Google Scholar; Christoyianni et al, 2002; Dacosta et al, 2002). Recent studies have shown that neural networks can be successfully used for analysis of dermatoscopic images (Rubegni et al., 2002aRubegni P. Burroni M. Cevenini G. et al.Digital dermoscopy analysis and artificial neural network for the differentiation of clinically atypical pigmented skin lesions: A retrospective study.J Invest Dermatol. 2002; 119: 471-474Crossref PubMed Scopus (103) Google Scholar;Rubegni et al., 2002bRubegni P. Cevenini G. Burroni M. et al.Automated diagnosis of pigmented skin lesions.Int J Cancer. 2002; 101: 576-580Crossref PubMed Scopus (143) Google Scholar). Here we have developed a neural network system for the automated classification of Raman spectra, allowing us to differentiate MM from other clinically similar skin tumors. First, all Raman spectra were preprocessed (see Patients and Methods). Then, examples of Raman spectra were used to train the neural network for learning the spectral patterns. The performance of the neural network was tested on an independent set of spectra, by prediction of lesion type and comparing it to the true class. Moreover, the importance of specific spectral bands for the classification of individual classes could be evaluated, thus making it possible to extract the most meaningful parts of spectra important for the neural network classification. The spectra of normal skin were similar to those previously reported (Barry et al., 1992Barry B.W. Edwards H.G.M. Williams A.C. Fourier transform Raman and infrared vibrational study of human skin: Assignment of spectral bands.J Raman Spectrosc. 1992; 23: 641-645Crossref Scopus (208) Google Scholar;Edwards et al., 1995Edwards H.G.M. Williams A.C. Barry B.W. Potential applications of FT-Raman Spectroscopy for dermatological diagnosis.J Mol Struct. 1995; 347: 379-388Crossref Scopus (67) Google Scholar;Gniadecka et al., 1998bGniadecka M. Faurskov Nielsen O. Christensen D.H. Wulf H.C. Structure of water, protein, and lipids in intact human skin, hair, and nail.J Invest Dermatol. 1998; 110: 393-398Crossref PubMed Scopus (268) Google Scholar;Gniadecka et al., 1998cGniadecka M. Faurskov Nielsen O. Christensen D.H. Wessel S. Heidenheim M. Wulf H.C. Water and protein structure in photoaged and chronically aged skin.J Invest Dermatol. 1998; 111: 1129-1133Crossref PubMed Scopus (167) Google Scholar). Peaks originating from vibrations within protein molecules and lipids were clearly resolved (Figure 1, Table I). In normal skin the positions of the amide I and III bands (1650 cm-1 and 1270 cm-1) and the presence of a well-developed band around 935 cm-1 suggested a helical protein structure. Pigmentation increased the background level of the spectra, which was most pronounced in the region 1800–2500 cm-1 as previously described for normal skin (Knudsen et al., 2002Knudsen L. Johansson C.K. Philipsen P.A. Gniadecka M. Wulf H.C. Natural variations and reproducibility of in vivo near-infrared Fourier transform Raman spectroscopy of normal human skin.J Raman Spectrosc. 2002; 33: 574-579Crossref Scopus (29) Google Scholar) Figure 2. This region does not contain protein and lipid bands.Table IMajor protein and lipid band positions of NIR-FT Raman spectra of normal skin (NOR), PN, MM, BCC, and SKAmide I in proteinsAmide III in proteinsδ(CH2)(CH3) in proteins and lipidsNOR1662 (1656–1668)1264 (1241–1287)1451 (1449–1453)PN1660 (1652–1668)aSignificant compared to NOR.1264 (1242–1286)1451 (1448–1454)MM1655 (1647–1663)aSignificant compared to NOR.1272 (1258–1286)aSignificant compared to NOR.,bSignificant compared to BCC.1449 (1443–1455)aSignificant compared to NOR.BCC1655 (1651–1659)aSignificant compared to NOR.1268 (1253–1283)1449 (1445–1453)aSignificant compared to NOR.SK1653 (1649–1657)aSignificant compared to NOR.1276 (1266–1286)aSignificant compared to NOR.1446 (1441–1451)aSignificant compared to NOR.Values represent means with 95% confidence intervals.a Significant compared to NOR.b Significant compared to BCC. Open table in a new tab Figure 2NIR-FT Raman spectra of normal skin (NOR), PN, MM, BCC, and SK. Note the influence of the pigmentation on the 1800–2500 cm-1 region.View Large Image Figure ViewerDownload (PPT) Values represent means with 95% confidence intervals. Spectra revealed clear-cut changes between the investigated types of tumors allowing differentiation of MM from other investigated lesions (Figure 1, Table I, Table II). MM presented with a decrease in the amide I band intensity resulting in flattening of the spectral area between 1500 cm-1 and 1800 cm-1 (Figure 1a, Table II). This suggested a change in molecular composition of proteins. Moreover, an increase in intensity of the band at 1300 cm-1 of the CH twisting and wagging in lipids and the band around 1310–1330 cm-1 was seen with a relative decrease in the amide III regions of proteins (Figure 1b, Table II).Table IIMajor band ratios in NIR-FT Raman spectra of normal skin (NOR), PN, MM, BCC, and SKAmide I in proteins, I1650 cm-1/I1450 cm-1Amide III in proteins, I1270 cm-1/I1320 cm-1C–C in proteins, I930 cm-1/I1000 cm-1NOR0.69 (0.54–0.84)1.15 (0.96–1.34)2.49 (1.24–3.74)PN0.62 (0.39–0.85)aSignificant compared to NOR.1.05 (0.98–1.12)aSignificant compared to NOR.2.24 (0.64–3.84)aSignificant compared to NOR.MM0.29 (-0.27–0.85)aSignificant compared to NOR.,bSignificant compared to BCC.0.99 (0.93–1.05)aSignificant compared to NOR.,bSignificant compared to BCC.0.77 (-0.29–1.83)aSignificant compared to NOR.BCC0.86 (0.51–1.21)aSignificant compared to NOR.0.94 (0.86–1.02)aSignificant compared to NOR.0.82 (0.10–1.54)aSignificant compared to NOR.SK0.81 (0.33–1.29)aSignificant compared to NOR.0.95 (0.82–1.08)aSignificant compared to NOR.0.85 (0.14–1.56)aSignificant compared to NOR.Values represent means with 95% confidence intervalsa Significant compared to NOR.b Significant compared to BCC. Open table in a new tab Values represent means with 95% confidence intervals BCC showed similar spectral changes in the region 1300–1340 cm-1 (Figure 1b, Table II). The emerging bands in this region were found around 1300 cm-1 and 1330 cm-1. Both MM and BCC spectra showed a decrease in protein band intensity around 940 cm-1 Figure 1c. In contrast to MM, BCC spectra did not show any decrease in intensity of the amide I band Figure 1a. SK presented with a prominent increase in intensity of the 1300 cm-1 band, making the discrimination from BCC possible Figure 1b. The band around 1450 cm-1 reflecting (CH2)(CH3) in lipids and proteins was shifted in MM, BCC, and SK (Figure 1d, Table I). Spectra of PN were almost identical to those of normal skin but consistently showed a decrease in intensity in the right wing of the amide I band (Figure 1a, Table II). This suggested minor changes in composition or conformational alterations in the proteins. Moreover, in the region from 1800 cm-1 to 2500 cm-1, an increase in intensity of the otherwise flat spectral region was observed Figure 2. This phenomenon occurred most probably due to an increase in pigmentation of the lesional skin, similar to that of highly pigmented skin and in MM spectra. Neural network analysis of Raman spectra achieved a diagnostic sensitivity of 85% and specificity of 99% for the diagnosis of MM Table III. BCC diagnosis was 97% and 98%, respectively. The confusion map, established by the neural network, presenting wrongly classified skin lesions is presented in Table IV. MM had a tendency to be misdiagnosed as BCC and PN, whereas PN were confused with normal skin. SK was diagnosed with 96% sensitivity and was most frequently confused with BCC. PN were diagnosed with 78% sensitivity.Table IIISpecificity and sensitivity of neural network analysis of Raman spectra for classification of normal skin (NOR), PN, MM, BCC, and SKSensitivitySpecificityNOR96%±394%±3PN78%±697%±2MM85%±599%±1BCC98%±298%±2SK96%±3100%±0Sensitivity and specificity are defined as inAltman, 1991Altman D.G. Practical Statistics for Medical Research. Chapman and Hall, London1991: 409-419Google Scholar and are shown±twice the SD of the approximated normal distribution to the binomial distribution. Open table in a new tab Table IVConfusion map of the neural network predictions of normal skin (NOR), PN, MM, BCC, and SKNORPNMMBCCSKNOR*96%20%0%0%0%PN*3%78%10%2%0%MM*0%2%85%0%0%BCC*1%0%5%98%4%SK*0%0%0%0%96%The top row indicates the correct labels, and the far left column (marked with asterisks) indicates network predictions. The confusion map presents the probability of different classes estimated by the neural network given the true class label. Open table in a new tab Sensitivity and specificity are defined as inAltman, 1991Altman D.G. Practical Statistics for Medical Research. Chapman and Hall, London1991: 409-419Google Scholar and are shown±twice the SD of the approximated normal distribution to the binomial distribution. The top row indicates the correct labels, and the far left column (marked with asterisks) indicates network predictions. The confusion map presents the probability of different classes estimated by the neural network given the true class label. The sensitivity map for MM in Figure 3 showed important spectral intensities at 1620–1670 cm-1, 1230–1300 cm-1, and 1430–1450 cm-1, which matched the spectral differentiation marked A, B, and D in Figure 1. Moreover, the sensitivity map pointed out new frequencies not used for visual inspection of the spectra: 2840–3000 cm-1, 2000–2350 cm-1, and 1000 cm-1, marked E, F, and G respectively in Figure 2. Vibrations reflected at these spectral regions are most probably characteristic for (E) CH stretching in proteins and lipids (Gniadecka et al., 1998aGniadecka M. Wulf H.C. Nielsen O.F. Christensen D.H. Hercogova J. Rossen K. Potential of Raman spectroscopy for in vitro and in vivo diagnosis of malignant melanoma.in: Heyns E.M. Proceedings of the XVI International Conference on Raman Spectroscopy. John Wiley and Sons, Chichester1998: 764-765Google Scholar), (F) skin fluorescence (Knudsen et al., 2002Knudsen L. Johansson C.K. Philipsen P.A. Gniadecka M. Wulf H.C. Natural variations and reproducibility of in vivo near-infrared Fourier transform Raman spectroscopy of normal human skin.J Raman Spectrosc. 2002; 33: 574-579Crossref Scopus (29) Google Scholar), and (G) ring vibrations in aromatic amino acid residues (Gniadecka et al., 1997aGniadecka M. Wulf H.C. Nymark Mortensen N. Faurskov Nielsen O. Christensen D.H. Diagnosis of basal cell carcinoma by Raman spectroscopy.J Raman Spectrosc. 1997; 28: 125-129Crossref Google Scholar). We have demonstrated that it is possible to differentiate MM from PN, BCC, and SK by neural network analysis of the near-infrared Fourier transform (NIR-FT) Raman spectra. The neural network analysis used spectral alterations reflecting changes in composition and structure of proteins and lipids. MM and BCC spectra showed similar patterns but were distinctive. Intensities in the amide III regions of proteins decreased relative to the lipid-specific bands. An emergence of the lipid component of the δ (CH2)(CH3) accounting for shifting of the 1450 cm-1 band occurred both in BCC and MM spectra. For SK, a major increase in intensity of the 1300 cm-1 band of twisting and wagging vibrations of lipids reflected lipid accumulation in the horny cysts of SK. The first Raman spectroscopy study on cancer recognition was performed on uterus, ovary, and cervix cancer tissue where the decrease in intensity of the amide III region was reported (Liu et al., 1992Liu C.H. Das B.B. ShaGlassman W.L. et al.Raman, fluorescence, and time-resolved light scattering as optical diagnostic techniques to separate diseased and normal biomedical media.J Photochem Photobiol B - Biol. 1992; 16: 187-209Crossref PubMed Scopus (203) Google Scholar). This phenomenon has also been observed for laryngeal cancer, where additionally increase in the lipid bands around 1310–1340 cm-1 was found (Stone et al., 2000Stone N. Tavroulaki P. Kendall C. Birchall M. Barr H. Raman spectroscopy for early detection of laryngeal malignancy: Preliminary results.Laryngoscope. 2000; 110: 1756-1763Crossref PubMed Scopus (192) Google Scholar). The peak around 1320 cm-1 can be attributed to both collagen and nucleic acids (DNA), more specifically to the purine bases guanine and adenine (Stone et al., 2000Stone N. Tavroulaki P. Kendall C. Birchall M. Barr H. Raman spectroscopy for early detection of laryngeal malignancy: Preliminary results.Laryngoscope. 2000; 110: 1756-1763Crossref PubMed Scopus (192) Google Scholar). Raman spectra obtained from MM and BCC show similar alterations in the 1310–1340 cm-1 region. Therefore it is most probable that malignant transformation triggers similar molecular changes independently of the tissue involved. Alterations of the amide bands in Raman spectra are attributed to the conformational changes of proteins. Collagen contributes predominantly to Raman spectra of the skin (Johansson et al., 2000Johansson C.K. Gniadecka M. Ullman S. Halberg P. Kobayasi T. Wulf H.C. Alterations in collagen structure in hypermobility and Ehlers–Danlos syndromes detected by Raman spectroscopy in vivo. Optical biopsy and tissue optics.Prog Biomed Optics Imaging. 2000; 1: 138-143Google Scholar) and therefore changes in its structure or its degradation could explain the decrease in the amide band intensities. Collagen degradation is influenced by matrix metalloproteinases (Yoshizaki et al., 2002Yoshizaki T. Sato H. Furukawa M. Recent advances in the regulation of matrix metalloproteinases activation: From basic research to clinical implication (Review).Oncol Rep. 2002; 9: 607-6011PubMed Google Scholar). Upregulation of metalloproteinases was found in invasive skin tumors like MM and BCC (Hofmann et al., 2000Hofmann U.B. Westphal J.R. Van Muijen G.N. Ruiter D.J. Matrix metalloproteinases in human melanoma.J Invest Dermatol. 2000; 115: 337-344Crossref PubMed Scopus (329) Google Scholar;Varani et al., 2000Varani J. Hattori Y. Chi Y. Schmidt T. Perone P. Zeigler M.E. Fader Johnson T.M. Collagenolytic and gelatinolytic matrix metalloproteinases and their inhibitors in basal cell carcinoma of skin: Comparison with normal skin.Br J Cancer. 2000; 82: 657-665Crossref PubMed Scopus (74) Google Scholar) but also in endometrial and laryngeal carcinoma (Magary et al., 2000Magary S.P. Ryan M.W. Tarnuzzer R.W. Kornberg L. Expression of matrix metalloproteinases and tissue inhibitor metalloproteinases in laryngeal and pharyngeal squamous cell carcinoma: A quantitative analysis.Otolaryngol Head Neck Surg. 2000; 122: 712-716PubMed Google Scholar;Di Nezza et al., 2002Di Nezza L.A. Misajon A. Zhang J. et al.Presence of active gelatinises in endometrial carcinoma and correlation of matrix metalloproteinase expression with increasing tumor grade and invasion.Cancer. 2002; 94: 1466-1475Crossref PubMed Scopus (139) Google Scholar). Cancer–stroma interactions (Ruiter et al., 2002aRuiter D. Bogenrieder T. Elder D. Herlyn M. Melanoma–stroma interactions: Structural and functional aspects.Lancet Oncol. 2002; 3: 35-43Abstract Full Text Full Text PDF PubMed Scopus (189) Google Scholar), where ground substance and fibroblasts are affected, are most probably reflected in the alterations of the protein, lipid, and water structure detected by Raman spectroscopy (Gniadecka et al, 2003). This is the first report where neural network analysis of Raman spectra has been employed for MM diagnosis. We have previously used this method for BCC diagnosis (Gniadecka et al., 1997aGniadecka M. Wulf H.C. Nymark Mortensen N. Faurskov Nielsen O. Christensen D.H. Diagnosis of basal cell carcinoma by Raman spectroscopy.J Raman Spectrosc. 1997; 28: 125-129Crossref Google Scholar). Neural networks and other computer methods like principal component analysis have been successfully used as an aid in the detection of gastrointestinal tract malignancies and breast cancer (Hanlon et al., 2000Hanlon E.B. Manoharan R. Koo T.W. et al.Prospects for in vivo Raman spectroscopy.Phys Med Biol. 2000; 45: R1-R59Crossref PubMed Scopus (721) Google Scholar;Christoyianni et al., 2002Christoyianni I. Koutras A. Dermatas E. Kokkanikis G. Computer aided diagnosis of breast cancer in digitized mommograms.Comput Med Imaging Graph. 2002; 26: 309Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar;Dacosta et al., 2002Dacosta R.S. Wilson B.C. Marcon N.E. New optical technologies for earlier endoscopic diagnosis of premalignant gastr
Referência(s)