Skin cancer diagnosis using NIR spectroscopy data of skin lesions in vivo using machine learning algorithms

• Autorzy: Rocha Matheus B. , Loss Flavio P. , da Cunha Pedro H. , Zanoni Madson Poltronieri , de Lima Leandro M. , Nascimento Isadora Tavares , Rezende Isabella , Canuto Tania R.P. , de Paula Vieira Luciana , Rossoni Renan , Santos Maria C.S. , Frasson Patricia Lyra , Romão Wanderson , Filgueiras Paulo R. , Krohling Renato A.

Identyfikatory

DOI

10.1016/j.bbe.2024.10.001

• Języki publikacji: EN

Abstrakty

EN

Skin lesions are classified in benign or malignant. Among the malignant, melanoma is a very aggressive cancer and the major cause of deaths. So, early diagnosis of skin cancer is very desired. In the last few years, there is a growing interest in computer aided diagnostic (CAD) of skin lesions. Near-Infrared (NIR) spectroscopy may provide an alternative source of information to automated CAD of skin lesions to be used with the modern techniques of machine learning and deep learning (MDL). One of the main limitations to apply MDL to spectroscopy is the lack of public datasets. Since there is no public dataset of NIR spectral data to skin lesions, as far as we know, an effort has been made and a new dataset named NIR-SC-UFES, has been collected, annotated and analyzed generating the gold-standard for classification of NIR spectral data to skin cancer. Next, the machine learning algorithms XGBoost, CatBoost, LightGBM, 1D-convolutional neural network (1D-CNN) and standard algorithms as SVM and PLS-DA were investigated to classify cancer and non-cancer skin lesions. Experimental results indicate that the best performance was obtained by LightGBM with pre-processing using standard normal variate (SNV), feature extraction and data augmentation with Generative Adversarial Networks (GAN) providing values of 0.839 for balanced accuracy, 0.851 for recall, 0.852 for precision, and 0.850 for F-score. The obtained results indicate the first steps in CAD of skin lesions aiming the automated triage of patients with skin lesions in vivo using NIR spectral data.

Słowa kluczowe

EN

machine learning algorithm; 1D convolutional neural network; skin cancer; near infrared spectroscopy; NIR

PL

algorytm uczenia maszynowego; sieć neuronowa konwolucyjna; rak skóry; spektroskopia bliskiej podczerwieni; NIR

• Wydawca: Nałęcz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of Sciences ; Elsevier
• Czasopismo: Biocybernetics and Biomedical Engineering
• Rocznik: 2024
• Tom: Vol. 44, no. 4
• Strony: 824--835
• Opis fizyczny: Bibliogr. 56 poz., rys., tab., wykr.

Twórcy

autor

Rocha Matheus B.

• Labcin - Nature Inspired Computing Lab, Vitória, Brazil
• PPGI - Graduate Program in Computer Science, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

Loss Flavio P.

• Labcin - Nature Inspired Computing Lab, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

da Cunha Pedro H.

• PPGQ - Graduate Program in Chemistry, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

Zanoni Madson Poltronieri

• Department of Chemistry, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

de Lima Leandro M.

• Labcin - Nature Inspired Computing Lab, Vitória, Brazil
• PPGI - Graduate Program in Computer Science, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

Nascimento Isadora Tavares

• Dermatological Assistance Program (PAD), Vitória, Brazil

autor

Rezende Isabella

• Dermatological Assistance Program (PAD), Vitória, Brazil

autor

Canuto Tania R.P.

• Dermatological Assistance Program (PAD), Vitória, Brazil
• Secretary of Health of the Espírito Santo State, Vitória, Brazil

autor

de Paula Vieira Luciana

• Dermatological Assistance Program (PAD), Vitória, Brazil
• Secretary of Health of the Espírito Santo State, Vitória, Brazil

autor

Rossoni Renan

• Dermatological Assistance Program (PAD), Vitória, Brazil

autor

Santos Maria C.S.

• Pathological Anatomy Unit of the University Hospital Cassiano Antônio Moraes (HUCAM), Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

Frasson Patricia Lyra

• Department of Specialized Medicine, Vitória, Brazil
• Dermatological Assistance Program (PAD), Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

Romão Wanderson

• PPGQ - Graduate Program in Chemistry, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil
• Federal Institute of Espírito Santo (IFES), Vitória, ES, 29040-780, Brazil

autor

Filgueiras Paulo R.

• PPGQ - Graduate Program in Chemistry, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

autor

Krohling Renato A.

• Labcin - Nature Inspired Computing Lab, Vitória, Brazil
• PPGI - Graduate Program in Computer Science, Vitória, Brazil
• Federal University of Espírito Santo (UFES), Vitória, ES, 29075-910, Brazil

Bibliografia

• [1] Marzuka AG, Book SE. Basal cell carcinoma: pathogenesis, epidemiology, clin-ical features, diagnosis, histopathology, and management. Yale J Biol Med 2015;88(2):167-79.
• [2] Gallagher RP, Lee TK. Adverse effects of ultraviolet radiation: A brief review.Prog Biophys Mol Biol 2006;92(1):119-31, UV exposure guidance: A balanced approach between health risks and health benefits of UV and Vitamin D. Proceedings of an International Workshop, International Commission on Non-ionizing Radiation Protection, Munich, Germany, 17-18 October, 2005.
• [3] Hamidi R, Peng D, Cockburn M. Efficacy of skin self-examination for the early detection of melanoma. Int J Dermatol 2010;49(2):126-34.
• [4] Chen Q, Hong Y. AliFuse: Aligning and fusing multi-modal medical data for computer-aided diagnosis. 2024, ArXivabs/2401.01074.
• [5] Pacheco AG, Krohling RA. The impact of patient clinical information on automated skin cancer detection. Comput Biol Med 2020;116:103545.
• [6] Zakeri A, Hokmabadi A. Improvement in the diagnosis of melanoma and dysplastic lesions by introducing ABCD-PDT features and a hybrid classifier.Biocybern Biomed Eng 2018;38(3):456-66.
• [7] Cheong KH, Tang KJW, Zhao X, Koh JEW, Faust O, Gururajan R, Ciaccio EJ, Rajinikanth V, Acharya UR. An automated skin melanoma detection system with melanoma-index based on entropy features. Biocybern Biomed Eng2021;41(3):997-1012.
• [8] Zhao J, Zeng H, Kalia S, Lui H. Using Raman spectroscopy to detect and diagnose skin cancer in vivo. Dermatol Clin 2017;35(4):495-504.
• [9] Ferreira Lima AM, Daniel CR, Navarro RS, Bodanese B, Pasqualucci CA, Pacheco MTT, Zângaro RA, Silveira L. Discrimination of non-melanoma skin cancer and keratosis from normal skin tissue in vivo and ex vivo by Raman spectroscopy. Vib Spectrosc 2019;100:131-41.
• [10] Araújo DC, Veloso AA, de Oliveira Filho RS, Giraud M-N, Raniero LJ, Ferreira LM, Bitar RA. Finding reduced Raman spectroscopy finger print of skin samples for melanoma diagnosis through machine learning. Artif Intell Med2021;120:102161.
• [11] Raypah ME, Muncan J, Sudik S, Omar AF, Mail MH, Tsenkova R, Seeni A. Implication of phenol red in quantification of cultured cancerous cells using near-infrared spectroscopy and aquaphotomics. Chemometr Intell Lab Syst2022;230:104669.
• [12] CloudMinds Inc. CloudMinds AI Raman spectrometer MR-5S user guide. 2018,User Guide for the CloudMinds AI Raman Spectrometer MR-5S.
• [13] Herold B, Kawano S, Sumpf B, Tillmann P, Walsh KB. VIS/NIR spectroscopy. OptMonit Fresh Process Agric Crop 2009;141-249.
• [14] Sandorfy C, Buchet R, Lachenal G. Principles of molecular vibrations fornear-infrared spectroscopy. In: Near-infrared spectroscopy in food science and technology. John Wiley & Sons, Ltd; 2006, p. 11-46, Ch. 2.
• [15] Siesler HW. Basic principles of near-infrared spectroscopy. In: Handbook of near-infrared analysis. CRC Press; 2007, p. 25-38.
• [16] Ramirez CAM, Greenop M, Ashton L, ur Rehman I. Applications of machine learning in spectroscopy. Appl Spectrosc Rev 2021;56(8-10):733-63.
• [17] Ng W, Minasny B, de Sousa Mendes W, Demattê JAM. The influence of training sample size on the accuracy of deep learning models for the prediction of soil properties with near-infrared spectroscopy data. SOIL 2020;6(2):565-78.
• [18] McIntosh LM, Jackson M, Mantsch HH, Mansfield JR, Crowson A, Toole JW. Near-infrared spectroscopy for dermatological applications. Vib Spectrosc2002;28(1):53-8.
• [19] Gniadecka M, Philipsen PA, Wessel S, Gniadecki R, Wulf HC, Sigurdsson S, Nielsen OF, Christensen DH, Hercogova J, Rossen K, Thomsen HK, Hansen LK. Melanoma diagnosis by Raman spectroscopy and neural networks: Structure alterations in proteins and lipids in intact cancer tissue. J Invest Dermatol2004;122(2):443-9.
• [20] Liu J, Osadchy M, Ashton L, Foster M, Solomon CJ, Gibson SJ. Deep convolutional neural networks for Raman spectrum recognition: a unified solution. Analyst 2017;142(21):4067-74.
• [21] Malek S, Melgani F, Bazi Y. One-dimensional convolutional neural networks for spectroscopic signal regression. J Chemom 2017;32(5):e2977.
• [22] Morais CLM, Lima KMG, Singh M, Martin FL. Tutorial: multivariate classification for vibrational spectroscopy in biological samples. Nat Protoc2020;15(7):2143-62.
• [23] Zeng W, Wang Q, Xia Z, Li Z, Qu H. Application of XGBoost algorithmin the detection of SARS-CoV-2 using Raman spectroscopy. J Phys Conf Ser2021;1775(1):012007.
• [24] Stuart BH. Infrared spectroscopy: Fundamentals and applications. John Wiley &Sons; 2004.
• [25] Brereton RG. Pattern recognition in chemometrics. Chemometr Intell Lab Syst 2015;149:90-6.
• [26] Yuanyuan C, Zhibin W. Quantitative analysis modeling of infrared spectroscopy based on ensemble convolutional neural networks. Chemometr Intell Lab Syst2018;181:1-10.
• [27] Ke G, Meng Q, Finley T, Wang T, Chen W, Ma W, Ye Q, Liu T-Y. LightGBM:A highly efficient gradient boosting decision tree. Adv Neural Inf Process Syst2017;30.
• [28] Chen T, Guestrin CE. XGBoost: A scalable tree boosting system. Knowl Disc DataMin 2016;785-94.
• [29] Prokhorenkova L, Gusev G, Vorobev A, Dorogush AV, Gulin A. CatBoost:unbiased boosting with categorical features. 2019,arXiv:1706.09516.
• [30] Pereira da Cunha PH, Zanoni MP, Santos FD, Nascimento IT, Rezende I,Canuto TRP, Vieira LdP, Santos MCS, Romão W, Frasson PHL, Krohling R, Filgueiras PR. NIR-SC-UFES: A portable NIR spectral dataset to skin cancer invivo. 2024, URLhttps://data.mendeley.com/datasets/j9773cyr3k/1.
• [31] Pacheco AG, Lima, R G, Salomão AS, Krohling B, Biral IP, de Angelo GG, Alves Jr FC, Esgario JG, Simora, C A, Castro PB, et al. PAD-UFES-20: askin lesion dataset composed of patient data and clinical images collected from smartphones. Data Brief 2020;32:106221.
• [32] Huck CW. Bio-applications of NIR spectroscopy. In: Ozaki Y, Huck C, Tsuchikawa S, Engelsen SB, editors. Near-infrared spectroscopy: Theory, spectral analysis, instrumentation, and applications. Singapore: Springer Singapore; 2021,p. 413-35.
• [33] Fendel S, Schrader B. Investigation of skin and skin lesions by NIR-FT-Raman spectroscopy. Fresenius’ J Anal Chem 1998;360(5):609-13.
• [34] Smulko J, Wrobel MS, Barman I. Noise in biological Raman spectroscopy. In:International conference on noise and fluctuations. ICNF, IEEE; 2015, p. 1-6.
• [35] Barnes RJ, Dhanoa MS, Lister SJ. Standard normal variate transformation and de-trending of near-infrared diffuse reflectance spectra. Appl Spectrosc1989;43(5):772-7.
• [36] van der Maaten L, Hinton G. Visualizing data using t-SNE. J Mach Learn Res2008;9(86):2579-605.
• [37] Bergstra J, Bardenet R, Bengio Y, Kégl B. Algorithms for hyper-parameter optimization. Adv Neural Inf Process Syst 2011;24.
• [38] Akiba T, Sano S, Yanase T, Ohta T, Koyama M. Optuna: A next-generation hyperparameter optimization framework. In: Proceedings of the 25th ACMSIGKDD international conference on knowledge discovery & data mining. NewYork, NY, USA: Association for Computing Machinery; 2019, p. 2623-31.
• [39] Abayomi-Alli OO, Damaševičius R, Wieczorek M, Woźniak M. Data augmentation using principal component resampling for image recognition by deep learning. In: Artificial intelligence and soft computing: 19th international conference, ICAISC2020, zakopane, Poland, October 12-14, 2020, proceedings, part II 19. Springer;2020, p. 39-48.
• [40] Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP. SMOTE: Synthetic minority over-sampling technique. J Artificial Intelligence Res 2002;16:321-57.
• [41] Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio Y. Generative adversarial nets. Adv Neural Inf Process Syst2014;27.
• [42] Pavlou E, Kourkoumelis N. Deep adversarial data augmentation for biomedical spectroscopy: Application to modelling Raman spectra of bone. Chemometr Intell Lab Syst 2022;228:104634.
• [43] Friedman JH. Stochastic gradient boosting. Comput Statist Data Anal2002;38:367-78.
• [44] Hancock JT, Khosgoftaar TM. CatBoost for big data: an interdisciplinary review.J Big Data 2020;7:1-45.
• [45] Rumelhart DE, Hinton GE, Williams RJ. Learning representations by back-propagating errors. Nature 1986;323(6088):533-6.
• [46] Goodfellow I, Bengio Y, Courville A. Deep learning. MIT Press; 2016,http://www.deeplearningbook.org.
• [47] Nair V, Hinton GE. Rectified linear units improve restricted Boltzmann machines.In: International conference on machine learning. Madison, WI, USA: Omni Press;2010, p. 807-14.
• [48] Gottfries J, Blennow K, Wallin A, Gottfries C. Diagnosis of dementias using partial least squares discriminant analysis. Dementia geriatric cognit disorders. J Big Data 1995;6(2):83-8.
• [49] Boser BE, Guyon IM, Vapnik VN. A training algorithm for optimal margin classifiers. In: Proceedings of the 5th annual workshop on computational learning theory. 1992, p. 144-52.
• [50] Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay E. Scikit-learn: Machine learning in python. JMach Learn Res 2011;12:2825-30.
• [51] Lundberg SM, Lee S-I. A unified approach to interpreting model predictions. Advin Neural Inf Process Syst 2017;30(NIP 2017):4765-74.
• [52] Shin A, Cai Q, Shu X-O, Gao Y-T, Zheng W. Genetic polymorphisms in the matrix metalloproteinase 12 gene (MMP12) and breast cancer risk and survival: the Shanghai breast cancer study. Breast Cancer Res 2005;7(4):R506.
• [53] Peña-Martín J, Belén García-Ortega M, Palacios-Ferrer JL, Díaz C, Ángel García M, Boulaiz H, Valdivia J, Jurado JM, Almazan-Fernandez FM, Arias San-tiago S, Vicente F, del Val C, Pérez del Palacio J, Marchal JA. Identification of novel biomarkers in the early diagnosis of malignant melanoma by untargeted liquid chromatography coupled to high-resolution mass spectrometry-based metabolomics: a pilot study. Br J Dermatol 2024;190(5):740-50.
• [54] Bratchenko IA, Bratchenko LA, Khristoforova YA, Moryatov AA, Kozlov SV, Zakharov VP. Classification of skin cancer using convolutional neural networks analysis of Raman spectra. Comput Methods Programs Biomed 2022;219:106755.
• [55] Krohling B, Castro PB, Pacheco AG, Krohling RA. A smartphone based application for skin cancer classification using deep learning with clinical images and lesion information. 2021, arXiv preprintarXiv:2104.14353.
• [56] Courtenay LA, Barbero-García I, Martínez-Lastras S, Del Pozo S, Corral M, González-Aguilera D. Using computational learning for non-melanoma skin cancer and actinic keratosis near-infrared hyperspectral signature classification.Photodiagnosis Photodyn Ther 2024;49:104269.

Uwagi

Opracowanie rekordu ze środków MNiSW, umowa nr POPUL/SP/0154/2024/02 w ramach programu "Społeczna odpowiedzialność nauki II" - moduł: Popularyzacja nauki (2025).