CC BY
0Recognition of malignant cutaneous melanoma by multimodal
analysis of optical biopsy data
I. Matveeva1
1- Samara National Research University, Moskovskoye shosse 34, Samara, 443086, Russia
m-irene-a@yandex.ru
The International Agency for Research on Cancer (IARC) estimates that approximately 325,000 new cases of malignant melanoma (MM) were diagnosed worldwide in 2020 and 57,000 people died from the disease. Moreover, researchers predict that the number of new cases of MM per year will increase by more than 50% from 2020 to 2040 [1]. Externally, MM may be similar to a pigmented nevus (PN). Therefore, the quality of diagnosing MM by visual examination largely depends on the level of qualifications and professional experience of the doctor and is 40-80% accurate [2]. The most complete clinical picture is provided by taking a biopsy, which involves taking a sample of the neoplasm and its further histological examination by a specialist pathologist [3]. Due to the "aggressive" behavior and high risks of metastasis due to external influence, such a procedure is usually not used [4]. Optical methods are recognized as promising tools for studying skin tissue. Such methods include dermatoscopy [5], Raman spectroscopy [6,7], hyperspectral imaging [8], etc. However, at the moment, optical biopsy methods do not exceed the accuracy of the "gold standard" of diagnosis, that is, histological examination. One of the ways to overcome the limited accuracy values is the combined use of several optical methods [9]. The previous research showed an increase in the classification accuracy of MM and PN by joint analysis of Raman spectra and dermoscopic images [5]. The aim of this research is to develop a method for identifying MM based on multimodal joint analysis of Raman scattering data, dermatoscopic images, and hyperspectral images. An in vivo study of skin neoplasms was carried out at the Samara Regional Clinical Oncology Center. Experimental skin Raman spectra were recorded using a portable setup that includes a laser source with a central wavelength of 785 nm. The spectra were recorded with a spectral resolution of 0.2 nm in the range from 837 to 920 nm, which corresponds to 792-1874 cm-1 [6]. Dermatoscopic images of skin neoplasms were obtained using a digital dermatoscope [5]. To record hyperspectral images, an acoustooptical hyperspectral camera was used, which makes it possible to obtain an image of the area under study at an arbitrarily set wavelength in the range of 440-750 nm with a spectral resolution of 2.5 nm and the spatial resolution of 0.14 mm [8]. Machine learning methods, in particular, logistic regression and convolutional neural networks, were used for analysis of the registered data. The classification model for MM and PN has shown an increase in accuracy compared to the analysis of Raman spectra, dermatoscopic images, or hyperspectral images alone. As a result, a comprehensive multimodal approach for the detection of MM, which takes into account both specific spectral characteristics of neoplasms and spatial inhomogeneities in the distribution of optical density, has been proposed. The studied approaches to the analysis of optical biopsy data can potentially be integrated into the software for automated screening diagnostics of skin.
[1] J. Ferlay, M. Colombet, I. Soerjomataram, D.M. Parkin, M. Piñeros, A. Znaor, F. Bray, Cancer statistics for the year 2020: An overview, International journal of cancer, vol. 149(4), pp. 778-789, 2021.
[2] H.A. Haenssle, C. Fink, R. Schneiderbauer, F. Toberer, T. Buhl, A. Blum, A. Kalloo, A.B.H. Hassen, L. Thomas, A. Enk, L. Uhlmann, Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists, Annals of oncology, vol. 29(8), pp. 1836-1842, 2018.
[3] G.V. Long, S.M. Swetter, A.M. Menzies, J.E. Gershenwald, R.A. Scolyer, Cutaneous melanoma, The Lancet, vol. 402(10400), pp. 485-502, 2023.
[4] R.J. Friedman, D.S. Rigel, A.W. Kopf, Early detection of malignant melanoma: the role of physician examination and self-examination of the skin., CA: a cancer journal for clinicians, vol. 35(3), pp. 130-151, 1985.
[5] I.A. Matveeva, A.I. Komlev, O.I. Kaganov, A.A. Moryatov, V.P. Zakharov, Multidimensional Analysis of Dermoscopic Images and Spectral Information for the Diagnosis of Skin Tumors, Journal of Biomedical Photonics & Engineering, vol. 10(1), pp. 010307, 2024.
[6] I.A. Bratchenko, L.A. Bratchenko, A.A. Moryatov, Y.A. Khristoforova, D.N. Artemyev, O.O. Myakinin, A.E. Orlov, S.V. Kozlov, V.P. Zakharov, In vivo diagnosis of skin cancer with a portable Raman spectroscopic device, Experimental Dermatology, vol. 30(5), pp. 652663, 2021.
[7] I. Matveeva, I. Bratchenko, Y. Khristoforova, L. Bratchenko, A. Moryatov, S. Kozlov, O. Kaganov, V. Zakharov, Multivariate curve resolution alternating least squares analysis of in vivo skin Raman spectra, Sensors, vol. 22(24), pp. 9588, 2022.
[8] B.V. Grechkin, V.O. Vinokurov, Y.A. Khristoforova, I.A. Matveeva, VGG convolutional neural network classification of hyperspectral images of skin neoplasms, Journal of Biomedical Photonics & Engineering, vol. 9(4), pp. 040304, 2023.
[9] L. Rey-Barroso, S. Peña-Gutiérrez, C. Yáñez, F.J. Burgos-Fernández, M. Vilaseca, S. Royo, Optical technologies for the improvement of skin cancer diagnosis: a review, Sensors, vol. 21(1), pp. 252, 2021.
