Characterisation of breast cancer molecular signature and treatment assessment with vibrational spectroscopy and chemometric approach

Kołodziej, Magdalena; Kaznowska, Ewa; Paszek, Sylwia; Cebulski, Józef; Barnaś, Edyta; Cholewa, M.; Vongsvivut, Jitraporn; Zawlik, Izabela

doi:10.1371/journal.pone.0264347

Cited by 7 publications

(5 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PR importance presented the collagen (amide III) influence, where the 1204 cm −1 is associated with higher intensities for breast carcinoma tissue [1]. Still in PR, it is noted an absence of carbohydrates peaks in breast cancer spectra, which may be related to the higher glucose metabolism in cancer cells [103].…”

Section: Two-dimensional Deep Learningmentioning

confidence: 95%

“…At last, the 1080-980 cm −1 region is pertinent to glycogen (1050-1020 cm −1 ) and symmetric PO -2 stretching in RNA and DNA (1100-1040). Glucose expressions have been linked to cancer cells during the neoplastic process, while DNA and RNA oscillation is directly associated with cancer diagnosis [103].…”

Section: Data Preprocessingmentioning

confidence: 99%

“…Presence of PO -2 stretching vibrations may be associated to DNA damage caused by reactive oxygen species [103]. In 1300-1270 cm −1 , there are mainly amide III and collagen vibrational modes, and also a CH 2 wagging vibration of phospholipids acyl chains.…”

Section: Data Preprocessingmentioning

confidence: 99%

“…Adenine bands, as in subtype and ER, are reported to be higher in patients with cancer. This is due to the higher accelerated metabolism of the cells, which entails in oscillatory deformations of the C H peak of adenine [103]. Overproduction of fatty acids, listed in subtype importance, facilitates the tumor evolution and the survival of cancerous cells [103].…”

Section: Two-dimensional Deep Learningmentioning

confidence: 99%

“…Differences in DNA and RNA vibration frequency, as in the band shown in #1 ER importance, is an important evaluation to discriminate between normal and cancer spectra [103]. PR importance presented the collagen (amide III) influence, where the 1204 cm −1 is associated with higher intensities for breast carcinoma tissue [1].…”

Section: Two-dimensional Deep Learningmentioning

confidence: 99%

See 4 more Smart Citations

Evaluation of breast cancer molecular subtypes using artificial intelligence in micro-FTIR hyperspectral images

Valle¹

View full text Add to dashboard Cite

I would like to express my deepest appreciation to:• Professor Denise Maria Zezell, my advisor, for her trust, knowledge, academic dedication, and friendship. She is a brave and intelligent woman and a role model for any researcher.• Professor Emerson Soares Bernardes, my co-advisor, for providing ideas and guidance on this important subject of breast cancer.• My parents, Sueli and Alfredo, for their unwavering support throughout this journey. This title is as much theirs as it is mine.• My family, friends, and especially my girlfriend, Marina, for their understanding in times of absence and for the motivational support to keep moving on.• My colleagues at the Laboratory of Biophotonics, especially Moises, Pedro, Ariane and Amanda for their teachings, partnership, and friendship.• My colleagues at the Lasers and Applications Center, for all the shared knowledge.• USP and IPEN for the opportunities and infrastructure provided, and for their contribution to science and education. • FAPESP (CEPID 05/51689-2, 17/50332-0), CAPES (Finance Code 001, PROCAD 88881.068505/2014-01) and CNPq (INCT-465763/2014-6, PQ-309902/2017-7, 142229/2019-9) for their financial support. • Colleagues from my former job at the Health Innovation Techcenter of Hospital Israelita Albert Einstein, for sharing with me their GPU workstation (a.k.a. Vivi) and supporting my academic activities.• Everyone who contributed to this research. Thank you! RESUMO DEL-VALLE, Matheus. Avaliação de subtipos moleculares de câncer de mama utilizando inteligência artificial em imagens hiperespectrais por micro-FTIR. 2023. 70 p. Tese (Doutorado em Tecnologia Nuclear) -Instituto de Pesquisas Energéticas e Nucleares -IPEN -CNEN/SP. São Paulo. O câncer de mama é o mais incidente no mundo. A avaliação do subtipo molecular e seus biomarcadores tem um papel fundamental para o prognóstico. Os biomarcadores utilizados são os Receptores de Estrogênio (ER), de Progesterona (PR), de tipo 2 do fator de Crescimento Epidérmico Humano (HER2), e Ki67. Com base nestes, os subtipos são classificados como Luminal A (LA), Luminal B (LB), subtipo HER2 e Triplo-Negativo (TNBC). O padrão-ouro desta análise é a histologia e imuno-histoquímica, técnicas semiquantitativas que apresentam variações inter-laboratorial e inter-observador. A técnica de micro-espectroscopia no Infravermelho por Transformada de Fourier (FTIR), que fornece imagens hiperspectrais com informações bioquímicas de tecidos biológicos, é aplicada em conjunto de inteligência artificial (IA) para avaliação de cânceres. Nesta tese, foram utilizadas vinte amostras de duas linhagens celulares de câncer de mama, BT-474 e SK-BR-3, para definição do número ótimo de varreduras co-adicionadas para técnicas de aprendizado de máquina (ML). Foram utilizados os modelos de Análise Discriminante Linear (LDA), Análise Discriminante por Mínimos Quadrados Parciais (PLS-DA), K-Vizinhos Mais Próximos (KNN), Máquinas de Vetores de Suporte (SVM), Floresta Aleatória (RF) e Aumento de Gradiente Extremo (XGB). Sessenta imagens hiperespectrais de...

show abstract

Section: Two-dimensional Deep Learningmentioning

confidence: 95%

Section: Data Preprocessingmentioning

confidence: 99%

Section: Data Preprocessingmentioning

confidence: 99%

Section: Two-dimensional Deep Learningmentioning

confidence: 99%

Section: Two-dimensional Deep Learningmentioning

confidence: 99%

See 3 more Smart Citations