Applications of Data Characteristic AI-Assisted Raman Spectroscopy in Pathological Classification.

Raman spectroscopy has been widely used for label-free biomolecular analysis of cells and tissues for pathological diagnosis in vitro and in vivo. AI technology facilitates disease diagnosis based on Raman spectroscopy, including machine learning (PCA and SVM), manifold learning (UMAP), and deep learning (ResNet and AlexNet). However, it is not clear how to optimize the appropriate AI classification model for different types of Raman spectral data. Here, we selected five representative Raman spectral data sets, including endometrial carcinoma, hepatoma extracellular vesicles, bacteria, melanoma cell, diabetic skin, with different characteristics regarding sample size, spectral data size, Raman shift range, tissue sites, Kullback-Leibler (KL) divergence, and significant Raman shifts (i.e., wavenumbers with significant differences between groups), to explore the performance of different AI models (e.g., PCA-SVM, SVM, UMAP-SVM, ResNet or AlexNet). For data set of large spectral data size, Resnet performed better than PCA-SVM and UMAP. By building data characteristic-assisted AI classification model, we optimized the network parameters (e.g., principal components, activation function, and loss function) of AI model based on data size and KL divergence etc. The accuracy improved from 85.1 to 94.6% for endometrial carcinoma grading, from 77.1 to 90.7% for hepatoma extracellular vesicles detection, from 89.3 to 99.7% for melanoma cell detection, from 88.1 to 97.9% for bacterial identification, from 53.7 to 85.5% for diabetic skin screening, and mean time expense of 5 s.

VHL mutation drives human clear cell renal cell carcinoma progression through PI3K/AKT-dependent cholesteryl ester accumulation.

BACKGROUND: Cholesteryl ester (CE) accumulation in intracellular lipid droplets (LDs) is an essential signature of clear cell renal cell carcinoma (ccRCC), but its molecular mechanism and pathological significance remain elusive. METHODS: Enabled by the label-free Raman spectromicroscopy, which integrated stimulated Raman scattering microscopy with confocal Raman spectroscopy on the same platform, we quantitatively analyzed LD distribution and composition at the single cell level in intact ccRCC cell and tissue specimens in situ without any processing or exogenous labeling. Since we found that commonly used ccRCC cell lines actually did not show the CE-rich signature, primary cancer cells were isolated from human tissues to retain the lipid signature of ccRCC with CE level as high as the original tissue, which offers a preferable cell model for the study of cholesterol metabolism in ccRCC. Moreover, we established a patient-derived xenograft (PDX) mouse model that retained the CE-rich phenotype of human ccRCC. FINDINGS: Surprisingly, our results revealed that CE accumulation was induced by tumor suppressor VHL mutation, the most common mutation of ccRCC. Moreover, VHL mutation was found to promote CE accumulation by upregulating HIFα and subsequent PI3K/AKT/mTOR/SREBPs pathway. Inspiringly, inhibition of cholesterol esterification remarkably suppressed ccRCC aggressiveness in vitro and in vivo with negligible toxicity, through the reduced membrane cholesterol-mediated downregulations of integrin and MAPK signaling pathways. INTERPRETATION: Collectively, our study improves current understanding of the role of CE accumulation in ccRCC and opens up new opportunities for treatment. FUNDING: This work was supported by National Natural Science Foundation of China (No. U23B2046 and No. 62027824), National Key R&D Program of China (No. 2023YFC2415500), Fundamental Research Funds for the Central Universities (No. YWF-22-L-547), PKU-Baidu Fund (No. 2020BD033), Peking University First Hospital Scientific and Technological Achievement Transformation Incubation Guidance Fund (No. 2022CX02), and Beijing Municipal Health Commission (No. 2020-2Z-40713).