Predictive data clustering of laser-induced breakdown spectroscopy for brain tumor analysis

Abstract

Limited by the lack of training spectral data in different kinds of tissues, the diagnostic accuracy of laser-induced breakdown spectroscopy (LIBS) is hard to reach the desired level with normal supervised learning identification methods. In this paper, we proposed to apply the predictive data clustering methods with supervised learning methods together to identify tissue information accurately. The meanshift clustering method is introduced to compare with three other clustering methods which have been used in LIBS field. We proposed the cluster precision (CP) score as a new criterion to work with Calinski-Harabasz (CH) score together for the evaluation of the clustering effect. The influences of principal component analysis (PCA) on all four kinds of clustering methods are also analyzed. PCA-meanshift shows the best clustering effect based on the comprehensive evaluation combined CH and CP scores. Based on the spatial location and feature similarity information provided by the predictive clustering, the PCA-Meanshift can improve diagnosis accuracy from less than 95% to 100% for all classifiers including support vector machine (SVM), k nearest neighbor (k-NN), soft independent modeling of class analogy (Simca) and random forests (RF) models.

Single-cell multimodal modeling with deep parametric inference

“…It is worth noting that to eliminate differences in clustering algorithms, the low-dimensional representations of these models are all fed into the Leiden clustering algorithm [13] with the same parameters. The results show that the DPI model outperforms SeuratV4 and TotalVI on both Calinski Harabasz score [14] and Silhouette score [15] evaluations (Fig. 3A).…”

Laser-Induced Breakdown Spectroscopy in Biological Samples: A Review of Experiments with Soft Tissues

“…Other cases of tissues with very similar compositions were glioma and the infiltrating tissues obtained in brain tumor surgery. In [57], this kind of sample was studied, with half embedded in paraffin for pathological examination and half prepared on glass slides for LIBS measurements. They consisted of 1200 spectra that were statistically analyzed.…”

Visualizing the cerebral distribution of chemical elements: A challenge met with LIBS elemental imaging

“…From this plasma, the emitted light is collected and analyzed with adapted spectrometers and the spectroscopic data are further converted into images i.e. elemental maps 14 . Working at frequency of 100 Hz (as LA-ICP-MS does), this strategy for imaging chemical elements is fast, and advantageously complementary to the conventional histological investigations performed with optical microscopy for biomedical purposes.…”