Machine learning of diffraction image patterns for accurate classification of cells modeled with different nuclear sizes

Accurate Classification of Human CD4+ T, CD8+ T, and CD19+ B Cells Isolated from Splenocytes by Cross-Polarized Diffraction Image Pairs

Diffraction imaging of cells allows rapid phenotyping by the response of intracellular molecules to coherent illumination. However, its ability to distinguish numerous types of human leukocytes remains to be investigated. Here, we show that accurate classification of three lymphocyte subtypes can be achieved with features extracted from cross-polarized diffraction image (p-DI) pairs. A deep neural network (DNN) of DINet-PS has been developed for feature extraction from and filtering of, in the angular frequency domain, p-DI pairs acquired from live lymphocytes isolated from human spleen tissues. We built the network in a dual-channel structure and incorporated two adaptive spectral filter blocks to actively suppress extracted features related to the noise component of light in p-DI pairs. The DINet-PS was trained with p-DI pairs acquired from 5311 CD4+ T, 3819 CD8+ T, and 4054 CD19+ B cells after preprocessing and rebelling of manually derived secondary labels and classification accuracy of 96.6 ± 0.40% has been achieved in hold-out test data sets among the three subtypes. Our results show the power of DNN to extract cell-related features from p-DI pairs and the potential of polarization diffraction imaging flow cytometry for accurate and label-free classification of lymphocyte subtypes in particular and leukocytes in general.

Multi-wavelength multi-direction laser light scattering for cell characterization using machine learning-based methods

Cell identification and analysis play a crucial role in many biology- and health-related applications. The internal and surface structures of a cell are complex and many of the features are sub-micron in scale. Well-resolved images of these features cannot be obtained using optical microscopy. Previous studies have reported that the single-cell angular laser-light scattering patterns (ALSP) can be used for label-free cell identification and analysis. The ALSP can be affected by cell properties and the wavelength of the probing laser. Two cell properties, cell surface roughness and the number of mitochondria, are investigated in this study. The effects of probing laser wavelengths (blue, green, and red) and the directions of scattered light collection (forward, side, and backward) are studied to determine the optimum conditions for distinguishing the two cell properties. Machine learning (ML) analysis has been applied to ALSP obtained from numerical simulations. The results of ML analysis show that the backward scattering is the best direction for characterizing the surface roughness, while the forward scattering is the best direction for differentiating the number of mitochondria. The laser light having red or green wavelength is found to perform better than that having the blue wavelength in differentiating the surface roughness and the number of mitochondria. This study provides important insights into the effects of probing laser wavelength on gaining information about cells from their ALSP.

Stain-free identification of cell nuclei using tomographic phase microscopy in flow cytometry

Quantitative Phase Imaging (QPI) has gained popularity in bioimaging because it can avoid the need for cell staining, which in some cases is difficult or impossible. However, as a result, QPI does not provide labelling of various specific intracellular structures. Here we show a novel computational segmentation method based on statistical inference that makes it possible for QPI techniques to identify the cell nucleus. We demonstrate the approach with refractive index tomograms of stain-free cells reconstructed through the tomographic phase microscopy in flow cytometry mode. In particular, by means of numerical simulations and two cancer cell lines, we demonstrate that the nucleus can be accurately distinguished within the stain-free tomograms. We show that our experimental results are consistent with confocal fluorescence microscopy (FM) data and microfluidic cytofluorimeter outputs. This is a significant step towards extracting specific three-dimensional intracellular structures directly from the phase-contrast data in a typical flow cytometry configuration.

Analysis of polarized diffraction images of human red blood cells: a numerical study

We carried out a systematic study on cross-polarized diffraction image (p-DI) pairs of 3098 mature red blood cells (RBCs) using optical cell models with varied morphology, refractive index (RI), and orientation. The influence of cell rotation on texture features of p-DI pairs characterized by the gray-level co-occurrence matrix (GLCM) algorithm was quantitatively analyzed. Correlations between the transverse diameters of RBCs with different RI values and scattering efficiency ratios of s- and p-polarized light were also investigated. The correlations remain strong even for RBCs with significant orientation variations. In addition, we applied a minimum redundancy maximum relevance (mRMR) algorithm to improve the performance of support vector machine (SVM) classifiers. It was demonstrated that a set of selected GLCM parameters allowed for an efficient solution of classification problems of RBCs based on morphology. For 1598 RBCs with varied shapes corresponding to normal or pathological cases, the accuracy of the SVM based classifications increased from 83.8% to 96.8% with the aid of mRMR. These results indicate the strong potential of p-DI data for rapid and accurate screening examinations of RGC shapes in routine clinical tests.

Integration of light scattering with machine learning for label free cell detection

Light scattering has been used for label-free cell detection. The angular light scattering patterns from the cells are unique to them based on the cell size, nucleus size, number of mitochondria, and cell surface roughness. The patterns collected from the cells can then be classified based on different image characteristics. We have also developed a machine learning (ML) method to classify these cell light scattering patterns. As a case study we have used this light scattering technique integrated with the machine learning to analyze staurosporine-treated SH-SY5Y neuroblastoma cells and compare them to non-treated control cells. Experimental results show that the ML technique can provide a classification accuracy (treated versus non-treated) of over 90%. The predicted percentage of the treated cells in a mixed solution is within 5% of the reference (ground-truth) value and the technique has the potential to be a viable method for real-time detection and diagnosis.