Hematoxylin and eosin (H&E) staining has been a gold standard for diagnosing cancer in histopathology. Nonetheless, H&E staining heavily requires time, resources and is limited in two-dimensional analyses. Here, we propose three-dimensional (3D) virtual staining of hematoxylin and eosin (H&E) from label-free refractive index (RI) images of histopathological slides. To achieve this, we integrated RI tomography, which provides 3D RI distribution without staining, with deep learning. By predicting brightfield (BF) images from RI images of thick colon slides, we generate virtually stained 3D H&E images. The accuracy of our approach is validated against conventional staining methods.
The Nucleophosmin 1 (NPM1) mutation rapidly progresses to acute myeloid leukemia, emphasizing the need for early diagnosis, especially in cases with low blast counts. Some low blast count instances may not undergo next-generation sequencing, causing delays and necessitating swift NPM1 mutation screening. This study utilizes cutting-edge label-free three-dimensional imaging with holotomography (HT) to identify NPM1 mutation in blasts. Machine learning and deep learning algorithms achieve precise single-cell and patient-specific predictions. HT's accurate detection of protein movement associated with NPM1 mutation holds great promise as a reliable and efficient tool for early detection in hematologic malignancy patients with low blast counts.
Induced pluripotent stem cells (iPSCs) hold the potential for personalized regenerative medicine. Yet accurately gauging the stemness of each colony remains a challenge since existing methods are inconsistent or harmful to iPSCs. Addressing this, we introduce holotomography (HT), a non-invasive microscopic technique. HT, through three-dimensional refractive index distributions, revealed iPSC structures at various scales as well as properties like volume, mass density, and lipid ratio. We identified altered properties in iPSCs exposed to differentiation agents, and then employed a machine-learning algorithm to detect reduced stemness from images. Through these results, HT emerges as a potential tool for iPSC quality maintenance.
Accurate and rapid evaluation of dynamic immune status is critical to determine therapeutic modalities for sepsis patients, which is impeded by the limitations of conventional diagnostic tools. Here, we employ refractive index tomography to quantitatively assess the immune status of human monocytes in a label-free manner. Measurement of refractive index tomograms enabled quantifications of three-dimensional morphological parameters, which revealed a clear increment in lipid droplets content and intracellular inhomogeneities as the septic stage progresses. We leveraged these observations to engineer a deep-learning-based algorithm that predicts the immune status of monocytes, showing over 99 % blind test accuracy.
Rapid identification of infectious pathogens can save lives and mitigate healthcare expenses. Yet the current turnaround time for microbial identification typically exceeds 24 hours, as the common methods require the cultivation of millions or more bacteria to detect the collective signal. In this study, we propose a hybrid framework of quantitative phase imaging and artificial neural network to facilitate rapid identification at an individual-cell level. Specifically, three-dimensional images of refractive index were acquired for individual bacteria, and an optimized artificial neural network determined the species based on the three-dimensional morphologies, securing 82.5% blind test accuracy at an individual-cell level.
We present a deep learning approach for the rapid resolution enhancement of optical diffraction tomography. Once our three-dimensional U-net-based convolutional neural network learns an image translation between raw tomograms and total-variation-regularized tomograms, the trained network can fill in the missing cone of a measured refractive index tomogram and improve its resolution within seconds. We demonstrate the feasibility and generalizability of our approach on various biological samples, including bacteria, WBC, and NIH3T3.
Rapid, label-free, volumetric, and automated assessment in microscopy is necessary to assess the dynamic interactions between lymphocytes and their targets through the immunological synapse (IS) and the relevant immunological functions. However, attempts to realize the automatic tracking of IS dynamics have been stymied by the limitations of imaging techniques and computational analysis methods. Here, we demonstrate the automatic three-dimensional IS tracking by combining optical diffraction tomography and deep-learning-based segmentation. The proposed approach enables quantitative spatiotemporal analyses of IS regarding morphological and biochemical parameters related to its protein densities, offering a novel complementary method to fluorescence microscopy for studies in immunology.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.