In biological dosimetry a radiation dose is estimated using the average number of chromosomal aberrations per peripheral blood lymphocytes. This analysis is still manually performed on 2D metaphase images depicting the 23 pairs of chromosomes because the false discovery rate of current automated detection systems is too high and variable because of sensitivity to small variations in image quality (chromosome spread, illumination variations …). Therefore, the current systems are only used to assist human experts. Designing more performant automatic and reliable chromosomal aberration detection systems has become of paramount importance to improve diagnosis speed and reduce human expertise time. Here, we propose a novel deep-learning method for automatic rare chromosomal aberration detection and uncertainty quantification. We formulate the problem as a unique regression problem requiring the minimization of a sparsity-promoting loss to reduce the false alarm rate. Furthermore, we select checkpoints at the end of each epoch during training to form a model ensemble. The resulting artificial experts are further analyzed to derive a consensus voting, similar to an agreement of human annotator rating, to provide trustworthy aberration detections and confidence intervals. A radiation dose curve is finally derived from deep learning-assisted counting of dicentrics and fragments in metaphase images, in high agreement with the reference hand-crafted curve in biological dosimetry with a promising dose estimation validation.
Dogs are a critical non-rodent species used in preclinical safety studies, particularly, in the pharmaceutical field, due to their physiological, metabolic, and immunological similarities to humans. As such, immunophenotyping of canine peripheral blood mononuclear cells (PBMCs) plays a crucial role in translational research, immune monitoring, and safety evaluations in drug development. However, the limited availability of canine-specific antibodies restricts detailed and accurate immune profiling, which is essential for advancing safety evaluations in drug development. To address this challenge, we developed a 15-marker panel for comprehensive mass cytometry-based immunophenotyping of cryopreserved canine PBMCs. This panel encompasses major leukocyte subsets, including B cells, CD4+ T helper cells, regulatory T cells, CD8+ cytotoxic T cells, memory T cell subsets, natural killer T cells, natural killer cells, dendritic cells, CD4+ monocytes, classical monocytes, and neutrophils. We utilized both extracellular and intracellular markers to facilitate in-depth immune profiling, despite the limited availability of canine-specific antibodies. The panel was thoroughly optimized in terms of marker selection, antibody clone validation, and metal isotope pairing. Additionally, by the use of mass cytometry, several channels remain unoccupied, providing flexibility for future panel expansion.
Flow cytometry is an essential component of routine hematological lab testing. Many computational methods have been proposed for the analysis of flow cytometry data, but most have focused on supervised learning for just one or a few specific disorders. To maximize clinical utility, we develop a method that enables identification of multiple common disorders and quality indicators. Our method includes a self-supervised pretraining component as well as a new, transformer-based model architecture. The self-supervised training algorithm is based on the DINO method while the model architecture is a relatively simple transformer encoder stack that includes a class (CLS) token, similar to BERT or vision-transformer models. Using a dataset of 52,625 samples obtained during routine clinical testing at our laboratory, we show that our pretraining method develops informative tube-level representations that clearly separate important diagnostic classes. We then evaluate performance on multiple downstream tasks, including sample viability estimation and five common hematological disorders. We compare our method to self-organizing maps, convolutional neural networks, attention-based multiple-instance learning models, and two varieties of set-transformer-based models, and demonstrate that our method delivers higher classification performance than other approaches.
Advances in spectral cytometry instrumentation and fluorescent reagents have led to the possibility of ultra-high-parameter panels exceeding 50 colors. However, panel size is limited in practice by unmixing-dependent spreading (UDS), a phenomenon which leads to a progressive deterioration of unmixed signal-to-noise ratios in panels that contain fluorochrome combinations with significant spectral overlap. So far, choosing spectrally compatible sets of fluorochromes that avoid UDS has been a complex and labor-intensive task involving substantial trial-and-error experimentation. Here, we provide a detailed explanation of UDS and practical strategies for handling UDS in large spectral panels. We describe the empirical hallmarks of UDS, demonstrate how to quantify its impact, and dissect its underlying mathematical cause in terms of spectral collinearity. We present practical tools derived from the regression literature that can be used to select optimal combinations of fluorochromes in a platform-agnostic fashion based on publicly available reference data, providing a general tool for spectral panel design.
