Biology subjects

Pirone, D.

Publications and source records attributed to Pirone, D..

11 recordsLinked to original sources

bioRxiv2026

Impact of Image Representation on Deep Learning-Based Single-Cell Classification by Holographic Imaging Flow Cytometry

Accurate cell type classification is essential for a wide range of biomedical applications, including disease diagnosis, drug discovery, and the study of cellular processes. Holographic imaging flow cytometry (HIFC) provides label-free quantitative phase imaging (QPI) of individual cells, enabling classification based on phase images. However, reconstructing holograms into phase images involves multi-step image processing, which introduces substantial computational overhead. The availability of diverse image representations across holographic reconstruction stages allows for flexible analytical strategies, enabling the optimization of trade-off between classification accuracy and computational efficiency. Moreover, deep learning offers an efficient alternative, accelerating the reconstruction process while performing accurate classification. However, despite its importance, this optimization challenge remains largely unexplored in the current literature. Here, we present the first systematic evaluation aimed at balancing classification accuracy with computational efficiency, highlighting how different image representations affect overall performance. We focus on a binary classification task discriminating natural killer cells from breast cancer cells. Six distinct classification pipelines are evaluated: direct processing of raw holograms, analysis of demodulated complex fields (CFs), refocused CFs, unwrapped phase images, and two deep learning-based methods that either replace the automatic refocusing stage or perform end-to-end hologram-to-phase reconstruction. For each strategy, we assess both computational cost and classification performance. Our results reveal a clear trade-off: reconstructed phase images provide the highest accuracy, whereas simpler representations or accelerated reconstruction methods significantly reduce processing time with minimal loss of accuracy. A Pareto analysis identifies the optimal set of strategies, offering practical guidelines for selecting image representations and processing pipelines based on available hardware and desired performance. Thus, this work offers a systematic framework for high-throughput deep learning classification in HIFC, serving as a potential reference for future biomedical applications.

biophysics↗

bioRxiv2026

Self-consistent automatic retrieval of single cell rotation enables highly reliable holo-tomographic flow cytometry

Holo-Tomographic Flow Cytometry (HTFC) is a powerful, label-free imaging technique that allows obtaining three-dimensional (3D) refractive index (RI) tomograms of individual flowing and rotating cells. Determining the unknown orientation of moving cells is crucial for tomographic reconstruction. There is significant potential to enhance previous methods by increasing accuracy, achieving full automation, and reducing the dependence on a priori information. Here, we present a novel self-consistent method for the accurate and automatic retrieval of unknown rotations of single cells. The proposed method retrieves the sequence of rotation angles by an iterative reprojection-based optimization algorithm of the reconstructed 3D RI tomogram. The experimental results demonstrate that the method proposed in this work achieves full automation and superior accuracy, thus surpassing the existing techniques. This advancement markedly improves the scalability of HTFC, enabling automatic and highly reliable analysis of single cells.

biophysics↗

bioRxiv2026

Profiling co-occurrent morphological phenotypes and their degree of expression severity in vacuolated cells by holo-tomographic flow cytometry and fractal analysis

Cells are complex systems characterized by large phenotype heterogeneity. Conventional single cell classification approaches usually separate cells expressing a certain phenotype (e.g. associated to a disease condition) from the healthy control. However, multiple phenotypes typically coexist within the same cell as a result of complex intracellular interactions, machineries, functioning and external stimuli. Here we use label-free optical microscopy, powered by AI, to investigate how morphological phenotypes co-occur within vacuolated cells. Cytoplasmic vacuoles are important hallmarks of several pathological states (e.g. lysosomal storage diseases, viral infections, cancer). We rely on Holo-Tomographic Flow Cytometry (HTFC) to obtain 3D refractive index tomograms of vacuolated cells in continuous flow. Then, we propose a strategy to reduce the dimensionality of the tomogram using cross-sectioning and minimum intensity projection maps. We extract a set of morphological, refractive index-based, and fractal parameters demonstrating that the complex heterogeneity of vacuole patterns can be captured and can foster classification based on interpretable features. For training the AI, biologist domain-experts provided annotation of the different morphological phenotypes expressed and ranked them in terms of expression severity from the tomographic observations. Thus, we introduce a pipeline for morphometric phenotype profiling, in which each cell is associated with a 7-digits classification code representing the combination of coexisting phenotypes it expresses and their expression severity levels.

bioengineering↗

bioRxiv2026

Label-Free Nucleoli Measurement by 3D Holo-Tomographic Flow Cytometry Using Biolens Phase Compensation

Holo-Tomographic Flow Cytometry (HTFC) holds the potential to transform cellular research and clinical screening through 3D label-free quantitative phase imaging (QPI) of flowing single cells. However, it has been limited by insufficient intracellular specificity in 3D refractive index (RI) distributions, since suspended cells act as highly aberrating spherical biolenses obscuring internal structures. Here, we show the Biolens Phase Compensation (BPC), a method that corrects phase aberrations in 2D QPI projections to transform the 3D RI tomogram. Working within this new 3D pseudo-RI space demonstrates for the first time the extraction of nucleoli in HTFC. Extensive validation against 2D fluorescence flow cytometry and 3D confocal microscopy demonstrates that BPC achieves reliable intranuclear specificity. Using statistically significant single-cell analysis, we provide multiplexed quantitative 3D measurements of nested intracellular compartments (cytoplasm, nucleoplasm, nucleoli). This approach extends label-free HTFC toward capabilities of gold-standard Fluorescence Microscopy, overcoming its well-known drawbacks in intracellular and intranuclear staining.

biophysics↗

bioRxiv2026

Establishing label-free quantitative X-ray dose-response profiling by Holo-Tomographic Flow Cytometry

Flow cytometry (FC) offers multiparametric analysis capabilities that can quantify cellular damage after exposure to cytotoxic agents. Here, we present a comprehensive study establishing a label-free quantitative X-ray dose-response profiling using a novel FC modality based on 3D Quantitative Phase Imaging, termed Holo-Tomographic Flow Cytometry (HTFC). This approach enables fully label-free 3D refractive index (RI) measurements, allowing detailed and quantitative characterization of the biophysical properties and morphology of living cells exposed to X-rays. By analyzing datasets of 3D RI tomograms from cells irradiated at graded doses, we identify intracellular biophysical markers that define a robust X-ray dose-response curve. Validation against standard clonogenic survival assays on three model cancer cell lines reveal a high correlation (>90%). HTFC not only eliminates labeling and operator bias but also markedly reduces experimental time from 1-2 weeks to 24 hours, offering a fully automated and objective readout. While clonogenic survival remains the benchmark for radiosensitivity assessment, our findings establish HTFC as a powerful label-free platform for fast assessment of radiation damage. This technology paves the way for predictive biosensors that can capture patient-specific responses, thereby supporting the transition from conventional, uniform radiotherapy protocols to personalized treatment strategies.

biophysics↗

bioRxiv2025

Holotomography-driven learning for in-silico staining of single cells in flow cytometry avoiding co-registration

Virtual staining is the current state-of-the-art computational technique to cleverly enhance intracellular specificity in unstained biological samples by using convolutional neural networks (CNNs) trained on co-registered pairs of unstained/stained images. While effective, this approach suffers from unpredictable biases inherent to fluorescence microscopy and encounters challenges when applied to flow cytometry data as it would require accurate co-registration on a huge number of images. Here, we present a novel method that exploits for the first time a Holotomography-driven learning to completely eliminate the need for co-registration. We demonstrate that training a CNN on a stain-free dataset of 3D refractive index tomograms of flowing cells elegantly unlocks stain-free intracellular specificity in quantitative phase imaging flow cytometry. This breakthrough, by circumventing the critical co-registration bottleneck, opens unprecedented perspectives for label-free, high-throughput imaging flow cytometry, offering a powerful new paradigm for advanced 2D and 3D single-cell analysis.

biophysics↗

bioRxiv2025

Quantitative profiling of lysosomal accumulation through label-free biomarkers via High-Content Holo-Tomographic Flow Cytometry

Lysosomal storage diseases (LSDs) are genetic disorders caused by enzyme deficiencies that lead to lysosomal dysfunction and progressive cell damage. Accurate visualization and quantification of lysosomes are essential for understanding disease progression and developing effective therapies. Here, for the first time, we successfully identified and characterized lysosomes using an innovative Holo-Tomographic Flow Cytometry (HTFC) technique, which allows label-free, high-content, and high-throughput 3D imaging of lysosomal compartments in single live cells. This breakthrough could revolutionize traditional gold-standard methods overcoming the actual limitations. Leveraging this technology, we discovered novel biomarkers of lysosomal accumulation in LSD-affected cells. In fact, by generating refractive index tomograms, we achieved accurate measurement and comprehensive 3D visualization of cytoplasmic lysosomal aggregation in suspended single cells. Through experimental validation and advanced computational analyses, we identified a quantitative correlation between the 3D lysosomal architecture and the efficacy of various therapeutic strategies, including genetic and pharmacological interventions. This work represents a significant advance in lysosomal research, paving the way for improved diagnostics and the development of targeted therapies for LSDs.

cell biology↗

bioRxiv2024

Simulating 3D refractive index distributions of suspended cells

The introduction of holo-tomographic flow cytometry has unlocked the possibility of imaging the 3D refractive index (RI) distribution of suspended cells while they flow and rotate in microfluidic streams. Similarly, approaches that optically trap and rotate the samples can image them in suspension conditions. Great effort has been spent in developing robust algorithms for the tomogram estimation, as well as denoising and 3D segmentation algorithms. However, the lack of a ground-truth dataset for suspended cells significantly hinders the development of image processing pipelines, limiting the advancement of the associated technology field. Here we propose a novel method for simulating 3D refractive index tomograms of suspended cells. We start with prior knowledge of the statistics of the 3D RI distribution and morphometry of various cell sub-compartments gathered from holo-tomographic flow cytometry experiments and combined with literature data to create realistic 3D distributions. Next, we introduce a simple method to obtain a simulation of the tomogram obtainable from each cell, which approximates the effects of the limited numerical aperture of the system, the speckle noise and the inversion algorithm itself. As a benchmark for the simulator, we created a shared dataset of RI tomograms with various levels of complexity, simulating yeast eukaryotic cells at different budding stages with various phenotypes of cytoplasmic vacuoles, and the presence of cytoplasmic vacuoles and lipid droplets in monocyte cells. We believe that this shared dataset and the simulation method will contribute to the development of a wide plethora of tomogram estimation and processing algorithms.

biophysics↗

bioRxiv2021

Stain-free nucleus identification in holographic learning flow cyto-tomography

Quantitative Phase Imaging (QPI) has gained popularity because it can avoid the staining step, which in some cases is difficult or impossible. However, QPI does not provide the well-known specificity to various parts of the cell (e.g., organelles, membrane). Here we show a novel computational segmentation method based on statistical inference that bridges the gap between the specificity of Fluorescence Microscopy (FM) and the label-free property of QPI techniques to identify the cell nucleus. We demonstrate application to stain-free cells reconstructed through the holographic learning and in flow cyto-tomography modality. In particular, by means of numerical simulations and two cancer cell lines, we demonstrate that the nucleus-like regions can be accurately distinguished within the stain-free tomograms. We show that our experimental results are consistent with confocal FM data and microfluidic cytofluorimeter outputs. This is a significant step towards extracting the three-dimensional (3D) intracellular specificity directly from the phase-contrast data in a typical flow cytometry configuration.

bioengineering↗

bioRxiv2021

Lipid droplets 3D full measurement by holographic in-flow tomography

The most recent discoveries in the biochemical field are highlighting the increasingly important role of lipid droplets (LDs) in several regulatory mechanisms in living cells. LDs are dynamic organelles and therefore their complete characterization in terms of number, size, spatial positioning and relative distribution in the cell volume can shed light on the roles played by LDs. Until now, fluorescence microscopy and transmission electron microscopy are assessed as the gold standard methods for identifying LDs due to their high sensitivity and specificity. However, such methods generally only provide 2D assays and partial measurements. Furthermore, both can be destructive and with low productivity, thus limiting analysis of large cell numbers in a sample. Here we demonstrate for the first time the capability of 3D visualization and the full LD characterization in high-throughput with a tomographic phase-contrast flow-cytometer, by using ovarian cancer cells and monocyte cell lines as models. A strategy for retrieving significant parameters on spatial correlations and LD 3D positioning inside each cell volume is reported. The information gathered by this new method could allow more in depth understanding and lead to new discoveries on how LDs are correlated to cellular functions.

biophysics↗

bioRxiv2021

Tracking-based rolling angles recovery method for holographic tomography of flowing cells

Holographic Tomography (HT) is an emerging label-free technique for microscopic bioimaging applications, that allows reconstructing the three-dimensional (3D) refractive index (RI) distribution of biological specimens. Recently, an in-flow HT technique has been proposed in which multiple digital holograms are recorded at different viewing angles around the sample while it flows and rotates within a microfluidic channel. However, unlike conventional HT methods, there is no a priori information about cell 3D orientations, that are instead requested to perform any tomographic algorithm. Here we investigate a tracking-based rolling angles recovery method, showing robustness against the samples features. It is based on a phase images similarity metric recently demonstrated, that exploits the local contrast phase measurements to recognize a full cell rotation within the microfluidic channel. Hence, the orientations of the flowing cells are retrieved from their positions, which are in turn computed through the 3D holographic tracking. The performances of the rolling angles recovery method have been assessed both numerically, by simulating a 3D cell phantom, and experimentally, by reconstructing the 3D RI tomograms of two cancer cells. Both the numerical and the experimental analysis have been performed at different spatial resolutions. This rolling angles recovery method, not depending on the cell shapes, the RI contents, and the optical experimental conditions, could pave the way to the study of circulating tumor cells (CTCs) in the challenging tool of liquid biopsy.

bioengineering↗