Label-Free Leukemia Monitoring by Computer Vision

Cold storage surpasses the impact of biological age and donor characteristics on red blood cell morphology classified by deep machine learning

Assessment of the morphology of red blood cells (RBCs) can improve clinical benefits following blood transfusion. Deep machine learning surpasses traditional microscopy-based classification methods, offering more accurate and consistent results while reducing time and labor intensity. RBCs from teenage males, teenage females, senior males, and senior females were biologically age-profiled or density-separated into dense/old RBCs (O-RBCs) and less-dense/young (Y-RBCs) throughout hypothermic storage and assessed using image flow cytometry with deep machine learning analysis. Regardless of biological age, morphology index decreased with hypothermic storage. Significant differences in RBC morphology index were not seen when comparing unseparated RBCs (U-RBCs), O-RBCs, and Y-RBCs, although the proportions of morphology subclasses revealed differences between RBCs groups from different donor groups and in samples with different biological age. Cold storage remains the most significant influence on morphology, although teenage male donors demonstrated slightly more susceptibility to storage lesions compared with senior males and females. Our work highlights that hypothermic storage most significantly impacts RBC morphology over biological age and donor characteristics, emphasizing the importance of storage effects on transfusion quality and safety.

Imaging flow cytometry as a novel approach for the diagnosis of heparin-induced thrombocytopenia

Heparin-induced thrombocytopenia (HIT) is an adverse reaction characterized by anti-PF4-heparin antibody generation and hypercoagulability. Imaging flow cytometry (IFC) provides a detailed morphological analysis of platelets, which change upon activation. We evaluated IFC-derived morphometric features to detect platelet activation and developed a functional assay for HIT diagnosis. We analysed blood samples from 42 patients with suspected HIT and extracted platelet size, shape and texture features using IFC. The morphological features were compared with CD62P expression, light transmission aggregometry (LTA) and a serotonin release assay (SRA) in terms of their ability to predict a HIT diagnosis. Five IFC-derived morphological features (area, circularity, contrast, diameter and major axis) significantly distinguished resting from activated platelets. The major axis feature performed best for HIT diagnosis, with a sensitivity of 89.3% and a specificity of 92.9% versus functional assays (LTA/SRA); this diagnostic performance was similar to that of CD62P expression on the same platelet donors. The area and diameter had similar specificity (92.9%) and a slightly lower sensitivity (85.7%). The morphological features associated with platelet activation might be effective markers for the diagnosis of HIT, matching platelet CD62P expression assay performance. The high-throughput IFC exploration of platelet activation offers new perspectives in label-free analysis and time-saving.

Blood cancer prediction model based on deep learning technique

Blood cancer is among the critical health concerns among people around the world and normally emanates from genetic and environmental issues. Early detection becomes essential, as the rate of death associated with it is high, to ensure that the rate of treatment success is up, and mortality reduced. This paper focuses on improving blood cancer diagnosis using advanced deep learning techniques like ResNetRS50, RegNetX016, AlexNet, Convnext, EfficientNet, Inception_V3, Xception, and VGG19. Among the models assessed, ResNetRS50 had better accuracy and speed with minimal error rates compared with other state-of-the-arts. This work will exploit the power of ResNetRS50 in contributing to early detection and reducing bad outcomes for blood cancer patients. Blood cancer is currently one of the deadliest diseases worldwide, resulting from a combination of genetic and non-genetic factors. It stands as a leading cause of cancer-related deaths in both developed and developing nations. Early detection of cancer is pivotal in reducing mortality rates, as it increases the likelihood of successful treatment and potential cure. The objective is to decrease mortality rates through early diagnosis of blood cancer, thus offering individuals a better chance of survival from this disease.

The Role of Machine Learning Approaches in Pediatric Oncology: A Systematic Review

To enhance patient outcomes in pediatric cancer, a better understanding of the medical and biological risk variables is required. With the growing amount of data accessible to research in pediatric cancer, machine learning (ML) is a form of algorithmic inference from sophisticated statistical techniques. In addition to highlighting developments and prospects in the field, the objective of this systematic study was to methodically describe the state of ML in pediatric oncology. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to search for relevant studies on four distinct databases (Scopus, Web of Science, PubMed, and Cochrane Library). A total of 1536 relevant studies were retrieved to the EndNote library (Clarivate, Philadelphia, USA) where duplicates were removed and the rest of the studies were assessed for eligibility based on titles, abstracts, and the availability of full-text articles. After assessing the studies for eligibility, we found 42 studies eligible for inclusion in this systematic review. We found nine studies on liquid tumors, 13 on solid tumors, and 20 on central nervous system (CNS) tumors. ML goals included classification, treatment response prediction, and dose optimization. Neural networks, k-nearest neighbors, random forests, support vector machines, and naive Bayes were among the techniques employed. The identified studies' strengths included treatment response prediction and automated analysis that matched or outperformed physician comparators. Significant variation in clinical applicability, criteria for reporting, limited sample numbers, and the absence of external validation cohorts were among the common issues. We found places where ML can improve clinical care in manners that would not be possible otherwise. Even though ML has great promise for enhancing pediatric cancer diagnosis, decision-making, and monitoring, the discipline is still in its infancy, and standards and recommendations will support future research to guarantee robust methodologic design and maximize therapeutic applicability.

Capturing cell heterogeneity in representations of cell populations for image-based profiling using contrastive learning

Image-based cell profiling is a powerful tool that compares perturbed cell populations by measuring thousands of single-cell features and summarizing them into profiles. Typically a sample is represented by averaging across cells, but this fails to capture the heterogeneity within cell populations. We introduce CytoSummaryNet: a Deep Sets-based approach that improves mechanism of action prediction by 30-68% in mean average precision compared to average profiling on a public dataset. CytoSummaryNet uses self-supervised contrastive learning in a multiple-instance learning framework, providing an easier-to-apply method for aggregating single-cell feature data than previously published strategies. Interpretability analysis suggests that the model achieves this improvement by downweighting small mitotic cells or those with debris and prioritizing large uncrowded cells. The approach requires only perturbation labels for training, which are readily available in all cell profiling datasets. CytoSummaryNet offers a straightforward post-processing step for single-cell profiles that can significantly boost retrieval performance on image-based profiling datasets.

Optofluidic time-stretch imaging flow cytometry with a real-time storage rate beyond 5.9 GB/s

Optofluidic time-stretch imaging flow cytometry (OTS-IFC) provides a suitable solution for high-precision cell analysis and high-sensitivity detection of rare cells due to its high-throughput and continuous image acquisition. However, transferring and storing continuous big data streams remains a challenge. In this study, we designed a high-speed streaming storage strategy to store OTS-IFC data in real-time, overcoming the imbalance between the fast generation speed in the data acquisition and processing subsystem and the comparatively slower storage speed in the transmission and storage subsystem. This strategy, utilizing an asynchronous buffer structure built on the producer-consumer model, optimizes memory usage for enhanced data throughput and stability. We evaluated the storage performance of the high-speed streaming storage strategy in ultra-large-scale blood cell imaging on a common commercial device. The experimental results show that it can provide a continuous data throughput of up to 5891 MB/s.

Accessible high-speed image-activated cell sorting

Over the past six decades, fluorescence-activated cell sorting (FACS) has become an essential technology for basic and clinical research by enabling the isolation of cells of interest in high throughput. Recent technological advancements have started a new era of flow cytometry. By combining the spatial resolution of microscopy with high-speed cell sorting, new instruments allow cell sorting based on simple image-derived parameters or sophisticated image analysis algorithms, thereby greatly expanding the scope of applications. In this review, we discuss the systems that are commercially available or have been described in enough methodological and engineering detail to allow their replication. We summarize their strengths and limitations and highlight applications that have the potential to transform various fields in basic life science research and clinical settings.

Leveraging deep learning for detecting red blood cell morphological changes in blood films from children with severe malaria anaemia

In sub-Saharan Africa, acute-onset severe malaria anaemia (SMA) is a critical challenge, particularly affecting children under five. The acute drop in haematocrit in SMA is thought to be driven by an increased phagocytotic pathological process in the spleen, leading to the presence of distinct red blood cells (RBCs) with altered morphological characteristics. We hypothesized that these RBCs could be detected systematically and at scale in peripheral blood films (PBFs) by harnessing the capabilities of deep learning models. Assessment of PBFs by a microscopist does not scale for this task and is subject to variability. Here we introduce a deep learning model, leveraging a weakly supervised Multiple Instance Learning framework, to Identify SMA (MILISMA) through the presence of morphologically changed RBCs. MILISMA achieved a classification accuracy of 83% (receiver operating characteristic area under the curve [AUC] of 87%; precision-recall AUC of 76%). More importantly, MILISMA's capabilities extend to identifying statistically significant morphological distinctions (p < 0.01) in RBCs descriptors. Our findings are enriched by visual analyses, which underscore the unique morphological features of SMA-affected RBCs when compared to non-SMA cells. This model aided detection and characterization of RBC alterations could enhance the understanding of SMA's pathology and refine SMA diagnostic and prognostic evaluation processes at scale.

Capturing cell heterogeneity in representations of cell populations for image-based profiling using contrastive learning

Image-based cell profiling is a powerful tool that compares perturbed cell populations by measuring thousands of single-cell features and summarizing them into profiles. Typically a sample is represented by averaging across cells, but this fails to capture the heterogeneity within cell populations. We introduce CytoSummaryNet: a Deep Sets-based approach that improves mechanism of action prediction by 30-68% in mean average precision compared to average profiling on a public dataset. CytoSummaryNet uses self-supervised contrastive learning in a multiple-instance learning framework, providing an easier-to-apply method for aggregating single-cell feature data than previously published strategies. Interpretability analysis suggests that the model achieves this improvement by downweighting small mitotic cells or those with debris and prioritizing large uncrowded cells. The approach requires only perturbation labels for training, which are readily available in all cell profiling datasets. CytoSummaryNet offers a straightforward post-processing step for single-cell profiles that can significantly boost retrieval performance on image-based profiling datasets.

Imaging Flow Cytometry and Convolutional Neural Network-Based Classification Enable Discrimination of Hematopoietic and Leukemic Stem Cells in Acute Myeloid Leukemia

Acute myeloid leukemia (AML) is a heterogenous blood cancer with a dismal prognosis. It emanates from leukemic stem cells (LSCs) arising from the genetic transformation of hematopoietic stem cells (HSCs). LSCs hold prognostic value, but their molecular and immunophenotypic heterogeneity poses challenges: there is no single marker for identifying all LSCs across AML samples. We hypothesized that imaging flow cytometry (IFC) paired with artificial intelligence-driven image analysis could visually distinguish LSCs from HSCs based solely on morphology. Initially, a seven-color IFC panel was employed to immunophenotypically identify LSCs and HSCs in bone marrow samples from five AML patients and ten healthy donors, respectively. Next, we developed convolutional neural network (CNN) models for HSC-LSC discrimination using brightfield (BF), side scatter (SSC), and DNA images. Classification using only BF images achieved 86.96% accuracy, indicating significant morphological differences. Accuracy increased to 93.42% when combining BF with DNA images, highlighting differences in nuclear morphology, although DNA images alone were inadequate for accurate HSC-LSC discrimination. Model development using SSC images revealed minor granularity differences. Performance metrics varied substantially between AML patients, indicating considerable morphologic variations among LSCs. Overall, we demonstrate proof-of-concept results for accurate CNN-based HSC-LSC differentiation, instigating the development of a novel technique within AML monitoring.

Comparative analysis of feature-based ML and CNN for binucleated erythroblast quantification in myelodysplastic syndrome patients using imaging flow cytometry data

Myelodysplastic syndrome is primarily characterized by dysplasia in the bone marrow (BM), presenting a challenge in consistent morphology interpretation. Accurate diagnosis through traditional slide-based analysis is difficult, necessitating a standardized objective technique. Over the past two decades, imaging flow cytometry (IFC) has proven effective in combining image-based morphometric analyses with high-parameter phenotyping. We have previously demonstrated the effectiveness of combining IFC with a feature-based machine learning algorithm to accurately identify and quantify rare binucleated erythroblasts (BNEs) in dyserythropoietic BM cells. However, a feature-based workflow poses challenges requiring software-specific expertise. Here we employ a Convolutional Neural Network (CNN) algorithm for BNE identification and differentiation from doublets and cells with irregular nuclear morphology in IFC data. We demonstrate that this simplified AI workflow, coupled with a powerful CNN algorithm, achieves comparable BNE quantification accuracy to manual and feature-based analysis with substantial time savings, eliminating workflow complexity. This streamlined approach holds significant clinical value, enhancing IFC accessibility for routine diagnostic purposes.

Label-free cell detection of acute leukemia using ghost cytometry

Early diagnosis and prompt initiation of appropriate treatment are critical for improving the prognosis of acute leukemia. Acute leukemia is diagnosed by microscopic morphological examination of bone marrow smears and flow cytometric immunophenotyping of bone marrow cells stained with fluorophore-conjugated antibodies. However, these diagnostic processes require trained professionals and are time and resource-intensive. Here, we present a novel diagnostic approach using ghost cytometry, a recently developed high-content flow cytometric approach, which enables machine vision-based, stain-free, high-speed analysis of cells, leveraging their detailed morphological information. We demonstrate that ghost cytometry can detect leukemic cells from the bone marrow cells of patients diagnosed with acute lymphoblastic leukemia and acute myeloid leukemia without relying on biological staining. The approach presented here holds promise as a precise, simple, swift, and cost-effective diagnostic method for acute leukemia in clinical practice.

PXPermute reveals staining importance in multichannel imaging flow cytometry

Imaging flow cytometry (IFC) allows rapid acquisition of numerous single-cell images per second, capturing information from multiple fluorescent channels. However, the traditional process of staining cells with fluorescently labeled conjugated antibodies for IFC analysis is time consuming, expensive, and potentially harmful to cell viability. To streamline experimental workflows and reduce costs, it is crucial to identify the most relevant channels for downstream analysis. In this study, we introduce PXPermute, a user-friendly and powerful method for assessing the significance of IFC channels, particularly for cell profiling. Our approach evaluates channel importance by permuting pixel values within each channel and analyzing the resulting impact on machine learning or deep learning models. Through rigorous evaluation of three multichannel IFC image datasets, we demonstrate PXPermute's potential in accurately identifying the most informative channels, aligning with established biological knowledge. PXPermute can assist biologists with systematic channel analysis, experimental design optimization, and biomarker identification.

DeepIFC: Virtual fluorescent labeling of blood cells in imaging flow cytometry data with deep learning

Imaging flow cytometry (IFC) combines flow cytometry with microscopy, allowing rapid characterization of cellular and molecular properties via high-throughput single-cell fluorescent imaging. However, fluorescent labeling is costly and time-consuming. We present a computational method called DeepIFC based on the Inception U-Net neural network architecture, able to generate fluorescent marker images and learn morphological features from IFC brightfield and darkfield images. Furthermore, the DeepIFC workflow identifies cell types from the generated fluorescent images and visualizes the single-cell features generated in a 2D space. We demonstrate that rarer cell types are predicted well when a balanced data set is used to train the model, and the model is able to recognize red blood cells not seen during model training as a distinct entity. In summary, DeepIFC allows accurate cell reconstruction, typing and recognition of unseen cell types from brightfield and darkfield images via virtual fluorescent labeling.

Assessment of stored red blood cells through lab-on-a-chip technologies for precision transfusion medicine

Transfusion of red blood cells (RBCs) is one of the most valuable and widespread treatments in modern medicine. Lifesaving RBC transfusions are facilitated by the cold storage of RBC units in blood banks worldwide. Currently, RBC storage and subsequent transfusion practices are performed using simplistic workflows. More specifically, most blood banks follow the "first-in-first-out" principle to avoid wastage, whereas most healthcare providers prefer the "last-in-first-out" approach simply favoring chronologically younger RBCs. Neither approach addresses recent advances through -omics showing that stored RBC quality is highly variable depending on donor-, time-, and processing-specific factors. Thus, it is time to rethink our workflows in transfusion medicine taking advantage of novel technologies to perform RBC quality assessment. We imagine a future where lab-on-a-chip technologies utilize novel predictive markers of RBC quality identified by -omics and machine learning to usher in a new era of safer and precise transfusion medicine.

Leveraging machine learning to identify acute myeloid leukemia patients and their chemotherapy regimens in an administrative database

BACKGROUND

Administrative datasets are useful for identifying rare disease cohorts such as pediatric acute myeloid leukemia (AML). Previously, cohorts were assembled using labor-intensive, manual reviews of patients' longitudinal chemotherapy data.

METHODS

We utilized a two-step machine learning (ML) method to (i) identify pediatric patients with newly diagnosed AML, and (ii) among the identified AML patients, their chemotherapy courses, in an administrative/billing database. Using 2558 patients previously manually reviewed, multiple ML algorithms were derived from 75% of the study sample, and the selected model was tested in the remaining hold-out sample. The selected model was also applied to assemble a new pediatric AML cohort and further assessed in an external validation, using a standalone cohort established by manual chart abstraction.

RESULTS

For patient identification, the selected Support Vector Machine model yielded a sensitivity of 0.97 and a positive predictive value (PPV) of 0.97 in the hold-out test sample. For course-specific chemotherapy regimen and start date identification, the selected Random Forest model yielded overall PPV greater than or equal to 0.88 and sensitivity greater than or equal to 0.86 across all courses in the test sample. When applied to new cohort assembly, ML identified 3016 AML patients with 10,588 treatment courses. In the external validation subset, PPV was greater than or equal to 0.75 and sensitivity was greater than or equal to 0.82 for patient identification, and PPV was greater than or equal to 0.93 and sensitivity was greater than or equal to 0.94 for regimen identifications.

CONCLUSION

A carefully designed ML model can accurately identify pediatric AML patients and their chemotherapy courses from administrative databases. This approach may be generalizable to other diseases and databases.

Typing of acute leukemia by intelligent optical time-stretch imaging flow cytometry on a chip

Acute leukemia (AL) is one of the top life-threatening diseases. Accurate typing of AL can significantly improve its prognosis. However, conventional methods for AL typing often require cell staining, which is time-consuming and labor-intensive. Furthermore, their performance is highly limited by the specificity and availability of fluorescent labels, which can hardly meet the requirements of AL typing in clinical settings. Here, we demonstrate AL typing by intelligent optical time-stretch (OTS) imaging flow cytometry on a microfluidic chip. Specifically, we employ OTS microscopy to capture the images of cells in clinical bone marrow samples with a spatial resolution of 780 nm at a high flowing speed of 1 m s^-1 in a label-free manner. Then, to show the clinical utility of our method for which the features of clinical samples are diverse, we design and construct a deep convolutional neural network (CNN) to analyze the cellular images and determine the AL type of each sample. We measure 30 clinical samples composed of 7 acute lymphoblastic leukemia (ALL) samples, 17 acute myelogenous leukemia (AML) samples, and 6 samples from healthy donors, resulting in a total of 227 620 images acquired. Results show that our method can distinguish ALL and AML with an accuracy of 95.03%, which, to the best of our knowledge, is a record in label-free AL typing. In addition to AL typing, we believe that the high throughput, high accuracy, and label-free operation of our method make it a potential solution for cell analysis in scientific research and clinical settings.

Detection of acute promyelocytic leukemia in peripheral blood and bone marrow with annotation-free deep learning

While optical microscopy inspection of blood films and bone marrow aspirates by a hematologist is a crucial step in establishing diagnosis of acute leukemia, especially in low-resource settings where other diagnostic modalities are not available, the task remains time-consuming and prone to human inconsistencies. This has an impact especially in cases of Acute Promyelocytic Leukemia (APL) that require urgent treatment. Integration of automated computational hematopathology into clinical workflows can improve the throughput of these services and reduce cognitive human error. However, a major bottleneck in deploying such systems is a lack of sufficient cell morphological object-labels annotations to train deep learning models. We overcome this by leveraging patient diagnostic labels to train weakly-supervised models that detect different types of acute leukemia. We introduce a deep learning approach, Multiple Instance Learning for Leukocyte Identification (MILLIE), able to perform automated reliable analysis of blood films with minimal supervision. Without being trained to classify individual cells, MILLIE differentiates between acute lymphoblastic and myeloblastic leukemia in blood films. More importantly, MILLIE detects APL in blood films (AUC 0.94 ± 0.04) and in bone marrow aspirates (AUC 0.99 ± 0.01). MILLIE is a viable solution to augment the throughput of clinical pathways that require assessment of blood film microscopy.

Combining imaging flow cytometry and machine learning for high-throughput schistocyte quantification: A SVM classifier development and external validation cohort

Background

Schistocyte counts are a cornerstone of the diagnosis of thrombotic microangiopathy syndrome (TMA). Their manual quantification is complex and alternative automated methods suffer from pitfalls that limit their use. We report a method combining imaging flow cytometry (IFC) and artificial intelligence for the direct label-free and operator-independent quantification of schistocytes in whole blood.

Methods

We used 135,045 IFC images from blood acquisition among 14 patients to extract 188 features with IDEAS® software and 128 features from a convolutional neural network (CNN) with Keras framework in order to train a support vector machine (SVM) blood elements’ classifier used for schistocytes quantification.

Finding

Keras features showed better accuracy (94.03%, CI: 93.75-94.31%) than ideas features (91.54%, CI: 91.21-91.87%) in recognising whole-blood elements, and together they showed the best accuracy (95.64%, CI: 95.39-95.88%). We obtained an excellent correlation (0.93, CI: 0.90-0.96) between three haematologists and our method on a cohort of 102 patient samples. All patients with schistocytosis (>1% schistocytes) were detected with excellent specificity (91.3%, CI: 82.0-96.7%) and sensitivity (100%, CI: 89.4-100.0%). We confirmed these results with a similar specificity (91.1%, CI: 78.8-97.5%) and sensitivity (100%, CI: 88.1-100.0%) on a validation cohort (n=74) analysed in an independent healthcare centre. Simultaneous analysis of 16 samples in both study centres showed a very good correlation between the 2 imaging flow cytometers (Y=1.001x).

Interpretation

We demonstrate that IFC can represent a reliable tool for operator-independent schistocyte quantification with no pre-analytical processing which is of most importance in emergency situations such as TMA.

Funding

None.

Artificial intelligence applied to musculoskeletal oncology: a systematic review

Developments in artificial intelligence have the potential to improve the care of patients with musculoskeletal tumors. We performed a systematic review of the published scientific literature to identify the current state of the art of artificial intelligence applied to musculoskeletal oncology, including both primary and metastatic tumors, and across the radiology, nuclear medicine, pathology, clinical research, and molecular biology literature. Through this search, we identified 252 primary research articles, of which 58 used deep learning and 194 used other machine learning techniques. Articles involving deep learning have mostly involved bone scintigraphy, histopathology, and radiologic imaging. Articles involving other machine learning techniques have mostly involved transcriptomic analyses, radiomics, and clinical outcome prediction models using medical records. These articles predominantly present proof-of-concept work, other than the automated bone scan index for bone metastasis quantification, which has translated to clinical workflows in some regions. We systematically review and discuss this literature, highlight opportunities for multidisciplinary collaboration, and identify potentially clinically useful topics with a relative paucity of research attention. Musculoskeletal oncology is an inherently multidisciplinary field, and future research will need to integrate and synthesize noisy siloed data from across clinical, imaging, and molecular datasets. Building the data infrastructure for collaboration will help to accelerate progress towards making artificial intelligence truly useful in musculoskeletal oncology.

Applications of Artificial Intelligence in Pediatric Oncology: A Systematic Review

PURPOSE

There is a need for an improved understanding of clinical and biologic risk factors in pediatric cancer to improve patient outcomes. Machine learning (ML) represents the application of computational inference from advanced statistical methods that can be applied to increasing amount of data available for study in pediatric oncology. The goal of this systematic review was to systematically characterize the state of ML in pediatric oncology and highlight advances and opportunities in the field.

METHODS

We conducted a systematic review of the Embase, Scopus, and MEDLINE databases for applications of ML in pediatric oncology. Query results from all three databases were aggregated and duplicate studies were removed.

RESULTS

A total of 42 unique articles that examined the applications of ML in pediatric oncology met inclusion criteria for review. We identified 20 studies of CNS tumors, 13 of solid tumors, and nine of leukemia. ML tasks included classification, prediction of treatment response, and dose optimization with a variety of methods being used including neural network, k-nearest neighbor, random forest, naive Bayes, and support vector machines. Strengths of the identified studies included matching or outperforming physician comparators via automated analysis and predicting therapeutic response. Common limitations included significant heterogeneity in reporting standards, clinical applicability, small sample sizes, and missing external validation cohorts.

CONCLUSION

We identified areas where ML can enhance clinical care in ways that may not otherwise be achievable. Although ML promises enormous potential in improving diagnostics, decision making, and monitoring for children with cancer, the field remains in early stages and future work will be aided by standards and guidelines to ensure rigorous methodologic design and maximizing clinical utility.

A weakly supervised deep learning approach for label-free imaging flow-cytometry-based blood diagnostics

The application of machine learning approaches to imaging flow cytometry (IFC) data has the potential to transform the diagnosis of hematological diseases. However, the need for manually labeled single-cell images for machine learning model training has severely limited its clinical application. To address this, we present iCellCnn, a weakly supervised deep learning approach for label-free IFC-based blood diagnostics. We demonstrate the capability of iCellCnn to achieve diagnosis of Sézary syndrome (SS) from patient samples on the basis of bright-field IFC images of T cells obtained after fluorescence-activated cell sorting of human peripheral blood mononuclear cell specimens. With a sample size of four healthy donors and five SS patients, iCellCnn achieved a 100% classification accuracy. As iCellCnn is not restricted to the diagnosis of SS, we expect such weakly supervised approaches to tap the diagnostic potential of IFC by providing automatic data-driven diagnosis of diseases with so-far unknown morphological manifestations.

Imaging flow cytometry reveals a subset of TdT negative T-ALL blasts with very low forward scatter on conventional flow cytometry

BACKGROUND

Studies in T-cell acute lymphoblastic leukemia (T-ALL) have shown that leukemic blast populations may display immunophenotypic heterogeneity. In the clinical setting, evaluation of measurable residual disease during treatment and follow-up is highly dependent on knowledge of the diversity of blast subsets. Here, we set out to evaluate whether variation in expression of the blast marker, TdT, in T-ALL blasts could correspond to differences in morphometric features.

METHODS

We investigated diagnostic bone marrow samples from six individual T-ALL patients run in parallel on imaging flow cytometry (IFC) and conventional flow cytometry (CFC).

RESULTS

Guided by the imagery available in IFC, we identified distinct TdT^neg and TdT^pos subpopulations with apparent differences in internal complexity. As TdT^neg blasts predominantly displayed very low forward scatter (FSC) on CFC, these subsets were initially excluded from routine analysis as debris, elements of small diameter, apoptotic, and/or dead cells. However, IFC-based morphometric analyses demonstrated that cell size and shape of TdT^neg blasts were comparable to the TdT^pos cells and without morphometric apoptotic hallmarks, supporting that the TdT^neg subpopulation corresponded to T-ALL blasts. Fluorescence in situ hybridization analyses substantiated the clinical relevance of TdT^neg FSC^very-low cells by retrieving known diagnostic cytogenetic abnormalities at comparable frequencies in purified TdT^neg FSC^very-low and TdT^pos FSC^int subsets.

CONCLUSION

We highlight this finding as knowledge of phenotypic heterogeneity is of crucial importance in the clinical setting for delineation and quantification of blast subpopulations of potential biological relevance. We argue that the IFC imagery may allow for visual verification and improvement of applied gating strategies.

Deep Learning in Biomedical Optics

This article reviews deep learning applications in biomedical optics with a particular emphasis on image formation. The review is organized by imaging domains within biomedical optics and includes microscopy, fluorescence lifetime imaging, in vivo microscopy, widefield endoscopy, optical coherence tomography, photoacoustic imaging, diffuse tomography, and functional optical brain imaging. For each of these domains, we summarize how deep learning has been applied and highlight methods by which deep learning can enable new capabilities for optics in medicine. Challenges and opportunities to improve translation and adoption of deep learning in biomedical optics are also summarized. Lasers Surg. Med. © 2021 Wiley Periodicals LLC.

Deepometry, a framework for applying supervised and weakly supervised deep learning to imaging cytometry

Deep learning offers the potential to extract more than meets the eye from images captured by imaging flow cytometry. This protocol describes the application of deep learning to single-cell images to perform supervised cell classification and weakly supervised learning, using example data from an experiment exploring red blood cell morphology. We describe how to acquire and transform suitable input data as well as the steps required for deep learning training and inference using an open-source web-based application. All steps of the protocol are provided as open-source Python as well as MATLAB runtime scripts, through both command-line and graphic user interfaces. The protocol enables a flexible and friendly environment for morphological phenotyping using supervised and weakly supervised learning and the subsequent exploration of the deep learning features using multi-dimensional visualization tools. The protocol requires 40 h when training from scratch and 1 h when using a pre-trained model.

An image-based flow cytometric approach to the assessment of the nucleus-to-cytoplasm ratio

The nucleus-to-cytoplasm ratio (N:C) can be used as one metric in histology for grading certain types of tumor malignancy. Current N:C assessment techniques are time-consuming and low throughput. Thus, in high-throughput clinical contexts, there is a need for a technique that can assess cell malignancy rapidly. In this study, we assess the N:C ratio of four different malignant cell lines (OCI-AML-5-blood cancer, CAKI-2-kidney cancer, HT-29-colon cancer, SK-BR-3-breast cancer) and a non-malignant cell line (MCF-10A -breast epithelium) using an imaging flow cytometer (IFC). Cells were stained with the DRAQ-5 nuclear dye to stain the cell nucleus. An Amnis ImageStreamX® IFC acquired brightfield/fluorescence images of cells and their nuclei, respectively. Masking and gating techniques were used to obtain the cell and nucleus diameters for 5284 OCI-AML-5 cells, 1096 CAKI-2 cells, 6302 HT-29 cells, 3159 SK-BR-3 cells, and 1109 MCF-10A cells. The N:C ratio was calculated as the ratio of the nucleus diameter to the total cell diameter. The average cell and nucleus diameters from IFC were 12.3 ± 1.2 μm and 9.0 ± 1.1 μm for OCI-AML5 cells, 24.5 ± 2.6 μm and 15.6 ± 2.1 μm for CAKI-2 cells, 16.2 ± 1.8 μm and 11.2 ± 1.3 μm for HT-29 cells, 18.0 ± 3.7 μm and 12.5 ± 2.1 μm for SK-BR-3 cells, and 19.4 ± 2.2 μm and 10.1 ± 1.8 μm for MCF-10A cells. Here we show a general N:C ratio of ~0.6-0.7 across varying malignant cell lines and a N:C ratio of ~0.5 for a non-malignant cell line. This study demonstrates the use of IFC to assess the N:C ratio of cancerous and non-cancerous cells, and the promise of its use in clinically relevant high-throughput detection scenarios to supplement current workflows used for cancer cell grading.

Image-based cell phenotyping with deep learning

A cell's phenotype is the culmination of several cellular processes through a complex network of molecular interactions that ultimately result in a unique morphological signature. Visual cell phenotyping is the characterization and quantification of these observable cellular traits in images. Recently, cellular phenotyping has undergone a massive overhaul in terms of scale, resolution, and throughput, which is attributable to advances across electronic, optical, and chemical technologies for imaging cells. Coupled with the rapid acceleration of deep learning-based computational tools, these advances have opened up new avenues for innovation across a wide variety of high-throughput cell biology applications. Here, we review applications wherein deep learning is powering the recognition, profiling, and prediction of visual phenotypes to answer important biological questions. As the complexity and scale of imaging assays increase, deep learning offers computational solutions to elucidate the details of previously unexplored cellular phenotypes.

Toward deep biophysical cytometry: prospects and challenges

The biophysical properties of cells reflect their identities, underpin their homeostatic state in health, and define the pathogenesis of disease. Recent leapfrogging advances in biophysical cytometry now give access to this information, which is obscured in molecular assays, with a discriminative power that was once inconceivable. However, biophysical cytometry should go 'deeper' in terms of exploiting the information-rich cellular biophysical content, generating a molecular knowledge base of cellular biophysical properties, and standardizing the protocols for wider dissemination. Overcoming these barriers, which requires concurrent innovations in microfluidics, optical imaging, and computer vision, could unleash the enormous potential of biophysical cytometry not only for gaining a new mechanistic understanding of biological systems but also for identifying new cost-effective biomarkers of disease.

Emerging use of machine learning and advanced technologies to assess red cell quality

Improving blood product quality and patient outcomes is an accepted goal in transfusion medicine research. Thus, there is an urgent need to understand the potential adverse effects on red blood cells (RBCs) during pre-transfusion storage. Current assessment techniques of these degradation events, termed "storage lesions", are subjective, labor-intensive, and complex. Here we describe emerging technologies that assess the biochemical, biophysical, and morphological characteristics of RBC storage lesions. Of these emerging techniques, machine learning (ML) has shown potential to overcome the limitations of conventional RBC assessment methods. Our previous work has shown that neural networks can extract chronological progressions of morphological changes in RBCs during storage without human input. We hypothesize that, with broader training and testing of multivariate data (e.g., varying donor factors and manufacturing methods), ML can further our understanding of clinical transfusion outcomes in multiple patient groups.

Identification and Staging of B-Cell Acute Lymphoblastic Leukemia Using Quantitative Phase Imaging and Machine Learning

Identification and classification of leukemia cells in a rapid and label-free fashion is clinically challenging and thus presents a prime arena for implementing new diagnostic tools. Quantitative phase imaging, which maps optical path length delays introduced by the specimen, has been demonstrated to discern cellular phenotypes based on differential morphological attributes. Rapid acquisition capability and the availability of label-free images with high information content have enabled researchers to use machine learning (ML) to reveal latent features. We developed a set of ML classifiers, including convolutional neural networks, to discern healthy B cells from lymphoblasts and classify stages of B cell acute lymphoblastic leukemia. Here, we show that the average dry mass and volume of normal B cells are lower than those of cancerous cells and that these morphologic parameters increase further alongside disease progression. We find that the relaxed training requirements of a ML approach are conducive to the classification of cell type, with minimal space, training time, and memory requirements. Our findings pave the way for a larger study on clinical samples of acute lymphoblastic leukemia, with the overarching goal of its broader use in hematopathology, where the prospect of objective diagnoses with minimal sample preparation remains highly desirable.

Computer-aided detection of COVID-19 from X-ray images using multi-CNN and Bayesnet classifier

Corona virus disease-2019 (COVID-19) is a pandemic caused by novel coronavirus. COVID-19 is spreading rapidly throughout the world. The gold standard for diagnosing COVID-19 is reverse transcription-polymerase chain reaction (RT-PCR) test. However, the facility for RT-PCR test is limited, which causes early diagnosis of the disease difficult. Easily available modalities like X-ray can be used to detect specific symptoms associated with COVID-19. Pre-trained convolutional neural networks are widely used for computer-aided detection of diseases from smaller datasets. This paper investigates the effectiveness of multi-CNN, a combination of several pre-trained CNNs, for the automated detection of COVID-19 from X-ray images. The method uses a combination of features extracted from multi-CNN with correlation based feature selection (CFS) technique and Bayesnet classifier for the prediction of COVID-19. The method was tested using two public datasets and achieved promising results on both the datasets. In the first dataset consisting of 453 COVID-19 images and 497 non-COVID images, the method achieved an AUC of 0.963 and an accuracy of 91.16%. In the second dataset consisting of 71 COVID-19 images and 7 non-COVID images, the method achieved an AUC of 0.911 and an accuracy of 97.44%. The experiments performed in this study proved the effectiveness of pre-trained multi-CNN over single CNN in the detection of COVID-19.

Objective assessment of stored blood quality by deep learning

Stored red blood cells (RBCs) are needed for life-saving blood transfusions, but they undergo continuous degradation. RBC storage lesions are often assessed by microscopic examination or biochemical and biophysical assays, which are complex, time-consuming, and destructive to fragile cells. Here we demonstrate the use of label-free imaging flow cytometry and deep learning to characterize RBC lesions. Using brightfield images, a trained neural network achieved 76.7% agreement with experts in classifying seven clinically relevant RBC morphologies associated with storage lesions, comparable to 82.5% agreement between different experts. Given that human observation and classification may not optimally discern RBC quality, we went further and eliminated subjective human annotation in the training step by training a weakly supervised neural network using only storage duration times. The feature space extracted by this network revealed a chronological progression of morphological changes that better predicted blood quality, as measured by physiological hemolytic assay readouts, than the conventional expert-assessed morphology classification system. With further training and clinical testing across multiple sites, protocols, and instruments, deep learning and label-free imaging flow cytometry might be used to routinely and objectively assess RBC storage lesions. This would automate a complex protocol, minimize laboratory sample handling and preparation, and reduce the impact of procedural errors and discrepancies between facilities and blood donors. The chronology-based machine-learning approach may also improve upon humans' assessment of morphological changes in other biomedically important progressions, such as differentiation and metastasis.

Image-based high-content screening in drug discovery

While target-based drug discovery strategies rely on the precise knowledge of the identity and function of the drug targets, phenotypic drug discovery (PDD) approaches allow the identification of novel drugs based on knowledge of a distinct phenotype. Image-based high-content screening (HCS) is a potent PDD strategy that characterizes small-molecule effects through the quantification of features that depict cellular changes among or within cell populations, thereby generating valuable data sets for subsequent data analysis. However, these data can be complex, making image analysis from large HCS campaigns challenging. Technological advances in image acquisition, processing, and analysis as well as machine-learning (ML) approaches for the analysis of multidimensional data sets have rendered HCS as a viable technology for small-molecule drug discovery. Here, we discuss HCS concepts, current workflows as well as opportunities and challenges of image-based phenotypic screening and data analysis.

Screening as a strategy to drive regenerative medicine research

In the field of regenerative medicine, optimization of the parameters leading to a desirable outcome remains a huge challenge. Examples include protocols for the guided differentiation of pluripotent cells towards specialized and functional cell types, phenotypic maintenance of primary cells in cell culture, or engineering of materials for improved tissue interaction with medical implants. This challenge originates from the enormous design space for biomaterials, chemical and biochemical compounds, and incomplete knowledge of the guiding biological principles. To tackle this challenge, high-throughput platforms allow screening of multiple perturbations in one experimental setup. In this review, we provide an overview of screening platforms that are used in regenerative medicine. We discuss their fabrication techniques, and in silico tools to analyze the extensive data sets typically generated by these platforms.