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.

Optical mesoscopy, machine learning, and computational microscopy enable high information content diagnostic imaging of blood films.

Automated image-based assessment of blood films has tremendous potential to support clinical haematology within overstretched healthcare systems. To achieve this, efficient and reliable digital capture of the rich diagnostic information contained within a blood film is a critical first step. However, this is often challenging, and in many cases entirely unfeasible, with the microscopes typically used in haematology due to the fundamental trade-off between magnification and spatial resolution. To address this, we investigated three state-of-the-art approaches to microscopic imaging of blood films which leverage recent advances in optical and computational imaging and analysis to increase the information capture capacity of the optical microscope: optical mesoscopy, which uses a giant microscope objective (Mesolens) to enable high-resolution imaging at low magnification; Fourier ptychographic microscopy, a computational imaging method which relies on oblique illumination with a series of LEDs to capture high-resolution information; and deep neural networks which can be trained to increase the quality of low magnification, low resolution images. We compare and contrast the performance of these techniques for blood film imaging for the exemplar case of Giemsa-stained peripheral blood smears. Using computational image analysis and shape-based object classification, we demonstrate their use for automated analysis of red blood cell morphology and visualization and detection of small blood-borne parasites such as the malarial parasite Plasmodium falciparum. Our results demonstrate that these new methods greatly increase the information capturing capacity of the light microscope, with transformative potential for haematology and more generally across digital pathology. © 2021 The Authors. The Journal of Pathology published by John Wiley & Sons, Ltd. on behalf of The Pathological Society of Great Britain and Ireland.

Structure-dependent amplification for denoising and background correction in Fourier ptychographic microscopy.

Fourier Ptychographic Microscopy (FPM) allows high resolution imaging using iterative phase retrieval to recover an estimate of the complex object from a series of images captured under oblique illumination. FPM is particularly sensitive to noise and uncorrected background signals as it relies on combining information from brightfield and noisy darkfield (DF) images. In this article we consider the impact of different noise sources in FPM and show that inadequate removal of the DF background signal and associated noise are the predominant cause of artefacts in reconstructed images. We propose a simple solution to FPM background correction and denoising that outperforms existing methods in terms of image quality, speed and simplicity, whilst maintaining high spatial resolution and sharpness of the reconstructed image. Our method takes advantage of the data redundancy in real space within the acquired dataset to boost the signal-to-background ratio in the captured DF images, before optimally suppressing background signal. By incorporating differentially denoised images within the classic FPM iterative phase retrieval algorithm, we show that it is possible to achieve efficient removal of background artefacts without suppression of high frequency information. The method is tested using simulated data and experimental images of thin blood films, bone marrow and liver tissue sections. Our approach is non-parametric, requires no prior knowledge of the noise distribution and can be directly applied to other hardware platforms and reconstruction algorithms making it widely applicable in FPM.

Expert-level automated malaria diagnosis on routine blood films with deep neural networks.

Over 200 million malaria cases globally lead to half a million deaths annually. Accurate malaria diagnosis remains a challenge. Automated imaging processing approaches to analyze Thick Blood Films (TBF) could provide scalable solutions, for urban healthcare providers in the holoendemic malaria sub-Saharan region. Although several approaches have been attempted to identify malaria parasites in TBF, none have achieved negative and positive predictive performance suitable for clinical use in the west sub-Saharan region. While malaria parasite object detection remains an intermediary step in achieving automatic patient diagnosis, training state-of-the-art deep-learning object detectors requires the human-expert labor-intensive process of labeling a large dataset of digitized TBF. To overcome these challenges and to achieve a clinically usable system, we show a novel approach. It leverages routine clinical-microscopy labels from our quality-controlled malaria clinics, to train a Deep Malaria Convolutional Neural Network classifier (DeepMCNN) for automated malaria diagnosis. Our system also provides total Malaria Parasite (MP) and White Blood Cell (WBC) counts allowing parasitemia estimation in MP/µL, as recommended by the WHO. Prospective validation of the DeepMCNN achieves sensitivity/specificity of 0.92/0.90 against expert-level malaria diagnosis. Our approach PPV/NPV performance is of 0.92/0.90, which is clinically usable in our holoendemic settings in the densely populated metropolis of Ibadan. It is located within the most populous African country (Nigeria) and with one of the largest burdens of Plasmodium falciparum malaria. Our openly available method is of importance for strategies aimed to scale malaria diagnosis in urban regions where daily assessment of thousands of specimens is required.

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.

Network topological determinants of pathogen spread.

How do we best constrain social interactions to decrease transmission of communicable diseases? Indiscriminate suppression is unsustainable long term and presupposes that all interactions carry equal importance. Instead, transmission within a social network has been shown to be determined by its topology. In this paper, we deploy simulations to understand and quantify the impact on disease transmission of a set of topological network features, building a dataset of 9000 interaction graphs using generators of different types of synthetic social networks. Independently of the topology of the network, we maintain constant the total volume of social interactions in our simulations, to show how even with the same social contact some network structures are more or less resilient to the spread. We find a suitable intervention to be specific suppression of unfamiliar and casual interactions that contribute to the network's global efficiency. This is, pathogen spread is significantly reduced by limiting specific kinds of contact rather than their global number. Our numerical studies might inspire further investigation in connection to public health, as an integrative framework to craft and evaluate social interventions in communicable diseases with different social graphs or as a highlight of network metrics that should be captured in social studies.

Content aware multi-focus image fusion for high-magnification blood film microscopy.

Automated digital high-magnification optical microscopy is key to accelerating biology research and improving pathology clinical pathways. High magnification objectives with large numerical apertures are usually preferred to resolve the fine structural details of biological samples, but they have a very limited depth-of-field. Depending on the thickness of the sample, analysis of specimens typically requires the acquisition of multiple images at different focal planes for each field-of-view, followed by the fusion of these planes into an extended depth-of-field image. This translates into low scanning speeds, increased storage space, and processing time not suitable for high-throughput clinical use. We introduce a novel content-aware multi-focus image fusion approach based on deep learning which extends the depth-of-field of high magnification objectives effectively. We demonstrate the method with three examples, showing that highly accurate, detailed, extended depth of field images can be obtained at a lower axial sampling rate, using 2-fold fewer focal planes than normally required.

Digital refocusing and extended depth of field reconstruction in Fourier ptychographic microscopy.

Fourier ptychography microscopy (FPM) is a recently developed microscopic imaging method that allows the recovery of a high-resolution complex image by combining a sequence of bright and darkfield images acquired under inclined illumination. The capacity of FPM for high resolution imaging at low magnification makes it particularly attractive for applications in digital pathology which require imaging of large specimens such as tissue sections and blood films. To date most applications of FPM have been limited to imaging thin samples, simplifying both image reconstruction and analysis. In this work we show that, for samples of intermediate thickness (defined here as less than the depth of field of a raw captured image), numerical propagation of the reconstructed complex field allows effective digital refocusing of FPM images. The results are validated by comparison against images obtained with an equivalent high numerical aperture objective lens. We find that post reconstruction refocusing (PRR) yields images comparable in quality to adding a defocus term to the pupil function within the reconstruction algorithm, while reducing computing time by several orders of magnitude. We apply PRR to visualize FPM images of Giemsa-stained peripheral blood films and present a novel image processing pipeline to construct an effective extended depth of field image which optimally displays the 3D sample structure in a 2D image. We also show how digital refocusing allows effective correction of the chromatic focus shifts inherent to the low magnification objective lenses used in FPM setups, improving the overall quality of color FPM images.

Comparison of Artificial Intelligence based approaches to cell function prediction.

Predicting Retinal Pigment Epithelium (RPE) cell functions in stem cell implants using non-invasive bright field microscopy imaging is a critical task for clinical deployment of stem cell therapies. Such cell function predictions can be carried out using Artificial Intelligence (AI) based models. In this paper we used Traditional Machine Learning (TML) and Deep Learning (DL) based AI models for cell function prediction tasks. TML models depend on feature engineering and DL models perform feature engineering automatically but have higher modeling complexity. This work aims at exploring the tradeoffs between three approaches using TML and DL based models for RPE cell function prediction from microscopy images and at understanding the accuracy relationship between pixel-, cell feature-, and implant label-level accuracies of models. Among the three compared approaches to cell function prediction, the direct approach to cell function prediction from images is slightly more accurate in comparison to indirect approaches using intermediate segmentation and/or feature engineering steps. We also evaluated accuracy variations with respect to model selections (five TML models and two DL models) and model configurations (with and without transfer learning). Finally, we quantified the relationships between segmentation accuracy and the number of samples used for training a model, segmentation accuracy and cell feature error, and cell feature error and accuracy of implant labels. We concluded that for the RPE cell data set, there is a monotonic relationship between the number of training samples and image segmentation accuracy, and between segmentation accuracy and cell feature error, but there is no such a relationship between segmentation accuracy and accuracy of RPE implant labels.

Data-driven malaria prevalence prediction in large densely populated urban holoendemic sub-Saharan West Africa.

Over 200 million malaria cases globally lead to half-million deaths annually. The development of malaria prevalence prediction systems to support malaria care pathways has been hindered by lack of data, a tendency towards universal "monolithic" models (one-size-fits-all-regions) and a focus on long lead time predictions. Current systems do not provide short-term local predictions at an accuracy suitable for deployment in clinical practice. Here we show a data-driven approach that reliably produces one-month-ahead prevalence prediction within a densely populated all-year-round malaria metropolis of over 3.5 million inhabitants situated in Nigeria which has one of the largest global burdens of P. falciparum malaria. We estimate one-month-ahead prevalence in a unique 22-years prospective regional dataset of > 9 × 104 participants attending our healthcare services. Our system agrees with both magnitude and direction of the prediction on validation data achieving MAE ≤ 6 × 10-2, MSE ≤ 7 × 10-3, PCC (median 0.63, IQR 0.3) and with more than 80% of estimates within a (+ 0.1 to - 0.05) error-tolerance range which is clinically relevant for decision-support in our holoendemic setting. Our data-driven approach could facilitate healthcare systems to harness their own data to support local malaria care pathways.

Accurate and interpretable classification of microspectroscopy pixels using artificial neural networks.

This paper addresses the problem of classifying materials from microspectroscopy at a pixel level. The challenges lie in identifying discriminatory spectral features and obtaining accurate and interpretable models relating spectra and class labels. We approach the problem by designing a supervised classifier from a tandem of Artificial Neural Network (ANN) models that identify relevant features in raw spectra and achieve high classification accuracy. The tandem of ANN models is meshed with classification rule extraction methods to lower the model complexity and to achieve interpretability of the resulting model. The contribution of the work is in designing each ANN model based on the microspectroscopy hypothesis about a discriminatory feature of a certain target class being composed of a linear combination of spectra. The novelty lies in meshing ANN and decision rule models into a tandem configuration to achieve accurate and interpretable classification results. The proposed method was evaluated using a set of broadband coherent anti-Stokes Raman scattering (BCARS) microscopy cell images (600 000  pixel-level spectra) and a reference four-class rule-based model previously created by biochemical experts. The generated classification rule-based model was on average 85% accurate measured by the DICE pixel label similarity metric, and on average 96% similar to the reference rules measured by the vector cosine metric.

Four-dimensional radiotherapeutic dose calculation using biomechanical respiratory motion description.

PURPOSE: Organ motion due to patient breathing introduces a technical challenge for dosimetry and lung tumor treatment by hadron therapy. Accurate dose distribution estimation requires patient-specific information on tumor position, size, and shape as well as information regarding the material density and stopping power of the media along the beam path. A new 4D dosimetry method was developed, which can be coupled to any motion estimation method. As an illustration, the new method was implemented and tested with a biomechanical model and clinical data. METHODS: First, an anatomical model of the lung and tumor was synthesized with deformable tetrahedral grids using computed tomography (CT) images. The CT attenuation values were estimated at the grid vertices. Respiratory motion was simulated biomechanically based on nonlinear finite element analysis. Contrary to classical image-based methods where motion is described using deformable image registration algorithms, the dose distribution was accumulated over tetrahedral meshes that are deformed using biomechanical modeling based on finite element analysis. RESULTS: The new method preserves the mass of the objects during simulation with an error between 1.6 and 3.6%. The new method was compared to an existing dose calculation method demonstrating significant differences between the two approaches and overall superior performance using the new method. CONCLUSION: A unified model of 4D radiotherapy respiratory effects was developed where biomechanical simulations are coupled with dose calculations. Promising results demonstrate that this approach has significant potential for the treatment for moving tumors.