Independent real-world application of a clinical-grade automated prostate cancer detection system.

Artificial intelligence (AI)-based systems applied to histopathology whole-slide images have the potential to improve patient care through mitigation of challenges posed by diagnostic variability, histopathology caseload, and shortage of pathologists. We sought to define the performance of an AI-based automated prostate cancer detection system, Paige Prostate, when applied to independent real-world data. The algorithm was employed to classify slides into two categories: benign (no further review needed) or suspicious (additional histologic and/or immunohistochemical analysis required). We assessed the sensitivity, specificity, positive predictive values (PPVs), and negative predictive values (NPVs) of a local pathologist, two central pathologists, and Paige Prostate in the diagnosis of 600 transrectal ultrasound-guided prostate needle core biopsy regions ('part-specimens') from 100 consecutive patients, and to ascertain the impact of Paige Prostate on diagnostic accuracy and efficiency. Paige Prostate displayed high sensitivity (0.99; CI 0.96-1.0), NPV (1.0; CI 0.98-1.0), and specificity (0.93; CI 0.90-0.96) at the part-specimen level. At the patient level, Paige Prostate displayed optimal sensitivity (1.0; CI 0.93-1.0) and NPV (1.0; CI 0.91-1.0) at a specificity of 0.78 (CI 0.64-0.89). The 27 part-specimens considered by Paige Prostate as suspicious, whose final diagnosis was benign, were found to comprise atrophy (n = 14), atrophy and apical prostate tissue (n = 1), apical/benign prostate tissue (n = 9), adenosis (n = 2), and post-atrophic hyperplasia (n = 1). Paige Prostate resulted in the identification of four additional patients whose diagnoses were upgraded from benign/suspicious to malignant. Additionally, this AI-based test provided an estimated 65.5% reduction of the diagnostic time for the material analyzed. Given its optimal sensitivity and NPV, Paige Prostate has the potential to be employed for the automated identification of patients whose histologic slides could forgo full histopathologic review. In addition to providing incremental improvements in diagnostic accuracy and efficiency, this AI-based system identified patients whose prostate cancers were not initially diagnosed by three experienced histopathologists. © 2021 The Authors. The Journal of Pathology published by John Wiley & Sons, Ltd. on behalf of The Pathological Society of Great Britain and Ireland.

Micro-CT imaging: Developing criteria for examining fetal skeletons in regulatory developmental toxicology studies - A workshop report.

During the past two decades the use and refinements of imaging modalities have markedly increased making it possible to image embryos and fetuses used in pivotal nonclinical studies submitted to regulatory agencies. Implementing these technologies into the Good Laboratory Practice environment requires rigorous testing, validation, and documentation to ensure the reproducibility of data. A workshop on current practices and regulatory requirements was held with the goal of defining minimal criteria for the proper implementation of these technologies and subsequent submission to regulatory agencies. Micro-computed tomography (micro-CT) is especially well suited for high-throughput evaluations, and is gaining popularity to evaluate fetal skeletons to assess the potential developmental toxicity of test agents. This workshop was convened to help scientists in the developmental toxicology field understand and apply micro-CT technology to nonclinical toxicology studies and facilitate the regulatory acceptance of imaging data. Presentations and workshop discussions covered: (1) principles of micro-CT fetal imaging; (2) concordance of findings with conventional skeletal evaluations; and (3) regulatory requirements for validating the system. Establishing these requirements for micro-CT examination can provide a path forward for laboratories considering implementing this technology and provide regulatory agencies with a basis to consider the acceptability of data generated via this technology.

Micro-computed tomography imaging and analysis in developmental biology and toxicology.

Micro-computed tomography (micro-CT) is a high resolution imaging technique that has expanded and strengthened in use since it was last reviewed in this journal in 2004. The technology has expanded to include more detailed analysis of bone, as well as soft tissues, by use of various contrast agents. It is increasingly applied to questions in developmental biology and developmental toxicology. Relatively high-throughput protocols now provide a powerful and efficient means to evaluate embryos and fetuses subjected to genetic manipulations or chemical exposures. This review provides an overview of the technology, including scanning, reconstruction, visualization, segmentation, and analysis of micro-CT generated images. This is followed by a review of more recent applications of the technology in some common laboratory species that highlight the diverse issues that can be addressed.

Digimouse: a 3D whole body mouse atlas from CT and cryosection data.

We have constructed a three-dimensional (3D) whole body mouse atlas from coregistered x-ray CT and cryosection data of a normal nude male mouse. High quality PET, x-ray CT and cryosection images were acquired post mortem from a single mouse placed in a stereotactic frame with fiducial markers visible in all three modalities. The image data were coregistered to a common coordinate system using the fiducials and resampled to an isotropic 0.1 mm voxel size. Using interactive editing tools we segmented and labelled whole brain, cerebrum, cerebellum, olfactory bulbs, striatum, medulla, masseter muscles, eyes, lachrymal glands, heart, lungs, liver, stomach, spleen, pancreas, adrenal glands, kidneys, testes, bladder, skeleton and skin surface. The final atlas consists of the 3D volume, in which the voxels are labelled to define the anatomical structures listed above, with coregistered PET, x-ray CT and cryosection images. To illustrate use of the atlas we include simulations of 3D bioluminescence and PET image reconstruction. Optical scatter and absorption values are assigned to each organ to simulate realistic photon transport within the animal for bioluminescence imaging. Similarly, 511 keV photon attenuation values are assigned to each structure in the atlas to simulate realistic photon attenuation in PET. The Digimouse atlas and data are available at http://neuroimage.usc.edu/Digimouse.html.

Cell confluency analysis on microcarriers by micro-flow imaging.

The productivity of cell culture-derived vaccines grown in anchorage-dependent animal cells is limited by bioreactor surface area. One way to increase the available surface area is by growing cells as monolayers on small spheres called microcarriers, which are approximately 100-250 µm in diameter. In order for microcarrier-based cell culture to be a success, it is important to understand the kinetics of cell growth on the microcarriers. Micro-flow imaging (MFI) is a simple and powerful technique that captures images and analyzes samples as they are drawn through a precision flow cell. In addition to providing size distribution and defect frequency data to compare microcarrier lots, MFI was used to generate hundreds of images to determine cell coverage and confluency on microcarriers. Same-day manual classification of these images provided upstream cell culture teams with actionable data that informed in-process decision making (e.g. time of infection). Additionally, an automated cell coverage algorithm was developed to increase the speed and throughput of the analyses.

Silencing Myostatin Using Cholesterol-conjugated siRNAs Induces Muscle Growth.

Short interfering RNAs (siRNAs) are a valuable tool for gene silencing with applications in both target validation and therapeutics. Many advances have recently been made to improve potency and specificity, and reduce toxicity and immunostimulation. However, siRNA delivery to a variety of tissues remains an obstacle for this technology. To date, siRNA delivery to muscle has only been achieved by local administration or by methods with limited potential use in the clinic. We report systemic delivery of a highly chemically modified cholesterol-conjugated siRNA targeting muscle-specific gene myostatin (Mstn) to a full range of muscles in mice. Following a single intravenous injection, we observe 85-95% knockdown of Mstn mRNA in skeletal muscle and >65% reduction in circulating Mstn protein sustained for >21 days. This level of Mstn knockdown is also accompanied by a functional effect on skeletal muscle, with animals showing an increase in muscle mass, size, and strength. The cholesterol-conjugated siRNA platform described here could have major implications for treatment of a variety of muscle disorders, including muscular atrophic diseases, muscular dystrophy, and type II diabetes.

Development and optimization of a high-throughput micro-computed tomography imaging method incorporating a novel analysis technique to evaluate bone mineral density of arthritic joints in a rodent model of collagen induced arthritis.

Rheumatoid arthritis (RA) is a chronic autoimmune disease resulting in joint inflammation, pain, and eventual bone loss. Bone loss and remodeling caused by symmetric polyarthritis, the hallmark of RA, is readily detectable by bone mineral density (BMD) measurement using micro-CT. Abnormalities in these measurements over time reflect the underlying pathophysiology of the bone. To evaluate the efficacy of anti-rheumatic agents in animal models of arthritis, we developed a high throughput knee and ankle joint imaging assay to measure BMD as a translational biomarker. A bone sample holder was custom designed for micro-CT scanning, which significantly increased assay throughput. Batch processing 3-dimensional image reconstruction, followed by automated image cropping, significantly reduced image processing time. In addition, we developed a novel, automated image analysis method to measure BMD and bone volume of knee and ankle joints. These improvements significantly increased the throughput of ex vivo bone sample analysis, reducing data turnaround from 5 days to 24 hours for a study with 200 rat hind limbs. Taken together, our data demonstrate that BMD, as quantified by micro-CT, is a robust efficacy biomarker with a high degree of sensitivity. Our innovative approach toward evaluation of BMD using optimized image acquisition and novel image processing techniques in preclinical models of RA enables high throughput assessment of anti-rheumatic agents offering a powerful tool for drug discovery.

Quantification of Cy-5 siRNA signal in the intra-vital multi-photon microscopy images.

Transgenic mice with Tie2- green fluorescent protein (GFP) are used as a model to study the kinetic distribution of the Cy5-siRNA delivered by lipid nanoparticles (LNP) into the liver. After the mouse is injected with the LNP, it undergoes a procedure of intra-vital multi-photon microscopy imaging over a period of two hours, during which the process for the nanoparticle to diffuse into the hepatocytes from the vasculature system is monitored. Since the images are obtained in-vivo, the quantification of Cy5 kinetics suffers from the moving field of view (FOV). A method is proposed to register the sequence of images through template matching. Based on the semi-automatic segmentations of the vessels in the common FOV, the registered images are segmented into three regions of interest (ROI) in which the Cy5 signals are quantified. Computation of the percentage signal strength in the ROIs over time allows for the analysis of the diffusion of Cy5-siRNA into the hepatocytes, and helps demonstrate the effectiveness of the Cy5-siRNA delivery vehicle.

Microscopic image analysis for quantitative characterization of muscle fiber type composition.

Skeletal muscles consist of muscle fibers that are responsible for contracting and generating force. Skeletal muscle fibers are categorized into distinct subtypes based on several characteristics such as contraction time, force production and resistance to fatigue. The composition of distinct muscle fibers in terms of their number and cross-sectional areas is characterized by a histological examination. However, manual delineation of individual muscle fibers from digitized muscle histology tissue sections is extremely time-consuming. In this study, we propose an automated image analysis system for quantitative characterization of muscle fiber type composition. The proposed system operates on digitized histological muscle tissue slides and consists of the following steps: segmentation of muscle fibers, registration of successive slides with distinct stains, and classification of muscle fibers into distinct subtypes. The performance of the proposed approach was tested on a dataset consisting of 25 image pairs of successive muscle histological cross-sections with different ATPase stain. Experimental results demonstrate a promising overall segmentation and classification accuracy of 89.1% in identifying muscle fibers of distinct subtypes.

Analysis of lipid nanoparticles by Cryo-EM for characterizing siRNA delivery vehicles.

Lipid nanoparticles are self-assembling, dynamic structures commonly used as carriers of siRNA, DNA, and small molecular therapeutics. Quantitative analysis of particle characteristics such as morphological features can be very informative as biophysical properties are known to influence biological activity, biodistribution, and toxicity. However, accurate characterization of particle attributes and population distributions is difficult. Cryo-Electron Microscopy (Cryo-EM) is a leading characterization method and can reveal diversity in particle size, shape and lamellarity, however, this approach is traditionally used for qualitative review or low throughput image analysis due to inherent EM micrograph contrast characteristics and artifacts in the images which limit extraction of quantitative feature values. In this paper we describe the development of a semiautomatic image analysis framework to facilitate reliable image enhancement, object segmentation, and quantification of nanoparticle attributes in Cryo-EM micrographs. We apply this approach to characterize two formulations of siRNA-loaded lipid nanoparticles composed of cationic lipid, cholesterol, and poly(ethylene glycol)-lipid, where the formulations differ only by input component ratios. We found Cryo-EM image analysis provided reliable size and morphology information as well as the detection of smaller particle populations that were not detected by standard dynamic light scattering (DLS) analysis.

Segmentation of skull and scalp in 3-D human MRI using mathematical morphology.

We present a new technique for segmentation of skull and scalp in T(1)-weighted magnetic resonance images (MRIs) of the human head. Our method uses mathematical morphological operations to generate realistic models of the skull, scalp, and brain that are suitable for electroencephalography (EEG) and magnetoencephalography (MEG) source modeling. We first segment the brain using our Brain Surface Extractor algorithm; using this, we can ensure that the brain does not intersect our skull segmentation. We next generate a scalp mask using a combination of thresholding and mathematical morphology. We use the scalp mask in our skull segmentation procedure, as it allows us to automatically exclude background voxels with intensities similar to those of the skull. We find the inner and outer skull boundaries using thresholding and morphological operations. Finally, we mask the results with the scalp and brain volumes to ensure closed and nonintersecting skull boundaries. Visual evaluation indicated accurate segmentations of the cranium at a gross anatomical level (other than small holes in the zygomatic bone in eight subjects) in all 44 MRI volumes processed when run using default settings. In a quantitative comparison with coregistered CT images as a gold standard, MRI skull segmentation accuracy, as measured using the Dice coefficient, was found to be similar to that which would be obtained using CT imagery with a registration error of 2-3 mm.