Rxivist combines preprints from bioRxiv with data from Twitter to help you find the papers being discussed in your field. Currently indexing 63,046 bioRxiv papers from 279,617 authors.
Most downloaded bioRxiv papers, all time
in category pathology
306 results found. For more information, click each entry to expand.
5,808 downloads pathology
A quantitative model to genetically interpret the histology in whole microscopy slide images is desirable to guide downstream immunohistochemistry, genomics, and precision medicine. We constructed a statistical model that predicts whether or not SPOP is mutated in prostate cancer, given only the digital whole slide after standard hematoxylin and eosin [H&E] staining. Using a TCGA cohort of 177 prostate cancer patients where 20 had mutant SPOP, we trained multiple ensembles of residual networks, accurately distinguishing SPOP mutant from SPOP non-mutant patients (test AUROC=0.74, p=0.0007 Fisher's Exact Test). We further validated our full metaensemble classifier on an independent test cohort from MSK-IMPACT of 152 patients where 19 had mutant SPOP. Mutants and non-mutants were accurately distinguished despite TCGA slides being frozen sections and MSK-IMPACT slides being formalin-fixed paraffin-embedded sections (AUROC=0.86, p=0.0038). Moreover, we scanned an additional 36 MSK-IMPACT patient having mutant SPOP, trained on this expanded MSK-IMPACT cohort (test AUROC=0.75, p=0.0002), tested on the TCGA cohort (AUROC=0.64, p=0.0306), and again accurately distinguished mutants from non-mutants using the same pipeline. Importantly, our method demonstrates tractable deep learning in this "small data" setting of 20-55 positive examples and quantifies each prediction's uncertainty with confidence intervals. To our knowledge, this is the first statistical model to predict a genetic mutation in cancer directly from the patient's digitized H&E-stained whole microscopy slide. Moreover, this is the first time quantitative features learned from patient genetics and histology have been used for content-based image retrieval, finding similar patients for a given patient where the histology appears to share the same genetic driver of disease i.e. SPOP mutation (p=0.0241 Kost's Method), and finding similar patients for a given patient that does not have have that driver mutation (p=0.0170 Kost's Method).
3,724 downloads pathology
This document represents a brief account of ongoing project for Diabetic Retinopathy Detection (DRD) through integration of state-of the art Deep Learning methods. We make use of deep Convolutional Neural Networks (CNNs), which have proven revolutionary in multiple fields of computer vision including medical imaging, and we bring their power to the diagnosis of eye fundus images. For training our models we used publicly available Kaggle data set. For testing we used portion of Kaggle data withheld from training and Messidor-2 reference standard. Neither withheld Kaggle images, nor Messidor-2 were used for training. For Messidor-2 we achieved sensitivity 99%, specificity 71%, and AUC 0.97. These results close to recent state-of-the-art models trained on much larger data sets and surpass average results of diabetic retinopathy screening when performed by trained optometrists. With continuous development of our Deep Learning models we expect to further increase the accuracy of the method and expand it to cataract and glaucoma diagnostics.
3,335 downloads pathology
Breast cancer is one of the main causes of cancer death worldwide. Early diagnostics significantly increases the chances of correct treatment and survival, but this process is tedious and often leads to a disagreement between pathologists. Computer-aided diagnosis systems showed potential for improving the diagnostic accuracy. In this work, we develop the computational approach based on deep convolution neural networks for breast cancer histology image classification. Hematoxylin and eosin stained breast histology microscopy image dataset is provided as a part of the ICIAR 2018 Grand Challenge on Breast Cancer Histology Images. Our approach utilizes several deep neural network architectures and gradient boosted trees classifier. For 4-class classification task, we report 87.2% accuracy. For 2-class classification task to detect carcinomas we report 93.8% accuracy, AUC 97.3%, and sensitivity/specificity 96.5/88.0% at the high-sensitivity operating point. To our knowledge, this approach outperforms other common methods in automated histopathological image classification. The source code for our approach is made publicly available at https://github.com/alexander-rakhlin/ICIAR2018.
3,124 downloads pathology
Sudden Infant Death Syndrome (SIDS) is the most common cause of postneonatal infant death. The allostatic load hypothesis posits that SIDS is the result of perinatal cumulative painful, stressful, or traumatic exposures that tax neonatal regulatory systems. To test it, we explored the relationships between SIDS and two common stressors, male neonatal circumcision (MNC) and prematurity, using latitudinal data from 15 countries and over 40 US states during the years 1999-2016. We used linear regression analyses and likelihood ratio tests to calculate the association between SIDS and the stressors. SIDS prevalence was significantly and positively correlated with MNC and prematurity rates. MNC explained 14.2% of the variability of SIDS's male bias in the US, reminiscent of the Jewish myth of Lilith, the killer of infant males. Combined, the stressors increased the likelihood of SIDS. Ecological analyses are useful to generate hypotheses but cannot provide strong evidence of causality. Biological plausibility is provided by a growing body of experimental and clinical evidence linking adversary preterm and early-life events with SIDS. Together with historical evidence, our findings emphasize the necessity of cohort studies that consider these environmental stressors with the aim of improving the identification of at-risk infants and reducing infant mortality.
2,866 downloads pathology
Over the years, several tumor biomarkers have been suggested to foresee the prognosis of oral squamous cell carcinoma (OSCC) patients. Here, we present a systematic review to identify, evaluate and summarize the evidence for OSCC reported markers. Eligible studies were identified through a literature search of MEDLINE/PubMed until January 2016. We included primary articles reporting overall survival, disease-free survival and cause-specific survival as outcomes. Our findings were analysed using REporting recommendations for tumor MARKer prognostic studies (REMARK), QuickGo tool and SciCurve trends. We found 41 biomarkers, mostly proteins evaluated by immunohistochemistry. The selected studies are of good quality, although, any study referred to a sample size determination. Considering the lack of follow-up studies, the molecules are still potential biomarkers. Further research is required to validate these biomarkers in well designed clinical cohort-based studies.
2,644 downloads pathology
Christopher R. Merritt, Giang T Ong, Sarah Church, Kristi Barker, Gary Geiss, Margaret Hoang, Jaemyeong Jung, Yan Liang, Jill McKay-Fleisch, Karen Nguyen, Kristina Sorg, Isaac Sprague, Charles Warren, Sarah Warren, Zoey Zhou, Daniel R. Zollinger, Dwayne L. Dunaway, Gordon B. Mills, Joseph M. Beechem
We have developed Digital Spatial Profiling (DSP), a non-destructive method for high-plex spatial profiling of proteins and RNA, using oligonucleotide detection technologies with unlimited multiplexing capability. The key breakthroughs underlying DSP are threefold: (1) multiplexed readout of proteins/RNA using oligo-tags; (2) oligo-tags attached to affinity reagents (antibodies/RNA probes) through a photocleavable (PC) linker; (3) photocleaving light projected onto the tissue sample to release PC-oligos in any spatial pattern. Here we show precise analyte reproducibility, validation, and cellular resolution using DSP. We also demonstrate biological proof-of-concept using lymphoid, colorectal tumor, and autoimmune tissue as models to profile immune cell populations, stroma, and cancer cells to identify factors specific for the diseased microenvironment. DSP utilizes the unlimited multiplexing capability of modern genomic approaches, while simultaneously providing spatial context of protein and RNA to examine biological questions based on analyte location and distribution.
2,264 downloads pathology
Multiplexing (mplx), labeling for multiple immunostains the very same cell or tissue section in situ, has raised considerable interest. The methods proposed include the use of labelled primary antibodies, spectral separation of fluorochromes, bleaching of the fluorophores or chromogens, blocking of previous antibody layers, all in various combinations. The major obstacles to the diffusion of this technique are high costs in custom antibodies and instruments, low throughput, scarcity of specialized skills or facilities. We have validated a method based on common primary and secondary antibodies and diffusely available fluorescent image scanners. It entails rounds of four-color indirect immunofluorescence, image acquisition and removal (stripping) of the antibodies, before another stain is applied. The images are digitally registered and the autofluorescence is subtracted. Removal of antibodies is accomplished by disulphide cleavage and a detergent or by a chaotropic salt treatment, this latter followed by antigen refolding. More than thirty different antibody stains can be applied to one single section from routinely fixed and embedded tissue. This method requires a modest investment in hardware and materials and uses freeware image analysis software. Mplx on routine tissue sections is a high throughput tool for in situ characterization of neoplastic, reactive, inflammatory and normal cells.
2,001 downloads pathology
Yifu Ding, Daniel J Vanselow, Maksim A Yakovlev, Spencer R Katz, Alex Y Lin, Darin P Clark, Phillip Vargas, Xuying Xin, Jean E Copper, Victor A Canfield, Khai C Ang, Yuxin Wang, Xianghui Xiao, Francesco De Carlo, Damian B. van Rossum, Patrick La Rivière, Keith C Cheng
Histological studies providing cellular insights into tissue architecture have been central to biological discovery and remain clinically invaluable today. Extending histology to three dimensions would be transformational for research and diagnostics. However, three-dimensional histology is impractical using current techniques. We have customized sample preparation, synchrotron X-ray tomographic parameters, and three-dimensional image analysis to allow for complete histological phenotyping using whole larval and juvenile zebrafish. The resulting digital zebrafish can be virtually sectioned and visualized in any plane. Whole-animal reconstructions at subcellular resolution also enable computational characterization of the zebrafish nervous system by region-specific detection of cell nuclei and quantitative assessment of individual phenotypic variation. Three-dimensional histological phenotyping has potential use in genetic and chemical screens, and in clinical and toxicological tissue diagnostics.
1,929 downloads pathology
Skeletal bone age assessment is a common clinical practice to diagnose endocrine and metabolic disorders in child development. In this paper, we describe a fully automated deep learning approach to the problem of bone age assessment using data from the 2017 Pediatric Bone Age Challenge organized by the Radiological Society of North America. The dataset for this competition is consisted of 12.6k radiological images. Each radiograph in this dataset is an image of a left hand labeled by the bone age and the sex of a patient. Our approach utilizes several deep neural network architectures trained end-to-end. We use images of whole hands as well as specific parts of a hand for both training and inference. This approach allows us to measure the importance of specific hand bones for the automated bone age analysis. We further evaluate performance of the method in the context of skeletal development stages. Our approach outperforms other common methods for bone age assessment.
1,920 downloads pathology
Adam Yala, Regina Barzilay, Laura Salama, Molly Griffin, Grace Sollender, Aditya Bardia, Constance Lehman, Julliette M Buckley, Suzanne B Coopey, Fernanda Polubriaginof, Judy E Garber, Barbara L Smith, Michele A Gadd, Michelle C Specht, Thomas M Gudewicz, Anthony Guidi, Alphonse Taghian, Kevin S Hughes
Purpose: Extracting information from Electronic Medical Record is a time-consuming and expensive process when done manually. Rule-based and machine learning techniques are two approaches to solving this problem. In this study, we trained a machine learning model on pathology reports to extract pertinent tumor characteristics, which enabled us to create a large database of attribute searchable pathology reports. This database can be used to identify cohorts of patients with characteristics of interest. Methods: We collected a total of 91,505 breast pathology reports from three Partners hospitals: Massachusetts General Hospital (MGH), Brigham and Womens Hospital (BWH), and Newton Wellesley Hospital (NWH), covering the period from 1978 to 2016. We trained our system with annotations from two datasets, consisting of 6,295 and 10,841 manually annotated reports. The system extracts 20 separate categories of information, including atypia types and various tumor characteristics such as receptors. We also report a learning curve analysis to show how much annotation our model needs to perform reasonably. Results: The model accuracy was tested on 500 reports that did not overlap with the training set. The model achieved accuracy of 90% for correctly parsing all carcinoma and atypia categories for a given patient. The average accuracy for individual categories was 97%. Using this classifier, we created a database of 91,505 parsed pathology reports. Conclusions: Our learning curve analysis shows that the model can achieve reasonable results even when trained on a few annotations. We developed a user-friendly interface to the database that allows physicians to easily identify patients with target characteristics and export the matching cohort. This model has the potential to reduce the effort required for analyzing large amounts of data from medical records, and to minimize the cost and time required to glean scientific insight from this data.
1,670 downloads pathology
The revolution in low-cost consumer photography and computation provides fertile opportunity for a disruptive reduction in the cost of biomedical imaging. Conventional approaches to low-cost microscopy are fundamentally restricted, however, to modest field of view (FOV) and/or resolution. We report a low-cost microscopy technique, implemented with a Raspberry Pi single-board computer and color camera combined with Fourier ptychography (FP), to computationally construct 25-megapixel images with sub-micron resolution. New image-construction techniques were developed to enable the use of the low-cost Bayer color sensor, to compensate for the highly aberrated re-used camera lens and to compensate for misalignments associated with the 3D-printed microscope structure. This high ratio of performance to cost is of particular interest to high-throughput microscopy applications, ranging from drug discovery and digital pathology to health screening in low-income countries. 3D models and assembly instructions of our microscope are made available for open source use.
1,448 downloads pathology
Rohan Munir, Kjetil Soreide, Rajan Ravindran, James J Powell, Ewen M. Harrison, Anya Adair, Stephen J. Wigmore, Rowan W. Parks, O James Garden, Lorraine Kirkpatrick, Lucy R Wall, Alan Christie, Ian Penman, Norma McAvoy, Vicki Save, Alan Stockman, David Worrall, Hamish Ireland, Graeme Weir, Neil Masson, Chris Hay, James-Gordon Smith, Damian James Mole
Introduction: The pattern of recurrence after surgical excision of pancreas cancer may guide alternative pre-operative strategies to either detect occult disease or need for chemotherapy. This study investigated patterns of recurrence after pancreatic surgery. Methods: Recurrence patterns were described in a series of resected pancreas cancers over a 2-year period and recurrence risk expressed as odds ratio (OR) with 95% confidence interval (C.I.). Survival was displayed by Kaplan-Meier curves. Results: Of 107 pancreas resections, 69 (65%) had pancreatic cancer. R0 resection was achieved in 21 of 69 (30.4%). Analysis was based on 66 patients who survived 30 days after surgery with median follow up 21 months. Recurrence developed in 41 (62.1%) patients with median time to first recurrence of 13.3 months (interquartile range 6.9, 20.8 months). Recurrence site was most frequently locoregional (n=28, 42%), followed by liver (n=23, 35%), lymph nodes (n=21, 32%), and lungs (n=13, 19%). In patients with recurrence, 9 of 41 had single site recurrence; the remaining 32 patients had more than one site of recurrence. Locoregional recurrence was associated with R+ resection (53% vs 25% for R+ vs R0, respectively; OR 3.5, 95% C.I. 1.1-11.2; P=0.034). Venous invasion was associated with overall recurrence risk (OR 3.3, 95% C.I. 1.1-9.4; P=0.025). In multivariable analysis, R-stage and adjuvant chemotherapy predicted longer survival. Discussion: The predominant locoregional recurrence pattern, multiple sites of recurrence and a high R+ resection rate reflect the difficulty in achieving initial local disease control.
1,426 downloads pathology
Non-small cell lung cancer is a leading cause of cancer death worldwide, and histopathological evaluation plays the primary role in its diagnosis. However, the morphological patterns associated with the molecular subtypes have not been systematically studied. To bridge this gap, we developed a quantitative histopathology analytic framework to identify the gene expression subtypes of non-small cell lung cancer objectively. We processed whole-slide histopathology images of lung adenocarcinoma (n=427) and lung squamous cell carcinoma patients (n=457) in The Cancer Genome Atlas. To establish neural networks for quantitative image analyses, we first build convolutional neural network models to identify tumor regions from adjacent dense benign tissues (areas under the receiver operating characteristic curves (AUC) > 0.935) and recapitulated expert pathologists' diagnosis (AUC > 0.88), with the results validated in an independent cohort (n=125; AUC > 0.85). We further demonstrated that quantitative histopathology morphology features identified the major transcriptomic subtypes of both adenocarcinoma and squamous cell carcinoma (P < 0.01). Our study is the first to classify the transcriptomic subtypes of non-small cell lung cancer using fully-automated machine learning methods. Our approach does not rely on prior pathology knowledge and can discover novel clinically-relevant histopathology patterns objectively. The developed procedure is generalizable to other tumor types or diseases.
1,392 downloads pathology
Dawn M Dudley, Matthew T Aliota, Emma Mohr, Andrea M Weiler, Gabrielle Lehrer-Brey, Kim L Weisgrau, Mariel S. Mohns, Meghan E Breitbach, Mustafa N Rasheed, Christina M. Newman, Dane D Gellerup, Louise H. Moncla, Jennifer Post, Nancy Schultz-Darken, Michele L. Schotkzo, Jennifer M Hayes, Josh A Eudailey, M Anthony Moody, Sallie R Permar, Shelby L. O’Connor, Eva G Rakasz, Heather A. Simmons, Saverio Capuano, Thaddeus G. Golos, Jorge E Osorio, Thomas C. Friedrich, David H. O’Connor
Infection with Asian lineage Zika virus has been associated with Guillain-Barre syndrome and fetal abnormalities 1-4, but the mechanisms and risk factors for these outcomes remain unknown. Here we show that rhesus macaques are susceptible to infection by an Asian-lineage virus closely related to strains currently circulating in the Americas. Following subcutaneous inoculation, Zika virus RNA was detected in plasma one day post-infection (dpi) in all animals (N = 8, including 2 animals infected during the first trimester of pregnancy). Plasma viral loads peaked above 1 x 105 viral RNA copies/mL in seven of eight animals. Viral RNA was also present in saliva, urine, and cerebrospinal fluid (CSF), consistent with case reports from infected humans. Viral RNA was cleared from plasma and urine by 21 dpi in non-pregnant animals. In contrast, both pregnant animals remained viremic longer, up to 57 days. In all animals, infection was associated with transient increases in proliferating natural killer cells, CD8+ T cells, CD4+ T cells, and plasmablasts. Neutralizing antibodies were detected in all animals by 21 dpi. Rechallenge of three non-pregnant animals with the Asian-lineage Zika virus 10 weeks after the initial challenge resulted in no detectable virus replication, suggesting that primary Zika virus infection elicits protective immunity against homologous virus strains. These data establish that Asian-lineage Zika virus infection of rhesus macaques provides a relevant animal model for studying pathogenesis in pregnant and non-pregnant individuals and evaluating potential interventions against human infection, including during pregnancy.
1,314 downloads pathology
Andrew J. Schaumberg, Wendy Juarez, Sarah J. Choudhury, Laura G. Pastrian, Bobbi S. Pritt, Mario Prieto Pozuelo, Ricardo Sotillo Sánchez, Khanh Ho, Nusrat Zahra, Betul Duygu Sener, Stephen Yip, Bin Xu, Srinivas Rao Annavarapu, Aurélien Morini, Karra A. Jones, Kathia Rosado-Orozco, S. Joseph Sirintrapun, Mariam Aly, Thomas J. Fuchs
Large-scale annotated image datasets like ImageNet and CIFAR-10 have been essential in developing and testing sophisticated new machine learning algorithms for natural vision tasks. Such datasets allow the development of neural networks to make visual discriminations that are done by humans in everyday activities, e.g. discriminating classes of vehicles. An emerging field -- computational pathology -- applies such machine learning algorithms to the highly specialized vision task of diagnosing cancer or other diseases from pathology images. Importantly, labeling pathology images requires pathologists who have had decades of training, but due to the demands on pathologists' time (e.g. clinical service) obtaining a large annotated dataset of pathology images for supervised learning is difficult. To facilitate advances in computational pathology, on a scale similar to advances obtained in natural vision tasks using ImageNet, we leverage the power of social media. Pathologists worldwide share annotated pathology images on Twitter, which together provide thousands of diverse pathology images spanning many sub-disciplines. From Twitter, we assembled a dataset of 2,746 images from 1,576 tweets from 13 pathologists from 8 countries; each message includes both images and text commentary. To demonstrate the utility of these data for computational pathology, we apply machine learning to our new dataset to test whether we can accurately identify different stains and discriminate between different tissues. Using a Random Forest, we report (i) 0.959 +- 0.013 Area Under Receiver Operating Characteristic [AUROC] when identifying single-panel human hematoxylin and eosin [H&E] stained slides that are not overdrawn and (ii) 0.996 +- 0.004 AUROC when distinguishing H&E from immunohistochemistry [IHC] stained microscopy images. Moreover, we distinguish all pairs of breast, dermatological, gastrointestinal, genitourinary, and gynecological [gyn] pathology tissue types, with mean AUROC for any pairwise comparison ranging from 0.771 to 0.879. This range is 0.815 to 0.879 if gyn is excluded. We report 0.815 +- 0.054 AUROC when all five tissue types are considered in a single multiclass classification task. Our goal is to make this large-scale annotated dataset publicly available for researchers worldwide to develop, test, and compare their machine learning methods, an important step to advancing the field of computational pathology.
1,220 downloads pathology
Idiopathic pulmonary fibrosis (IPF) is an intractable interstitial lung disease for which no curative treatment is available except for lung transplantation. Its pathogenesis is unclear, but a role for injury to type 2 alveolar epithelial cells is hypothesized. Recessive mutations in some, but not all genes implicated in Hermansky-Pudlak Syndrome (HPS) cause HPS-associated interstitial pneumonia (HPSIP), a clinical entity similar to IPF. We previously reported that mutation in HPS1 in embryonic stem cells-derived 3D lung organoids caused fibrotic changes. Here we show that introduction of all HPS mutations associated with HPSIP (HPS1, 2 and 4) promote fibrosis in lung organoids, while mutation in HSP8, which is not associated with HPSIP, does not. Furthermore, genome-expression analysis of epithelial cells derived from these organoids revealed significant overlap with similar analyses of both affected and unaffected lung tissue of non-HPS IPF patients. Importantly, this analysis showed upregulation of interleukin-11 in HPS-mutant fibrotic organoids and in fibrotic and unaffected lung tissue from IPF patients. Furthermore, IL-11 induced fibrosis in WT organoids, while its deletion prevented fibrosis in fibrotic HPS4-mutant organoids, suggesting IL-11 as a therapeutic target in IPF and HPSIP. hPSC-derived 3D lung organoids are therefore a valuable resource to model fibrotic lung disease.
1,196 downloads pathology
Neuropathologists assess vast brain areas to identify diverse and subtly-differentiated morphologies. Standard semi-quantitative scoring approaches, however, are coarse-grained and lack precise neuroanatomic localization. We report a proof-of-concept deep learning pipeline identifying specific neuropathologies--amyloid plaques and cerebral amyloid angiopathy--in immunohistochemically-stained archival slides. Using automated segmentation of stained objects and a cloud-based interface, we annotated >70,000 plaque candidates from 43 whole slide images (WSIs) to train and evaluate convolutional neural networks. Networks achieved strong plaque classification on a 10-WSI hold-out set (0.993 and 0.743 areas under the receiver operating characteristic and precision recall curve, respectively). Prediction confidence maps visualized morphology distributions for WSIs at high resolution. Resulting plaque-burden scores correlated well with established semi-quantitative scores on a 30-WSI blinded hold-out. Finally, saliency mapping demonstrated that networks learned patterns agreeing with accepted pathologic features. This scalable means to augment a neuropathologist's ability may suggest a route to neuropathologic deep phenotyping.
1,106 downloads pathology
John M. Starbuck, Sergi Llambrich, Ruben González, Julia Albaigès, Anna Sarlé, Jens Wouters, Alejandro González, Xavier Sevillano, James Sharpe, Rafael de la Torre, Mara Dierssen, Greetje Vande Velde, Neus Martinez-Abadias
In Down syndrome (DS), the overall genetic imbalance caused by trisomy of chromosome 21 leads to a complex pleiotropic phenotype that involves a recognizable set of facial traits. Several studies have shown the potential of epigallocatechin-3-gallate (EGCG), a green tea flavanol, as a therapeutic tool for alleviating different developmental alterations associated with DS, such as cognitive impairment, skull dysmorphologies, and skeletal deficiencies. Here we provide for the first time experimental and clinical evidence of the potential benefits of EGCG treatment to facial morphology. Our results showed that mouse models treated with low dose of EGCG during pre- and postnatal development improved facial dysmorphology. However, the same treatment at high dose produced disparate facial morphology changes with an extremely wide and abnormal range of variation. Our observational study in humans revealed that EGCG treatment since early in development is associated with intermediate facial phenotypes and significant facial improvement scores. Overall, our findings suggest a potential beneficial effect of ECGC on facial development, which requires further research to pinpoint the optimal dosages of EGCG that reliably improve DS phenotypes. Current evidence warns against the non-prescribed intake of this supplement as a health-promoting measure.
1,053 downloads pathology
Modern digital pathology departments have grown to produce whole-slide image data at petabyte scale, an unprecedented treasure chest for medical machine learning tasks. Unfortunately, most digital slides are not annotated at the image level, hindering large-scale application of supervised learning. Manual labeling is prohibitive, requiring pathologists with decades of training and outstanding clinical service responsibilities. This problem is further aggravated by the United States Food and Drug Administration's ruling that primary diagnosis must come from a glass slide rather than a digital image. We present the first end-to-end framework to overcome this problem, gathering annotations in a nonintrusive manner during a pathologist's routine clinical work: (i) microscope-specific 3D-printed commodity camera mounts are used to video record the glass-slide-based clinical diagnosis process; (ii) after routine scanning of the whole slide, the video frames are registered to the digital slide; (iii) motion and observation time are estimated to generate a spatial and temporal saliency map of the whole slide. Demonstrating the utility of these annotations, we train a convolutional neural network that detects diagnosis-relevant salient regions, then report accuracy of 85.15% in bladder and 91.40% in prostate, with 75.00% accuracy when training on prostate but predicting in bladder, despite different pathologists examining the different tissues. When training on one patient but testing on another, AUROC in bladder is 0.7929+-0.1109 and in prostate is 0.9568+-0.0374. Our tool is available at https://bitbucket.org/aschaumberg/deepscope
1,045 downloads pathology
Background: As a large number of digital histopathological images have been accumulated, there is a growing demand of content-based image retrieval (CBIR) in pathology for educational, diagnostic, or research purposes. However, no CBIR systems in digital pathology are publicly available. Results: We developed a web application, the Luigi system, which retrieves similar histopathological images from various cancer cases. Using deep texture representations computed with a pre-trained convolutional neural network as an image feature in conjunction with an approximate nearest neighbor search method, the Luigi system provides fast and accurate results for any type of tissue or cell without the need for further training. In addition, users can easily submit query images of an appropriate scale into the Luigi system and view the retrieved results using our smartphone application. The cases stored in the Luigi database are obtained from The Cancer Genome Atlas with rich clinical, pathological, and molecular information. We tested the Luigi system and the smartphone application by querying typical cancerous regions from four cancer types, and confirmed successful retrieval of relevant images with both applications. Conclusions: The Luigi system will help students, pathologists, and researchers easily retrieve histopathological images of various cancers similar to those of the query image. Luigi is freely available at https://luigi-pathology.com/.
- Top preprints of 2018
- Paper search
- Author leaderboards
- Overall metrics
- The API
- Email newsletter
- 21 May 2019: PLOS Biology has published a community page about Rxivist.org and its design.
- 10 May 2019: The paper analyzing the Rxivist dataset has been published at eLife.
- 1 Mar 2019: We now have summary statistics about bioRxiv downloads and submissions.
- 8 Feb 2019: Data from Altmetric is now available on the Rxivist details page for every preprint. Look for the "donut" under the download metrics.
- 30 Jan 2019: preLights has featured the Rxivist preprint and written about our findings.
- 22 Jan 2019: Nature just published an article about Rxivist and our data.
- 13 Jan 2019: The Rxivist preprint is live!