Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

Pathology Pathology

Automatic Segmentation, Localization, and Identification of Vertebrae in 3D CT Images Using Cascaded Convolutional Neural Networks

ArXiv Preprint

This paper presents a method for automatic segmentation, localization, and identification of vertebrae in arbitrary 3D CT images. Many previous works do not perform the three tasks simultaneously even though requiring a priori knowledge of which part of the anatomy is visible in the 3D CT images. Our method tackles all these tasks in a single multi-stage framework without any assumptions. In the first stage, we train a 3D Fully Convolutional Networks to find the bounding boxes of the cervical, thoracic, and lumbar vertebrae. In the second stage, we train an iterative 3D Fully Convolutional Networks to segment individual vertebrae in the bounding box. The input to the second networks have an auxiliary channel in addition to the 3D CT images. Given the segmented vertebra regions in the auxiliary channel, the networks output the next vertebra. The proposed method is evaluated in terms of segmentation, localization, and identification accuracy with two public datasets of 15 3D CT images from the MICCAI CSI 2014 workshop challenge and 302 3D CT images with various pathologies introduced in [1]. Our method achieved a mean Dice score of 96%, a mean localization error of 8.3 mm, and a mean identification rate of 84%. In summary, our method achieved better performance than all existing works in all the three metrics.

Naoto Masuzawa, Yoshiro Kitamura, Keigo Nakamura, Satoshi Iizuka, Edgar Simo-Serra

2020-09-29

General General

Dental Characteristics of Different Types of Cleft and Non-cleft Individuals.

In Frontiers in cell and developmental biology

Objective : The objective of this study was to compare the novel artificial intelligence (A.I.)-driven lateral cephalometric (Late. Ceph.) analysis of 14 different dental characteristics (DC) among different types of cleft lip and palate (CLP) and non-cleft (NC) individuals.

Materials and Methods : A retrospective study was conducted on 123 individuals [31 = NC, 29 = BCLP (bilateral cleft lip and palate), 41 = UCLP (unilateral cleft lip and palate), 9 = UCLA (unilateral cleft lip and alveolus), and 13 = UCL (unilateral cleft lip)] with an average age of 14.77 years. Demographic details were gathered from the clinical records. A novel artificial intelligence-driven Webceph software has been used for the Late. Ceph. analysis. A total of 14 different types of angular and linear DC measurements were analyzed and compared among groups. Two-way ANOVA and multiple-comparison statistics tests were applied to see the differences between gender and among different types of CLP versus NC subjects.

Results : Of the 14 DC tested, no significant gender disparities were found (p > 0.05). In relation to different types of CLP versus NC subjects, 8 over 14 DC were statistically significant (p < 001 to p = 0.03). Six other DC variables show insignificant (p > 0.05) noteworthy alterations in relation to type of CLP.

Conclusion : Based on the results, type of CLP revealed significantly altered DC compared to NC. Among different types of CLP, BCLP exhibited a maximum alteration in different DC.

Alam Mohammad Khursheed, Alfawzan Ahmed Ali

2020

bilateral cleft lip and palate, dental characteristics, incisal display, non-syndromic cleft lip and palate, overbite, overjet, unilateral cleft lip and palate

General General

Named Entity Recognition and Relation Detection for Biomedical Information Extraction.

In Frontiers in cell and developmental biology

The number of scientific publications in the literature is steadily growing, containing our knowledge in the biomedical, health, and clinical sciences. Since there is currently no automatic archiving of the obtained results, much of this information remains buried in textual details not readily available for further usage or analysis. For this reason, natural language processing (NLP) and text mining methods are used for information extraction from such publications. In this paper, we review practices for Named Entity Recognition (NER) and Relation Detection (RD), allowing, e.g., to identify interactions between proteins and drugs or genes and diseases. This information can be integrated into networks to summarize large-scale details on a particular biomedical or clinical problem, which is then amenable for easy data management and further analysis. Furthermore, we survey novel deep learning methods that have recently been introduced for such tasks.

Perera Nadeesha, Dehmer Matthias, Emmert-Streib Frank

2020

artificial intelligence, deep learning, information extraction, named entity recognition, natural language processing, relation detection, text analytics, text mining

General General

Toward a Closed Loop, Integrated Biocompatible Biopolymer Wound Dressing Patch for Detection and Prevention of Chronic Wound Infections.

In Frontiers in bioengineering and biotechnology

Chronic wound infections represent a significant burden to healthcare providers globally. Often, chronic wound healing is impeded by the presence of infection within the wound or wound bed. This can result in an increased healing time, healthcare cost and poor patient outcomes. Thus, there is a need for dressings that help the wound heal, in combination with early detection of wound infections to support prompt treatment. In this study, we demonstrate a novel, biocompatible wound dressing material, based on Polyhydroxyalkanoates, doped with graphene platelets, which can be used as an electrochemical sensing substrate for the detection of a common wound pathogen, Pseudomonas aeruginosa. Through the detection of the redox active secondary metabolite, pyocyanin, we demonstrate that a dressing can be produced that will detect the presence of pyocyanin across clinically relevant concentrations. Furthermore, we show that this sensor can be used to identify the presence of pyocyanin in a culture of P. aeruginosa. Overall, the sensor substrate presented in this paper represents the first step toward a new dressing with the capacity to promote wound healing, detect the presence of infection and release antimicrobial drugs, on demand, to optimized healing.

Ward Andrew C, Dubey Prachi, Basnett Pooja, Lika Granit, Newman Gwenyth, Corrigan Damion K, Russell Christopher, Kim Jongrae, Chakrabarty Samit, Connolly Patricia, Roy Ipsita

2020

Polyhydroxyalkanoates, Pseudomonas aeruginosa, artificial intelligence, biopolymer, electrochemical, graphene, pyocyanin, wound dressing

General General

Automated Detection of Acute Lymphoblastic Leukemia From Microscopic Images Based on Human Visual Perception.

In Frontiers in bioengineering and biotechnology

Microscopic image analysis plays a significant role in initial leukemia screening and its efficient diagnostics. Since the present conventional methodologies partly rely on manual examination, which is time consuming and depends greatly on the experience of domain experts, automated leukemia detection opens up new possibilities to minimize human intervention and provide more accurate clinical information. This paper proposes a novel approach based on conventional digital image processing techniques and machine learning algorithms to automatically identify acute lymphoblastic leukemia from peripheral blood smear images. To overcome the greatest challenges in the segmentation phase, we implemented extensive pre-processing and introduced a three-phase filtration algorithm to achieve the best segmentation results. Moreover, sixteen robust features were extracted from the images in the way that hematological experts do, which significantly increased the capability of the classifiers to recognize leukemic cells in microscopic images. To perform the classification, we applied two traditional machine learning classifiers, the artificial neural network and the support vector machine. Both methods reached a specificity of 95.31%, and the sensitivity of the support vector machine and artificial neural network reached 98.25 and 100%, respectively.

Bodzas Alexandra, Kodytek Pavel, Zidek Jan

2020

acute leukemia, automated leukemia detection, blood smear image analysis, cell segmentation, image processing, leukemic cell identification, machine learning

General General

Exploring the Contribution of Proprioceptive Reflexes to Balance Control in Perturbed Standing.

In Frontiers in bioengineering and biotechnology

Humans control balance using different feedback loops involving the vestibular system, the visual system, and proprioception. In this article, we focus on proprioception and explore the contribution of reflexes based on force and length feedback to standing balance. In particular, we address the questions of how much proprioception alone could explain balance control, and whether one modality, force or length feedback, is more important than the other. A sagittal plane neuro-musculoskeletal model was developed with six degrees of freedom and nine muscles in each leg. A controller was designed using proprioceptive reflexes and a dead zone. No feedback control was applied inside the dead zone. Reflexes were active once the center of mass moved outside the dead zone. Controller parameters were found by solving an optimization problem, where effort was minimized while the neuro-musculoskeletal model should remain standing upright on a perturbed platform. The ground was perturbed with random square pulses in the sagittal plane with different amplitudes and durations. The optimization was solved for three controllers: using force and length feedback (base model), using only force feedback, and using only length feedback. Simulations were compared to human data from previous work, where an experiment with the same perturbation signal was performed. The optimized controller yielded a similar posture, since average joint angles were within 5 degrees of the experimental average joint angles. The joint angles of the base model, the length only model, and the force only model correlated weakly (ankle) to moderately with the experimental joint angles. The ankle moment correlated weakly to moderately with the experimental ankle moment, while the hip and knee moment were only weakly correlated, or not at all. The time series of the joint angles showed that the length feedback model was better able to explain the experimental joint angles than the force feedback model. Changes in time delay affected the correlation of the joint angles and joint moments. The objective of effort minimization yielded lower joint moments than in the experiment, suggesting that other objectives are also important in balance control, which cause an increase in effort and thus larger joint moments.

Koelewijn Anne D, Ijspeert Auke J

2020

balance control, neuromusculoskeletal simulation, perturbed standing, proprioception, reflexes