Manual classification of flow-limitation using flow shape and gold standard signals

AbstractLongitudinal and perpendicular changes in the blood vessels of the vocal fold have been related to the advancement from benign to malignant laryngeal cancer stages. The combination of Contact Endoscopy (CE) and Narrow Band Imaging (NBI) provides intraoperative realtime visualization of vascular pattern in Larynx. The evaluation of these vascular patterns in CE+NBI images is a subjective process leading to differentiation difficulty and subjectivity between benign and malignant lesions. The main objective of this work is to compare multi-observer classification versus automatic classification of laryngeal lesions. Six clinicians visually classified CE+NBI images into benign and malignant lesions. For the automatic classification of CE+NBI images, we used an algorithm based on characterizing the level of the vessel’s disorder. The results of the manual classification showed that there is no objective interpretation, leading to difficulties to visually distinguish between benign and malignant lesions. The results of the automatic classification of CE+NBI images on the other hand showed the capability of the algorithm to solve these issues. Based on the observed results we believe that, the automatic approach could be a valuable tool to assist clinicians to classifying laryngeal lesions.

Download Full-text

Validity of classification of distal radial fractures in the Swedish fracture register

BMC Musculoskeletal Disorders ◽

10.1186/s12891-021-04473-5 ◽

2021 ◽

Vol 22 (1) ◽

Author(s):

Malena Bergvall ◽

Carl Bergdahl ◽

Carl Ekholm ◽

David Wennergren

Keyword(s):

Gold Standard ◽

Reference Group ◽

Observer Agreement ◽

Distal Radial Fractures ◽

Young Males ◽

Small Peak ◽

National Quality ◽

Fracture Register ◽

Standard Classification

Abstract Background Distal radial fractures (DRF) are one of the most common fractures with a small peak in incidence among young males and an increasing incidence with age among women. The reliable classification of fractures is important, as classification provides a framework for communicating effectively on clinical cases. Fracture classification is also a prerequisite for data collection in national quality registers and for clinical research. Since its inception in 2011, the Swedish Fracture Register (SFR) has collected data on more than 490,000 fractures. The attending physician classifies the fracture according to the AO/OTA classification upon registration in the SFR. Previous studies regarding the classification of distal radial fractures (DRF) have shown difficulties in inter- and intra-observer agreement. This study aims to assess the accuracy of the registration of DRF in adults in the SFR as it is carried out in clinical practice. Methods A reference group of three experienced orthopaedic trauma surgeons classified 128 DRFs, randomly retrieved from the SFR, at two classification sessions 6 weeks apart. The classification the reference group agreed on was regarded as the gold standard classification for each fracture. The accuracy of the classification in the SFR was defined as the agreement between the gold standard classification and the classification in the SFR. Inter- and intra-observer agreement was evaluated and the degree of agreement was calculated as Cohen’s kappa. Results The accuracy of the classification of DRF in the SFR was kappa = 0.41 (0.31–0.51) for the AO/OTA subgroup/group and kappa = 0.48 (0.36–0.61) for the AO/OTA type. This corresponds to moderate agreement. Inter-observer agreement ranged from kappa 0.22–0.48 for the AO/OTA subgroup/group and kappa 0.48–0.76 for the AO/OTA type. Intra-observer agreement ranged from kappa 0.52–0.70 for the AO/OTA subgroup/group and kappa 0.71–0.76 for the AO/OTA type. Conclusions The study shows moderate accuracy in the classification of DRF in the SFR. Although the degree of accuracy for DRF appears to be lower than for other fracture locations, the accuracy shown in the current study is similar to that in previous studies of DRF.

Download Full-text

Researching the Research: Applying Machine Learning Techniques to Dissertation Classification

Journal of Computer Science Research ◽

10.30564/jcsr.v2i4.2230 ◽

2020 ◽

Vol 2 (4) ◽

Author(s):

Suzanna Schmeelk

Keyword(s):

Machine Learning ◽

Full Text ◽

Doctoral Program ◽

Machine Learning Algorithms ◽

Machine Learning Techniques ◽

Full Time ◽

Learning Techniques ◽

Manual Classification ◽

Machine Learning Tool

This research examines industry-based dissertation research in a doctoral computing program through the lens of machine learning algorithms to understand topics explored by senior and experienced full-time working professionals (EFWPs). Our research categorizes dissertation by both their abstracts and by their full-text using the Graplab Create library from Apple’s Turi. We also compare the dissertation categorizations using IBM’s Watson Discovery deep machine learning tool. Our research provides perspectives on the practicality of the manual classification of technical documents; and, it provides insights into the: (1) categories of academic work created by EFWPs in a Computing doctoral program, (2) viability of automated categorization versus human abstraction, and (3) differences in categorization algorithms.

Download Full-text

Towards a better understanding of the low recall of insertion variants with short-read based variant callers

10.1101/2020.06.09.142232 ◽

2020 ◽

Author(s):

Wesley Delage ◽

Julien Thevenon ◽

Claire Lemaitre

Keyword(s):

Gold Standard ◽

Insertion Site ◽

Structural Variants ◽

Genomic Context ◽

Breakpoint Junction ◽

Short Read ◽

Depth Analysis ◽

Long Read ◽

The Impact

AbstractSince 2009, numerous tools have been developed to detect structural variants (SVs) using short read technologies. Insertions >50 bp are one of the hardest type to discover and are drastically underrepresented in gold standard variant callsets. The advent of long read technologies has completely changed the situation. In 2019, two independent cross technologies studies have published the most complete variant callsets with sequence resolved insertions in human individuals. Among the reported insertions, only 17 to 37% could be discovered with short-read based tools. In this work, we performed an in-depth analysis of these unprecedented insertion callsets in order to investigate the causes of such failures. We have first established a precise classification of insertion variants according to four layers of characterization: the nature and size of the inserted sequence, the genomic context of the insertion site and the breakpoint junction complexity. Because these levels are intertwined, we then used simulations to characterize the impact of each complexity factor on the recall of several SV callers. Simulations showed that the most impacting factor was the insertion type rather than the genomic context, with various difficulties being handled differently among the tested SV callers, and they highlighted the lack of sequence resolution for most insertion calls. Our results explain the low recall by pointing out several difficulty factors among the observed insertion features and provide avenues for improving SV caller algorithms and their [email protected]

Download Full-text

PV Modeling of Medical Imaging Systems

Benford's Law ◽

10.23943/princeton/9780691147611.003.0018 ◽

2015 ◽

Author(s):

John Chiverton ◽

Kevin Wells

Keyword(s):

Medical Imaging ◽

Gold Standard ◽

Ground Truth ◽

Imaging Data ◽

Nuclear Medicine Imaging ◽

Phantom Imaging ◽

Positron Emission ◽

Pet Ct ◽

Many Sources

This chapter applies a Bayesian formulation of the Partial Volume (PV) effect, based on the Benford distribution, to the statistical classification of nuclear medicine imaging data: specifically Positron Emission Tomography (PET) acquired as part of a PET-CT phantom imaging procedure. The Benford distribution is a discrete probability distribution of great interest for medical imaging, because it describes the probabilities of occurrence of single digits in many sources of data. The chapter thus describes the PET-CT imaging and post-processing process to derive a gold standard. Moreover, this chapter uses it as a ground truth for the assessment of a Benford classifier formulation. The use of this gold standard shows that the classification of both the simulated and real phantom imaging data is well described by the Benford distribution.

Download Full-text

Cohomology of Homotopy 2-types

An Invitation to Computational Homotopy ◽

10.1093/oso/9780198832973.003.0005 ◽

2019 ◽

pp. 379-422

Author(s):

Graham Ellis

Keyword(s):

Perturbation Theory ◽

Projective Plane ◽

Integral Homology ◽

Homotopy Classes ◽

Cw Complex ◽

Crossed Modules ◽

Homological Perturbation Theory ◽

Homological Perturbation ◽

Manual Classification

This chapter introduces some of the basic ingredients in the classification of homotopy 2-types and describes datatypes and algorithms for implementing them on a computer. These are illustrated using computer examples involving: the fundamental crossed modules of a CW-complex, cat-1-groups, simplicial groups, Moore complexes, the Dold-Kan correspondence, integral homology of simplicial groups, homological perturbation theory. A manual classification of homotopy classes of maps from a surface to the projective plane is also included.

Download Full-text

Application of BERT to Enable Gene Classification Based on Clinical Evidence

BioMed Research International ◽

10.1155/2020/5491963 ◽

2020 ◽

Vol 2020 ◽

pp. 1-13

Author(s):

Yuhan Su ◽

Hongxin Xiang ◽

Haotian Xie ◽

Yong Yu ◽

Shiyan Dong ◽

...

Keyword(s):

Clinical Evidence ◽

Genetic Mutations ◽

Data Presentation ◽

Sequencing Technologies ◽

Practical Tool ◽

Logarithmic Loss ◽

Manual Classification ◽

High Repeatability ◽

Generation Sequencing

The identification of profiled cancer-related genes plays an essential role in cancer diagnosis and treatment. Based on literature research, the classification of genetic mutations continues to be done manually nowadays. Manual classification of genetic mutations is pathologist-dependent, subjective, and time-consuming. To improve the accuracy of clinical interpretation, scientists have proposed computational-based approaches for automatic analysis of mutations with the advent of next-generation sequencing technologies. Nevertheless, some challenges, such as multiple classifications, the complexity of texts, redundant descriptions, and inconsistent interpretation, have limited the development of algorithms. To overcome these difficulties, we have adapted a deep learning method named Bidirectional Encoder Representations from Transformers (BERT) to classify genetic mutations based on text evidence from an annotated database. During the training, three challenging features such as the extreme length of texts, biased data presentation, and high repeatability were addressed. Finally, the BERT+abstract demonstrates satisfactory results with 0.80 logarithmic loss, 0.6837 recall, and 0.705 F -measure. It is feasible for BERT to classify the genomic mutation text within literature-based datasets. Consequently, BERT is a practical tool for facilitating and significantly speeding up cancer research towards tumor progression, diagnosis, and the design of more precise and effective treatments.

Download Full-text

Impact of ICD10 and secular changes on electronic medical record rheumatoid arthritis algorithms

Rheumatology ◽

10.1093/rheumatology/keaa198 ◽

2020 ◽

Vol 59 (12) ◽

pp. 3759-3766 ◽

Cited By ~ 1

Author(s):

Sicong Huang ◽

Jie Huang ◽

Tianrun Cai ◽

Kumar P Dahal ◽

Andrew Cagan ◽

...

Keyword(s):

Language Processing ◽

Gold Standard ◽

International Classification Of Diseases ◽

Data Mart ◽

Classification Of Diseases ◽

Icd Codes ◽

Rule Based Approach ◽

Citrullinated Peptide ◽

Better Than

Abstract Objective The objective of this study was to compare the performance of an RA algorithm developed and trained in 2010 utilizing natural language processing and machine learning, using updated data containing ICD10, new RA treatments, and a new electronic medical records (EMR) system. Methods We extracted data from subjects with ≥1 RA International Classification of Diseases (ICD) codes from the EMR of two large academic centres to create a data mart. Gold standard RA cases were identified from reviewing a random 200 subjects from the data mart, and a random 100 subjects who only have RA ICD10 codes. We compared the performance of the following algorithms using the original 2010 data with updated data: (i) a published 2010 RA algorithm; (ii) updated algorithm, incorporating ICD10 RA codes and new DMARDs; and (iii) published algorithm using ICD codes only, ICD RA code ≥3. Results The gold standard RA cases had mean age 65.5 years, 78.7% female, 74.1% RF or antibodies to cyclic citrullinated peptide (anti-CCP) positive. The positive predictive value (PPV) for ≥3 RA ICD was 54%, compared with 56% in 2010. At a specificity of 95%, the PPV of the 2010 algorithm and the updated version were both 91%, compared with 94% (95% CI: 91, 96%) in 2010. In subjects with ICD10 data only, the PPV for the updated 2010 RA algorithm was 93%. Conclusion The 2010 RA algorithm validated with the updated data with similar performance characteristics as the 2010 data. While the 2010 algorithm continued to perform better than the rule-based approach, the PPV of the latter also remained stable over time.

Download Full-text

Role of Combined Pleural Fluid Cholesterol and Total Protein in Differentiation of Exudates and Transudates

Nepalese Medical Journal ◽

10.3126/nmj.v3i1.28330 ◽

2020 ◽

Vol 3 (1) ◽

pp. 282-285

Author(s):

Anupam Bista ◽

Suman Thapa ◽

Prasant Subedi ◽

Kiran Manandhar

Keyword(s):

Pleural Fluid ◽

Standard Method ◽

Total Protein ◽

Gold Standard ◽

Pleural Effusions

Introduction: Light's criteria had been the standard method for distinguishing exudative and transudative pleural effusions which misidentify 15-20% of transudates as exudates. This study aims to find out the role of combined pleural fluid cholesterol and total protein in distinguishing exudative from transudative pleural effusions and its applicability in Nepalese populations. Materials and Methods: Patients with pleural effusions were enrolled for the study. The combined pleural fluid cholesterol and total protein were compared with Light’s criteria and also compared with the diagnosis on discharge to find out their usefulness in categorizing the pleural effusions. Results: A total of 81 patients enrolled in the study, 42 (51.9%) were male. Based on Light’s criteria, 88.8% pleural effusions were found to be exudates and 11.1% were found to be transudates. Within the criteria, Light’s criteria categorized more pleural fluids as exudates than the diagnosis on discharge. Based on pleural fluid cholesterol >60mg/dL and protein >3g/dL for the classification of exudative and transudative pleural fluid, 62.9% out of 81 samples felled under the exudates and 37.03% pleural effusions under transudates with the sensitivity 87.9% and specificity 100%. Conclusions: Though Light’s criteria remain the gold standard to differentiate transudates and exudates, combined pleural fluid cholesterol and total protein give nearly comparable results without the need for simultaneous blood investigations.

Download Full-text

Questionnaire-based diagnosis of benign paroxysmal positional vertigo

Neurology ◽

10.1212/wnl.0000000000008876 ◽

2019 ◽

Vol 94 (9) ◽

pp. e942-e949 ◽

Cited By ~ 7

Author(s):

Hyo-Jung Kim ◽

Jeong-Mi Song ◽

Liqun Zhong ◽

Xu Yang ◽

Ji-Soo Kim

Keyword(s):

Confidence Interval ◽

Gold Standard ◽

Benign Paroxysmal Positional Vertigo ◽

Class Iii ◽

Positional Vertigo ◽

National University ◽

Seoul National University ◽

Sensitivity Specificity ◽

Paroxysmal Positional Vertigo

ObjectivesTo develop a simple questionnaire for self-diagnosis of benign paroxysmal positional vertigo (BPPV).MethodsWe developed a questionnaire that consisted of 6 questions, the first 3 to diagnose BPPV and the next 3 to determine the involved canal and type of BPPV. From 2016 to 2017, 578 patients with dizziness completed the questionnaire before the positional tests, a gold standard for diagnosis of BPPV, at the Dizziness Clinic of Seoul National University Bundang Hospital.ResultsOf the 578 patients, 200 were screened to have BPPV and 378 were screened to have dizziness/vertigo due to disorders other than BPPV. Of the 200 patients with a questionnaire-based diagnosis of BPPV, 160 (80%) were confirmed to have BPPV with positional tests. Of the 378 patients with a questionnaire-based diagnosis of non-BPPV, 24 (6.3%) were found to have BPPV with positional tests. Thus, the sensitivity, specificity, and precision of the questionnaires for the diagnosis of BPPV were 87.0%, 89.8%, and 80.0% (121 of 161, 95% confidence interval 74.5%–85.5%). Of the 200 patients with a questionnaire-based diagnosis of BPPV, 30 failed to respond to the questions 4 through 6 to determine the involved canal and type of BPPV. The questionnaire and positional tests showed the same results for the subtype and affected side of BPPV in 121 patients (121 of 170, 71.2%).ConclusionThe accuracy of questionnaire-based diagnosis of BPPV is acceptable.Classification of evidenceThis study provides Class III evidence that, in patients with dizziness, a questionnaire can diagnose BPPV with a sensitivity of 87.0% and a specificity of 89.8%.

Download Full-text