Diagnosis Expert Research Articles

38 Plasma proteomic signature of motoric cognitive risk syndrome

OBJECTIVES/GOALS: Motoric cognitive risk (MCR) is a pre-dementia syndrome characterized by slow gait and subjective cognitive complaints. In the Atherosclerosis Risk in Communities (ARIC) study, we aim to (1) identify plasma proteins and protein modules associated with MCR and (2) compare the proteomic signature of MCR to that of mild cognitive impairment (MCI). METHODS/STUDY POPULATION: Nondemented ARIC participants were classified by MCR status (yes/no) according to a memory questionnaire and 4-meter walk. MCI status (yes/no) was classified by expert diagnosis using standardized criteria. We measured 4,877 proteins in plasma collected at ARIC Visit 5 (late-life) and Visit 2 (midlife) utilizing the SomaScan4 proteomic assay. Multivariable logistic regression”adjusted for demographic variables, kidney function, cardiovascular risk factors, and APOE4 status”related each protein to MCR at late-life. An FDR corrected P RESULTS/ANTICIPATED RESULTS: Proteome-wide association study among 4076 ARIC participants (mean age=75; 58% women, 17% Black, 4% MCR+, 21% MCI+; MCR+ and MCI+ groups overlapped) at late-life identified 26 MCR-associated proteins involved in metabolism, vascular/visceral smooth muscle, and extracellular matrix organization. At an uncorrected P DISCUSSION/SIGNIFICANCE: This proteomic characterization of MCR identifies novel plasma proteins and networks, both distinct from and overlapping with those of MCI, thus highlighting the partially divergent mechanisms underlying these pre-dementia syndromes. These findings may be leveraged toward dementia prognostication and targeted therapeutic approaches.

Evaluation of New Instruments for Screening and Diagnosis of Tics and Tic Disorders in a Well-Characterized Sample of Youth with Tics and Recruited Controls

ABSTRACT Tics and tic disorders can significantly impact children, but limited screening tools and diagnostic challenges may delay access to care. The current study attempted to address these gaps by evaluating sensitivity and specificity of the Motor or Vocal Inventory of Tics (MOVeIT), a tic symptom screener, and the Description of Tic Symptoms (DoTS), a brief diagnostic assessment for tic disorders. Children (n = 100, age 6–17 years old) with tic disorders attending a Tourette specialty clinic and a community-recruited sample without tics completed a gold-standard assessment by a tic expert; these evaluations were compared to child self-report and parent and teacher report versions of the MOVeIT, and child and parent versions of the DoTS. The parent and child MOVeIT met or exceeded pre-specified 85% sensitivity and specificity criteria for detecting the presence of tics when compared to a gold-standard tic expert diagnosis. The Teacher MOVeIT had lower sensitivity (71.4%) but good specificity (95.7%) for identifying any tic symptoms compared to gold standard. For determination of the presence or absence of any tic disorder, sensitivity of both parent and child DoTS was 100%; specificity of the parent DoTS was 92.7% and child DoTS specificity was 75.9%. More work may be needed to refine the teacher MOVeIT, but it is also recognized that tic expression may vary by setting. While the MOVeIT and DoTS parent and child questionnaires demonstrated adequate sensitivity and specificity for determining the presence of tics and tic disorders in this well-defined sample, additional testing in a general population is warranted.

Artificial intelligence assisted diagnoses of fine-needle aspiration of breast diseases: a single-center experience

Abstract: Purpose: Since 2010, physicians from Afghanistan have been uploading images of histological and cytological specimens to a telemedicine internet platform (iPath network) for expert evaluation. From this collective work, all cases with fine-needle aspirations (FNA) of mammary gland diseases were extracted and analyzed. The aim of the present retrospective feasibility study is to investigate the utility of artificial intelligence assisted diagnoses in fine-needle aspiration (FNA) of breast diseases.Material and Methods: A total of 3304 microphotographic images from 438 patients of smears from FNA of the mammary gland were available for this study. Telemedical expert diagnoses from 4 experienced cytopathologists were available in all 438 cases. Their diagnosis (malignant tumor of the mammary gland or benign mammary gland disease) was set as the gold standard. AI analysis was performed using i) clinical context data and ii) two different image recognition methods to determine the probability values for the presence of malignant breast tumor. Youden index and AUC (area under the curve) were used to evaluate test performance. Results: A score for invasive breast cancer (IBC) calculated from contextual variables agreed with the expert diagnosis (accuracy) in 85.2% and with the two image recognition systems in 78.4% and 65.2%. This simplifies health healthcare management of breast diseases in low income countries as in many patients the less expensive and less time-consuming technique of FNA may replace a histological examination.Conclusion: Image classification and analysis of context variables can be used to test the validity and plausibility of cytologic diagnoses, especially when cytologic interpretation has to be performed by people who are inexperienced in cytopathology.

Ultrasonography performed by an infectiologist in hip and knee prosthetic joint and native joint infections

IntroductionClinical ultrasonography (US) by infectiologists has only recently been developing, and as now there is little literature on the subject. Our study focuses on the conditions and diagnostic performance of clinical ultrasound imaging by infectiologists in cases of hip and knee prosthetic and native joint infection. MethodsA retrospective study carried out between June 1st 2019 and March 31st 2021 in the University Hospital of Bordeaux, South-Western France. We measured US sensitivity (Se), specificity (Sp), positive predictive value (PPV) and negative predictive value (NPV), combined or not with the analysis of articular fluid, compared to the MusculoSketetal Infection Society (MSIS) score in prosthetic joints, or to expert diagnosis in native joints. ResultsFifty-four patients underwent US by an infectiologist in an infectious disease ward, including 11 (20.4%) for native joint and 43 (79.6%) for prosthetic joint. Joint effusion and/or periarticular collection were highlighted in 47 (87%) patients, and US led to 44 punctures. In all patients (n = 54), Se, Sp, PPV and NPV of US alone were 91%, 19%, 64% and 57%, respectively. When US was combined with fluid analysis, Se, Sp, PPV, NPV were 68%, 100%, 100%, 64% in all patients (n = 54), 86%, 100%, 100%, 60% in acute arthritis (n = 17) and 50%, 100%, 100% and 65% respectively in non-acute arthritis (n = 37). ConclusionThese results suggest that US by infectiologists effectively diagnoses osteoarticular infections (OAIs). This approach has many applications in infectiology routines. Consequently, it would be interesting to define the contents of a first level of infectiologist competence in US clinical practice.

The development of a novel natural language processing tool to identify pediatric chest radiograph reports with pneumonia.

Chest radiographs are frequently used to diagnose community-acquired pneumonia (CAP) for children in the acute care setting. Natural language processing (NLP)-based tools may be incorporated into the electronic health record and combined with other clinical data to develop meaningful clinical decision support tools for this common pediatric infection. We sought to develop and internally validate NLP algorithms to identify pediatric chest radiograph (CXR) reports with pneumonia. We performed a retrospective study of encounters for patients from six pediatric hospitals over a 3-year period. We utilized six NLP techniques: word embedding, support vector machines, extreme gradient boosting (XGBoost), light gradient boosting machines Naïve Bayes and logistic regression. We evaluated their performance of each model from a validation sample of 1,350 chest radiographs developed as a stratified random sample of 35% admitted and 65% discharged patients when both using expert consensus and diagnosis codes. Of 172,662 encounters in the derivation sample, 15.6% had a discharge diagnosis of pneumonia in a primary or secondary position. The median patient age in the derivation sample was 3.7 years (interquartile range, 1.4-9.5 years). In the validation sample, 185/1350 (13.8%) and 205/1350 (15.3%) were classified as pneumonia by content experts and by diagnosis codes, respectively. Compared to content experts, Naïve Bayes had the highest sensitivity (93.5%) and XGBoost had the highest F1 score (72.4). Compared to a diagnosis code of pneumonia, the highest sensitivity was again with the Naïve Bayes (80.1%), and the highest F1 score was with the support vector machine (53.0%). NLP algorithms can accurately identify pediatric pneumonia from radiography reports. Following external validation and implementation into the electronic health record, these algorithms can facilitate clinical decision support and inform large database research.

Fault logic circuit design of switching power supply based on soft reference circuit module

The basic idea of the circuit fault diagnosis method based on fault logic is to establish a computer expert database system first, connect the computer to the GPIB interface card, and then connect each intelligent instrument through the GPIB standard bus. Through the connection of the tested object adapter and the tested object, a circuit fault expert diagnosis and test platform are formed. Based on the traditional fault simulation model, this paper improves the fault logic circuit of switching power supply, designs the system model of fault simulation, deals with the logic structure and circuit hierarchy respectively, and produces the final output through the logic gate circuit by comparing the results of the fault logic circuit and the fault-free logic circuit. The designed switching power supply fault logic circuit works in the switching state, which has the characteristics of low loss and high energy conversion efficiency. Through the simulation analysis, the output is connected to the comparator, which plays the role of the whole wave. The simulation results are in line with the parameter indicators. The functional test results of the chip show that the function of the module is basically normal, which has practical application value and good portability and can be used as a reference for similar electrical designs.

Neuromorphic processor-oriented hybrid Q-format multiplication with adaptive quantization for tiny YOLO3

Deep neural networks (DNNs) have delivered unprecedented achievements in the modern Internet of Everything society, encompassing autonomous driving, expert diagnosis, unmanned supermarkets, etc. It continues to be challenging for researchers and engineers to develop a high-performance neuromorphic processor for deployment in edge devices or embedded hardware. DNNs’ superpower derives from their enormous and complex network architecture, which is computation-intensive, time-consuming, and energy-heavy. Due to the limited perceptual capacity of humans, accurate processing results from DNNs require a substantial amount of computing time, making them redundant in some applications. Utilizing adaptive quantization technology to compress the DNN model with sufficient accuracy is crucial for facilitating the deployment of neuromorphic processors in emerging edge applications. This study proposes a method to boost the development of neuromorphic processors by conducting fixed-point multiplication in a hybrid Q-format using an adaptive quantization technique on the convolution of tiny YOLO3. In particular, this work integrates the sign-bit check and bit roundoff techniques into the arithmetic of fixed-point multiplications to address overflow and roundoff issues within the convolution’s adding and multiplying operations. In addition, a hybrid Q-format multiplication module is developed to assess the proposed method from a hardware perspective. The experimental results prove that the hybrid multiplication with adaptive quantization on the tiny YOLO3’s weights and feature maps possesses a lower error rate than alternative fixed-point representation formats while sustaining the same object detection accuracy. Moreover, the fixed-point numbers represented by Q(6.9) have a suboptimal error rate, which can be utilized as an alternative representation form for the tiny YOLO3 algorithm-based neuromorphic processor design. In addition, the 8-bit hybrid Q-format multiplication module exhibits low power consumption and low latency in contrast to benchmark multipliers.

Diagnostic Performance of a Machine Learning Algorithm (Asthma/Chronic Obstructive Pulmonary Disease [COPD] Differentiation Classification) Tool Versus Primary Care Physicians and Pulmonologists in Asthma, COPD, and Asthma/COPD Overlap

The differential diagnosis of asthma and chronic obstructive pulmonary disease (COPD) poses a challenge in clinical practice and its misdiagnosis results in inappropriate treatment, increased exacerbations, and potentially death. To investigate the diagnostic accuracy of the Asthma/COPD Differentiation Classification (AC/DC) tool compared with primary care physicians and pulmonologists in asthma, COPD, and asthma-COPD overlap. The AC/DC machine learning-based diagnostic tool was developed using 12 parameters from electronic health records of more than 400,000 patients aged 35 years and older. An expert panel of three pulmonologists and four general practitioners from five countries evaluated 119 patient cases from a prospective observational study and provided a confirmed diagnosis (n= 116) of asthma (n= 53), COPD (n= 43), asthma-COPD overlap (n= 7), or other (n= 13). Cases were then reviewed by 180 primary care physicians and 180 pulmonologists from nine countries and by the AC/DC tool, and diagnostic accuracies were compared with reference to the expert panel diagnoses. Average diagnostic accuracy of the AC/DC tool was superior to that of primary care physicians (median difference, 24%; 95% posterior credible interval: 17% to 29%; P < .0001) and was noninferior and superior (median difference, 12%; 95% posterior credible interval: 6% to 17%; P < .0001 for noninferiority and P= .0006 for superiority) to that of pulmonologists. Average diagnostic accuracies were 73%, 50%, and 61% by AC/DC tool, primary care physicians, and pulmonologists versus expert panel diagnosis, respectively. The AC/DC tool demonstrated superior diagnostic accuracy compared with primary care physicians and pulmonologists in the diagnosis of asthma and COPD in patients aged 35 years and greater and has the potential to support physicians in the diagnosis of these conditions in clinical practice.

Multigranulation-based double-quantitative rough sets for multi-source event-based decision systems

For the multi-source event-based decision system (MSEDS), each source exhibits its own one-sidedness and limitations. How to fuse the information contained in multiple sources to make a comprehensive decision has become an important research field. In this paper, the related concepts of support characteristic levels in one source are introduced from the viewpoint of single quantification and double quantification. The multigranulation-based single-quantitative rough set models, the related properties, and decision rules for single source based on the presented single-quantitative support characteristic levels are discussed. How to construct multigranulation-based single-quantitative and double-quantitative rough approximations in MSEDS and to analyze their corresponding properties and decision-making rules is addressed. To interpret the proposed multigranulation-based single-quantitative rough sets and double-quantitative rough sets in MSEDS, an illustrative example about an MSEDS formed by medical expert diagnosis is investigated, which is helpful to apply these developments to practical event-based decision-making problems.

A Prospective Approach to Integration of AI Fracture Detection Software in Radiographs into Clinical Workflow.

Gleamer BoneView© is a commercially available AI algorithm for fracture detection in radiographs. We aim to test if the algorithm can assist in better sensitivity and specificity for fracture detection by residents with prospective integration into clinical workflow. Radiographs with inquiry for fracture initially reviewed by two residents were randomly assigned and included. A preliminary diagnosis of a possible fracture was made. Thereafter, the AI decision on presence and location of possible fractures was shown and changes to diagnosis could be made. Final diagnosis of fracture was made by a board-certified radiologist with over eight years of experience, or if available, cross-sectional imaging. Sensitivity and specificity of the human report, AI diagnosis, and assisted report were calculated in comparison to the final expert diagnosis. 1163 exams in 735 patients were included, with a total of 367 fractures (31.56%). Pure human sensitivity was 84.74%, and AI sensitivity was 86.92%. Thirty-five changes were made after showing AI results, 33 of which resulted in the correct diagnosis, resulting in 25 additionally found fractures. This resulted in a sensitivity of 91.28% for the assisted report. Specificity was 97.11, 84.67, and 97.36%, respectively. AI assistance showed an increase in sensitivity for both residents, without a loss of specificity.

A predictive and normative model for the “smart city” implementation concept in expert discourse

The smart city concept strikes a discourse in the academic, media, political field in the present period. Discursive topics for that matter are the concept implementation model of choice determination, including its main structural parameters and the horizons of future develop-ment establishment. The challenge is the lack of a comprehensive approach to provide the effectiveness of social elements in the model of implementation of the concept of the smart city. The research issue includes an expert diagnosis parameter of the citizen as a subject. Set of parameters characterizing the social space of the city, namely, activity, identity, norms of trust and solidarity of the population, which together reflect the level of citizens’ subjectivity is proposed to take as a system indicator of the model of concept’s realization. The author model of the smart city concept implementation complements and expands the quadruplehelix model. The purpose of the work is aimed at predictive-normative model diagnostics by the expert community of the city for its integration into the management practice. Representatives of authorities, businesses, non-profit organizations of six cities as experts (Tyumen, Tobolsk, Khanty-Mansiysk, Surgut, Salekhard, Novy Urengoy) evaluated the implementation, prospects of the approach in their cities and the possibility of including the subjective model in management practice. The survey uncovered some degree of contradictions in the positions of experts on the predictive and normative model for the concept’s implementation. Most experts admit the increasing city dwellers subjectivity, but do not see opportunities and conditions for new forms of its institutionalization. As whole expert attitudes are not predictive, but rather demonstrate a prolongation of the existing provisions and trends of urban development. Nevertheless, the study showed that the parameters of the subject model are not in conflict with the existing model and can be used in the practice of urban management.

Automatic detection of active and inactive multiple sclerosis plaques using the Bayesian approach in susceptibility-weighted imaging.

Susceptibility-weighted imaging (SWI) is efficient in detecting multiple sclerosis (MS) plaques and evaluating the level of disease activity. To automatically detect active and inactive MS plaques in SWI images using a Bayesian approach. A 1.5-T scanner was used to evaluate 147 patients with MS. The area of the plaques along with their active or inactive status were automatically identified using a Bayesian approach. Plaques were given an orange color if they were active and a blue color if they were inactive, based on the preset signal intensity. Experimental findings show that the proposed method has a high accuracy rate of 91% and a sensitivity rate of 76% for identifying the type and area of plaques. Inactive plaques were properly identified in 87% of cases, and active plaques in 76% of cases. The Kappa analysis revealed an 80% agreement between expert diagnoses based on contrast-enhanced and FLAIR images and Bayesian inferences in SWI. The results of our study demonstrated that the proposed method has good accuracy for identifying the MS plaque area as well as for identifying the types of active or inactive plaques in SWI. Therefore, it might be helpful to use the proposed method as a supplemental tool to accelerate the specialist's diagnosis.

257. Coinfection Characteristics among Children with COVID-19

Abstract Background Coinfection in patients with SARS-CoV-2 has been associated with greater complications. We describe the clinical characteristics and outcomes of 126 pediatric patients with COVID-19 and viral, bacterial, or fungal coinfection. Methods We retrospectively reviewed and analyzed electronic data of all pediatric patients who tested positive for SARS-CoV-2 from April 16, 2020, to April 15, 2022, in our center. Confirmation of COVID-19 was based on positive RT-PCR. Viral coinfections (VC) were identified using a multiplex RT-PCR respiratory viral panel, bacterial coinfection (BC) was determined by positive bacterial culture (blood, bronchoalveolar lavage, sputum, urine) or clinical/radiological manifestations and antimicrobial assessment by a pediatric Infectious Diseases expert and fungal coinfection (FC) diagnosis based on Consensus definitions of invasive fungal disease. Results During the study period, among 400 pediatric patients with COVID-19, 126 children had coinfection. Children &amp;gt;10 years were the most affected age group. Underlying disease was present in 69%, hematological malignancies were the most common (17.5%). BC was detected in 76.9% (n=97), bacterial pneumonia (54.6%) was the main diagnosis, followed by oncologic patients with initial febrile neutropenia and posterior SARS-CoV-2 detection (14,4%). Unusual BC as congenital syphilis w detected; acute appendicitis was the initial presentation of COVID-19 in 8 patients. VC was identified in 15.87% (n=20), prevailing rhinovirus (9.5%) and adenovirus (3.96%), One FC presented as proven pulmonary aspergillosis (0.8%). B-V and B-F coinfection were detected in 2 patients. Fever and cough were the most common symptoms, higher fever &amp;gt;40°C was mostly observed in the BC group (3%). Twenty-seven patients with BC (27.8%) were admitted to intensive care, with the OR 0.7 IR 95% (0.611–1.008), 4.1% died. One ICU admission was observed in the VC group (5%) and all VC cases resolved without complications. Conclusion Pediatric patients with COVID-19 coinfection, especially BC were common in our center representing nearly one-third of the infected children, including unusual coinfections. BC was identified as a risk factor for ICU admission OR 0.7 IR 95% (0.611–1.008). Favorable outcomes were observed in most cases. Disclosures All Authors: No reported disclosures.

LANGaware: Robust Multimodal Machine Learning Methods for the Early Detection of Neurodegenerative and Psychiatric Diseases

AbstractBackgroundMultiple neurodegenerative and psychiatric diseases can affect speech, manifesting as subtle changes in spoken language. We perform a rigorous evaluation of a Machine Learning (ML) pipeline on a large‐scale multiaxial experimental setting, to investigate its capacity for detecting early signs of cognitive decline as manifested in speech. Using a diverse dataset of speech samples spanning different languages, patient cohorts and cognitive status diagnoses, we propose that the developed pipeline showcases robust performance, facilitating efficient decision support for early cognitive decline detection.MethodWe build upon previous research efforts on data‐driven cognitive decline prediction. Random forest models are constructed from a dataset of more than 3000 audio recordings and ASR‐generated transcripts, constituting cognitive assessment test tasks (e.g. picture description, everyday activity recounting, etc.). A multimodal NLP and audio analysis procedure derives sophisticated predictive biomarkers, using three methods for feature selection, feature decorrelation preprocessing and rigorous, grid search‐based hyperparameter optimization. We introduce novel experimental axes of investigation, including the spoken language (English and Greek), single and multi‐task verbal task settings, and different neurodegenerative diseases for the pathological cohorts (Alzheimer’s Disease or Dementia, along with age and education‐matched healthy cohorts). Four different NLP analysis tools for tokenization, syntax and dependency extraction are considered, using both ML and Deep Learning (DL) techniques. Model stability and generalization is supported via 5‐fold cross‐validation and the use of a separate test set.ResultResults are validated against medical expert diagnosis. Αverage performance across all investigative axes reaches high scores for accuracy (0.73), sensitivity (0.72), specificity (0.74), precision (0.77), F1 (0.74), and AUC (0.78). Additional findings suggest that single‐task models outperform multitask ones (e.g. reaching AUC scores of 0.85). Moreover, cognitive decline detection is more efficient for pathological cohorts of Alzheimer's, versus patients with Dementias.ConclusionThe broad empirical evaluation verifies the effectiveness of the deployed ML system, adding evidence in favor of multimodal AI solutions for predicting early cognitive decline and showcasing their robustness in diverse evaluation settings. Future work includes leveraging DL methods for classification and evaluation on differential diagnosis discrimination tasks.

Comparison of Naive Bayes and Dempster Shafer Methods in Expert System for Early Diagnosis of COVID-19

COVID-19 is a respiratory infection disease caused by the corona virus. Transmission of this virus can spread very quickly so that the number of cases of the corona virus continues to grow and becomes an epidemic that spreads not only in Indonesia but also in other countries in the world. The purpose of this study is to build an expert system that is able to diagnose Covid-19 early by using a comparison of the Nave Bayes method and the Dempster Shafer method. The amount of data used in this study is 550 data, consisting of 500 training data and 50 testing data. While the variables used are symptoms related to COVID-19 as many as 17 symptoms consisting of G01, G02, G03, G04, G05, G06, G07, G08, G09, G10, G11, G12, G13, G14, G15, G16, G17. The diagnostic data consists of Suspected (PDP), Non-Suspected, and Close Contact (ODP). The results of the percentage test by comparing system diagnoses with expert diagnoses, for the nave Bayes method it has an accuracy of 96% with 48 diagnoses according to expert diagnoses from 50 tested data. Meanwhile, the Dempster Shafer method has an accuracy of 40% with 20 diagnoses according to expert diagnoses from 50 tested data. Based on the results of this study, the Naive Bayes and Dempster Shafer methods can be applied to an expert system for early diagnosis of COVID-19, from the results of the system testing the Naive Bayes method has better accuracy than the Dempster Shafer method.

Extranodal Marginal Zone Lymphomas Show Recurrent Mutations in DNA Repair Genes, Cancer-Associated Proliferative Signaling and NOTCH1 Signaling Pathways, Regardless of Anatomic Site

Extranodal marginal zone lymphomas (EMZL) are usually indolent B cell lymphomas arising in acquired mucosa-associated lymphoid tissue (MALT) as a result of chronic antigen stimulation from infection or autoimmune processes. EMZLs occurring at different MALT sites have shared and unique genetic events including site-associated chromosomal structural changes and mutations. While large scale genetic analyses of EMZL are lacking, previous small studies have shown that EMZL often show activation of NF-κB through chronic stimulation of the B cell receptor and other cooperating genetic events that include somatic mutations, most frequently in TNFAIP3. Here, we report the largest whole exome sequencing (WES) and copy number (CN) analysis to date in EMZL. Formalin-fixed paraffin-embedded (FFPE) samples of EMZL were collected as part of the Atlas of Blood Cancer Genomes (ABCG) consortium. All cases underwent central pathology review by expert hematopathologists and diagnoses were confirmed as defined by the 2016 WHO classification. We defined global CN gains and losses at log2 fold change value of ±0.3, a value used by many CN tools. DNA and RNA WES were performed using the Illumina platform and DNA and RNA reads aligned to the GRCh38 genome and transcriptome, respectively. Exonic variants were identified and filtered using normal samples and population-based databases to identify putative driver mutations which were then aggregated at the gene level. Mutational analysis was done on samples that passed quality filtering. 225 unique diagnostic specimens were collected from 24 academic hematopathology departments. Cases originated from North America (USA and Canada, 86%), Europe (8%), and Asia (Singapore and China, 6%). 72 tumors were excluded due to inadequate mean coverage related to size of the biopsy or its quality. CN analysis and WES to identify non-silent mutations was performed in 153 unique EMZL including ocular adnexal (31), stomach (26), lung (23), salivary gland (23), and other extranodal locations (50). Average patient age was 63 years (range 53-71 years), female: male ratio of 1.6. Mutation calls identified 718 variants (613 missense mutations and 105 truncating mutations) across 59 putative driver genes detected in ≥5% cases across all the anatomic locations. Among the most commonly mutated genes, we detected genes previously reported to be altered by targeted sequencing in EMZL (e.g., SPEN (12%), TNFAIP3 (A20) (8%), CREBBP (8%), TBL1XR1 (7%), and others) confirming our filtering strategy. A total of 44 focal CN alterations (24 gains and 20 CN losses) were detected for our set of recurrently mutated genes. Global CN gains included chromosome 3 (6%), 6p (7%), 8 (3%), 12 (4%), and 18 (5%) and losses in 21p (10%). Previously described tumor suppressor genes such as TNFAIP3 showed a typical pattern of alterations consisting of CN losses (6q23.3 deletion in 6%) and/or mostly truncating mutations (8%). Combining DNA mutations and CN alterations, the number of genetic lesions per tumor sample in the 59 putative driver genes ranged from 1 to 21, with an average load of 4.9 aberrations per case, consistent with previous genetic studies in other types of MZL lymphoma. The most frequent mutations across all sites were IGLL5 (22%), KMT2C (15%), KMT2D (14%), SPEN (12%), PRDM2 (11%), and NCOR1 (10%). Analysis using Elsevier Pathway Collection, BioPlanet 2019 and Go Biological Process 2021 revealed adjusted statistically significant enrichment of aberrations in genes in NOTCH signaling, DNA repair, cancer-associated sustaining of proliferative signaling, cancer associated histone methylation and positive regulation of transcription, by DNA-templated (GO:0045893) and RNA polymerase II (GO:0045944). Many of the mutated genes were previously not reported in EMZL. Some of the mutations were shared by tumors at different anatomic locations while some were more unique. Mutations in TET2, previous reported as specific for thyroid EMZL, were also detected in ocular adnexa, gastric and salivary gland MZL. Mutations in PTPRD, previously suggested as specific for NMZL, were detected also in EMZL. Overall, this analysis of a large number of samples from different anatomic locations further refines the mutational landscape in EMZL and provides novel clues on pathogenesis. More detailed analyses of mutations, CN variations and RNA expression will be presented at the meeting.

Comparing Fuzzy Logic Mamdani and Naïve Bayes for Dental Disease Detection

Background: Dental disease detection is essential for the diagnosis of dental diseases. Objective: This research compares the Mamdani fuzzy logic and Naïve Bayes in detecting dental diseases. Methods: The first is to process data on dental disease symptoms and dental support tissues based on complaints of toothache consulted with experts at a community health centre (puskesmas). The second is to apply the Mamdani fuzzy logic and the Naïve Bayes to the proposed expert system. The third is to provide recommended decisions about dental diseases based on the symptom data inputted into the expert system. Patient data were collected at the North Cilacap puskesmas between July and December 2021. Results: The Mamdani fuzzy logic converts uncertain values into definite values, and the Naïve Bayes method classifies the type of dental disease by calculating the weight of patients’ answers. The methods were tested on 67 patients with dental disease complaints. The accuracy rate of the Mamdani fuzzy logic was 85.1%, and the Naïve Bayes method was 82.1%. Conclusion: The prediction accuracy was compared to the expert diagnoses to determine whether the Mamdani fuzzy logic method is better than the Naïve Bayes method. Keywords: Dental Disease, Expert System, Mamdani Fuzzy Logic, Naïve Bayes, Prediction

MACHINE LEARNING-BASED CLINICAL DECISION SUPPORT SYSTEM

This article aims to design a self-organizing decision support system for early detection of important physiological events. The proposed system consists of pre-processing, clustering and diagnostic systems based on self-organizing fuzzy logic modelling. Clustering methods have been used in empirical pattern analysis. Especially when the available information is incomplete or the data model is subject to ambiguity. This applies primarily to medical/clinical data. A clustering engine can be viewed as unsupervised learning from a given data set. This module analyses a patient's vital signs to identify important relationships, patterns, and clusters among medical data. Then, use self-organizing fuzzy logic modelling to detect symptoms and events early. Based on the clustering results, we observed a high degree of agreement between the system's interpretation and the human expert's diagnosis of physiological events and signs in abnormal sign detection. Keywords: Patient monitoring; Early diagnosis of clinical events; Clinical decision support system; Self-organising fuzzy system; Machine learning; Clustering analysis.

A Single-Center, Observational Study of 607 Children and Young People Presenting With Differences of Sex Development (DSD).

Differences of sex development (DSD) represent a wide range of conditions presenting at different ages to various health professionals. Establishing a diagnosis, supporting the family, and developing a management plan are important. We aimed to better understand the presentation and prevalence of pediatric DSD. A retrospective, observational cohort study was undertaken in a single tertiary pediatric center of all children and young people (CYP) referred to a DSD multidisciplinary team over 25 years (1995-2019). In total, 607 CYP (520 regional referrals) were included. Data were analyzed for diagnosis, sex-assignment, age and mode of presentation, additional phenotypic features, mortality, and approximate point prevalence. Among the 3 major DSD categories, sex chromosome DSD was diagnosed in 11.2% (68/607) (most commonly 45,X/46,XY mosaicism), 46,XY DSD in 61.1% (371/607) (multiple diagnoses often with associated features), while 46,XX DSD occurred in 27.7% (168/607) (often 21-hydroxylase deficiency). Most children (80.1%) presented as neonates, usually with atypical genitalia, adrenal insufficiency, undescended testes or hernias. Those presenting later had diverse features. Rarely, the diagnosis was made antenatally (3.8%, n = 23) or following incidental karyotyping/family history (n = 14). Mortality was surprisingly high in 46,XY children, usually due to complex associated features (46,XY girls, 8.3%; 46,XY boys, 2.7%). The approximate point prevalence of neonatal referrals for investigation of DSD was 1 in 6347 births, and 1 in 5101 overall throughout childhood. DSD represent a diverse range of conditions that can present at different ages. Pathways for expert diagnosis and management are important to optimize care.

ADL-CDF: A Deep Learning Framework for COVID-19 Detection from CT Scans Towards an Automated Clinical Decision Support System.

The emergence of deep learning has paved to solve many problems in the real world. COVID-19 pandemic, since the late 2019, has been affecting lives of people across the globe. Chest CT scan images are used to detect it and know its severity in patients. The problem with many existing solutions in COVID-19 detection using CT scan images is that inability to detect the infection when it is in initial stages. As the infection can exist on varied scales, there is need for more comprehensive approach that can ascertain the disease at all scales. Towards this end, we proposed a deep learning-based framework known as Automated Deep Learning-based COVID-19 Detection Framework (ADL-CDF). It does not need a human medical expert in diagnosis as it is capable of detecting automatically. The framework is assisted by two algorithms that involve image processing and deep learning. The first algorithm known as Region of Interest (ROI)-based Image Filtering (ROI-IF) which analyses given input CT scan images of a patient and discards the ones where ROI is missing. This algorithm minimizes time taken for processing besides reducing false positive rate. The second algorithm is known as Multi-Scale Feature Selection algorithm that fits into the deep learning framework’s pipeline to leverage detection performance of the ADL-CDF. The proposed framework is evaluated against ResNet50V2 and Xception. Our empirical study revealed that our model outperforms the state of the art.