Severity Classification Research Articles

The Relationship between Tooth Brushing Techniques and Gingival Status in Patients Who Wear Fixed Orthodontics at a Private Dental Clinic in Makassar City

Background: Dental and oral hygiene is an important factor that must be maintained by orthodontic users. An effort to maintain healthy teeth and mouth is by brushing your teeth. Patients who use a deaf orthodontic device must be able to maintain their oral cavity to avoid the onset of caries and marginal gingivitis. Because there is a relationship between the brushing technique and gingival status in patients who wear fixed orthodontics. Research objectives: To find out the relationship between tooth brushing techniques and gingival status in patients who wear fixed orthodontics at a private dental clinic in Makassar City in 2024. Methods: The type of research used in this study is quantitative research using a survey approach. Survey research is a research method that takes samples from a population. The data and information collected by the researcher came from respondents. Results : The results of the study were obtained that there were several patients who had gingival status with different categories, including the mild category as many as 5 gingiva indexes, and the moderate category as many as 4 gingiva indexes, while the severe category had 4 gingiva indexes. It can be seen that orthodontic users are very susceptible to experiencing poor oral hygiene caused by respondents who do not maintain their oral hygiene. Conclusion : There is a relationship between the tooth brushing technique and gingival status in patients who wear fixed orthodontics at a private dental clinic in Makassar City in 2024.

Discriminative validity and interpretability of the mood and feelings questionnaire

BackgroundUsing the Mood and Feelings Questionnaire (MFQ) to differentiate between depression severity levels remains unexplored. We explored the discriminative validity of the MFQ to identify an optimal cut-off MFQ score to distinguish between subthreshold-to-mild and moderate-to-severe depression severity levels. MethodsAn observational cross-sectional design was used in a sample (N = 67) of help-seeking youth (ages 13 to 18, inclusive) experiencing depressive symptoms. The MFQ was administered verbatim to youth by a research analyst over the phone. Youth were then grouped into subthreshold-to-mild or moderate-to-severe depression severity categories based on scores received on the Kiddie Schedule for Affective Disorders and Schizophrenia-Depression Rating Scale. Receiver Operating Characteristic curve analyses were conducted, with area under the curve (AUC) and Youden Index (J) as primary indices. We hypothesized that the lower limit of the 95 % confidence interval for the area under the curve would be ≥0.70. ResultsThe primary analysis yielded an AUC of 0.85 (95 % CI: 0.763–0.947) and an optimal cut-off of ≥43 (J = 0.60, positive predictive value = 91.4 %, negative predictive value = 62.5 %, sensitivity = 72.7 %, specificity = 87.0 %). LimitationsOur study collected a small sample, and as such cannot identify how subgroup classification (e.g., based on race or gender) may moderate outcomes. Further, unknown measurement error of the predictor and reference variable measures can bias the estimates. ConclusionsOur preliminary findings highlight the potential for the MFQ to support clinical decision-making relevant to adolescents experiencing varying severities of depressive symptoms in secondary care settings.

Thoracic computed tomography (CT) image-based identification and severity classification of COVID-19 cases using vision transformer (ViT)

In this research, we developed a two-stage deep learning (DL) model using Vision Transformer (ViT) to detect COVID-19 and assess its severity from thoracic CT images. In the first stage, we utilized a pre-trained ViT model (ViT_B/32) and a custom CNN model to classify CT images as COVID-19 or non-COVID-19. The ViT model achieved superior performance with a fivefold cross-validated accuracy of 99.7%, compared to the custom CNN’s 98%. In the second stage, we employed a ViT-based U-Net model (Vision Transformer for Biomedical Image Segmentation, VITBIS) to segment lung and infection regions in COVID-19 positive CT images, determining the infection severity. This model uses transformers with attention mechanisms in both the encoder and decoder. The lung segmentation network achieved an Intersection Over Union (IOU) of 95.8% and a sensitivity of 99.67%, while the lesion segmentation network attained an IOU of 94% and a sensitivity of 98.3%.

Treatment-induced anogenital melanosis is a very frequent finding in patients with vulvar lichen sclerosus.

Pigmented lesions such as melanosis have rarely been reported in patients with vulvar lichen sclerosus (VLS) that is typically characterized by hypopigmented lesions. We aimed to analyze systematically anogenital melanosis in a large cohort of VLS patients. We analyzed the clinical data of 198 female patients with VLS. The anogenital lesions of all patients were professionally photographed in a standardized position and illumination. Severity classification of architectural findings followed an easy-to-use clinical score. A modified Melasma Area and Severity Index and an image analysis software were used to evaluate the area and intensity of pigmentation. According to the clinical score, 79 (198/39.9%) patients showed grade 1 disease, 78 (198/39.4%) grade 2, 37 (198/18.7%) grade 3, and 4 (198/2%) grade 4 disease. About 111 (56.1%) of the 198 patients had anogenital melanosis with a median modified Melasma Area and Severity Index of 3.6 (0.4-14). Univariate analysis revealed that anogenital melanosis was positively correlated with the use of topical estrogens (P = .0018) and negatively correlated with the use of pulsed high-dose corticosteroids plus low-dose methotrexate (PHDC-LDM, P = .021). On multivariable analysis, the use of topical hormone therapy turned out to be a strong independent predictor for the presence of anogenital melanosis (odds ratio: 4.57, 95% confidence interval: 1.66-12.57, P = .0033), whereas PHDC-LDM use was an independent predictor for the absence of anogenital melanosis (odds ratio: 0.35, 95% confidence interval: 0.15-0.84, P = .018). The study includes the retrospective monocentric design. Anogenital melanosis is a very frequent and so far, under-reported clinical finding in VLS patients. It is likely caused by the use of topical estrogens employed for VLS treatment. In contrast, patients with more severe disease and PHDC-LDM treatment appear to develop less likely anogenital melanosis.

Spectral sensor‐based device for real‐time detection and severity estimation of groundnut bud necrosis virus in tomato

AbstractA machine learning‐based approach was utilized to develop a device for groundnut bud necrosis virus (GBNV) disease severity detection and estimation in tomato plants (Solanum lycopersicum L.). The study involved inoculating tomato plants with GBNV, monitoring changes in morphological and spectral characteristics, evaluating machine learning algorithms (decision tree [DT] classifier) for analysis and classification of disease severity, and developing and validating a device for disease detection and severity estimation. Spectral data analysis revealed distinct patterns in reflectance, with notable peaks observed in the 680 and 760 nm bands, while reflectance remained low and constant beyond 900 nm. Machine learning techniques, specifically a DT model, were employed to classify disease severity based on spectral data with high accuracy (95.01% training accuracy and 93.65% testing accuracy). The model identified the near‐infrared band as highly correlated (correlation coefficient of 0.82) with disease severity. Furthermore, a compact handheld device integrating a spectral sensor, organic light‐emitting diode display, and Raspberry Pi 3B was developed for real‐time disease severity estimation. The device demonstrated robust performance, accurately predicting disease severity at different growth stages, even in the absence of visible symptoms. Additionally, disease severity percentages obtained via reverse transcription polymerase chain reaction were used to validate the accuracy of the device's estimations. Its responsive nature, with estimated response times ranging from milliseconds to seconds, facilitates timely interventions in agricultural settings. Overall, this interdisciplinary approach, combining spectral analysis, machine learning, and device development, presents a promising solution for efficient disease monitoring and management in agriculture, contributing to enhanced crop health and food security.

Monocentric Retrospective Analysis of Clinical Outcomes, Complications, and Adjacent Segment Disease in 507 Patients Undergoing Anterior Cervical Discectomy and Fusion for Degenerative Cervical Myelopathy

BackgroundDegenerative Cervical Myelopathy (DCM) is a leading cause of non-traumatic spinal cord injury. Surgery aims to arrest neurological decline and improve conditions, but controversies surround risks and benefits in elderly patients, outcomes in mild myelopathy, and the risk of adjacent segment disease (ASD). MethodsRetrospective data of patients who underwent ACDF for DCM in our hospital were collected. Patients were stratified by preoperative mJOA (mild, moderate, severe) and age (Under 70, Over 70). Clinical outcomes, complications, and ASD rate were analyzed. We evaluated the relationship between mJOA recovery rate and the risk of complications and various preoperative parameters. Results507 consecutive patients were included in the study, with a mean follow-up of 43.52 months (12-71). Improvement in all outcome variables was observed in mild, moderate and severe myelopathy categories, with elderly patients showing a lower improvement. Except for age, no other variable correlated with mJOA recovery rate. We observed 45 complications (11.1% of patients), with 14 in the U70 group and 31 in the O70 group (p-value<0.001). Age, Charlson Comorbidity index and ASA score were found to be predictors of complications. Fourteen patients (2.8% of total), mean age 54.2, developed radiological and clinical ASD. Most had cranial-level ASD with Pfirmann grade >= 2 before index surgery. ConclusionsMost myelopathic patients improve after ACDF. Elderly patients show a lower improvement and higher complication rates than younger counterparts. ASD rates are low, and younger patients with preexisting cranial level alterations are more susceptible.

Automated Lung Cancer Diagnosis Applying Butterworth Filtering, Bi-Level Feature Extraction, and Sparce Convolutional Neural Network to Luna 16 CT Images.

Accurate prognosis and diagnosis are crucial for selecting and planning lung cancer treatments. As a result of the rapid development of medical imaging technology, the use of computed tomography (CT) scans in pathology is becoming standard practice. An intricate interplay of requirements and obstacles characterizes computer-assisted diagnosis, which relies on the precise and effective analysis of pathology images. In recent years, pathology image analysis tasks such as tumor region identification, prognosis prediction, tumor microenvironment characterization, and metastasis detection have witnessed the considerable potential of artificial intelligence, especially deep learning techniques. In this context, an artificial intelligence (AI)-based methodology for lung cancer diagnosis is proposed in this research work. As a first processing step, filtering using the Butterworth smooth filter algorithm was applied to the input images from the LUNA 16 lung cancer dataset to remove noise without significantly degrading the image quality. Next, we performed the bi-level feature selection step using the Chaotic Crow Search Algorithm and Random Forest (CCSA-RF) approach to select features such as diameter, margin, spiculation, lobulation, subtlety, and malignancy. Next, the Feature Extraction step was performed using the Multi-space Image Reconstruction (MIR) method with Grey Level Co-occurrence Matrix (GLCM). Next, the Lung Tumor Severity Classification (LTSC) was implemented by using the Sparse Convolutional Neural Network (SCNN) approach with a Probabilistic Neural Network (PNN). The developed method can detect benign, normal, and malignant lung cancer images using the PNN algorithm, which reduces complexity and efficiently provides classification results. Performance parameters, namely accuracy, precision, F-score, sensitivity, and specificity, were determined to evaluate the effectiveness of the implemented hybrid method and compare it with other solutions already present in the literature.

Assessment of the Quality of Cause-of-Death Data in Serbia for 2005-2019 Vital Statistics Performance Index Estimation.

We aimed to evaluate the quality of the cause of death (COD) concerning mortality patterns and completeness of death registration to identify areas for improvement in Serbia. COD data collected from the mortality register in Serbia from 2005 to 2019 (1540615 deaths) were analyzed with the software Analysis of National Causes of Death for Action. The Vital Statistics Performance Index for Quality (VSPI(Q)) is estimated for the overall COD data quality. The completeness of death certification was higher than 98%. Usable underlying COD was registered in 57%, 24.1% with an unusable and 18.6% with insufficiently specified COD. The VSPI(Q) was 67.2%, denoting medium quality. The typical error was using intermediate COD (24.7% of all deaths), while 13.2% and 8.5% of all garbage codes (GC) belonged to the Very High and High Severity classes. The leading underlying COD is unspecified cardiomyopathy. The analysis revealed that 39.1% of GC has been redistributed to non-communicable diseases, 2.5% to external causes and 1.1% to communicable diseases. In the 15 years' worth of data analyzed, the true underlying COD, in many cases, was ill-defined, indicating that COD data at the national level could be distorted. The additional and continuous professional education of medical students as well as physicians is needed. It should focus on the most common GC among the leading COD and acquiring skills in certifying external causes of death.

Stress, Anxiety and Depression in First-year Medical Undergraduates and Coping Strategies Employed during COVID-19-induced Online Learning: A Cross-sectional Study from North India

Abstract Background: The COVID-19 pandemic, due to its rising uncertainties, was expected to negatively impact the overall psychological functioning of students undergoing first year of medical professional education with changing teaching styles and environmental restraints. Objective: The objective was to assess the impact of this stressful period on the first-year Bachelor of Medicine and Bachelor of Surgery (MBBS) students at an Institute of National importance in North India, by measuring their various emotional states such as stress, anxiety, depression and coping strategies. Methods: This was a descriptive cross-sectional study undertaken in first-year MBBS students who were admitted in February 2021 due to delay in the admission process. Their physical classes were suspended due to the onset of second wave of COVID-19 and underwent a significant shift from conventional classes to a hybrid teaching module involving online classes, minimal face-to-face contact with teachers and almost no practical exposure which is in contrast to the conventional teaching practices. Data collection was done during and in the aftermath of second wave of COVID-19 using the Depression, Anxiety and Stress Scale (DASS-21) to assess emotional state and the Brief-Coping Orientation to Problems Experienced Inventory to measure the coping strategies employed. Results: One hundred and twenty-five students were approached and 80 consented to participate. The 80 students were analysed. Elevated depressive symptomatology scores were found in 46% of students, of which 14% had scores falling in severe category. More than 62% had elevated anxiety scores, of which 26% had severe range of scores, and 37% students had elevated stress scores, of which more than 10% had severe range of scores. Venting, humour and self-blame were found to be the predominant facets of coping styles employed and they had a significant statistical correlation with all three subscales of DASS-21. Conclusion: The study found a significant prevalence of depressive, anxiety and stress symptomatology. These findings emphasise the need for mental health interventions and support programmes for medical college MBBS students to help them cope with the psychological impact of academic and personal stressors.

Automated lost circulation severity classification and mitigation system using explainable Bayesian optimized ensemble learning algorithms

Lost circulation and mud losses cause 10 to 20% of the cost of drilling operations under extreme pressure and temperature conditions. Therefore, this research introduces an integrated system for an automated lost circulation severity classification and mitigation system (ALCSCMS). This proposed system allows decision makers to reliability predict lost circulation severity (LCS) based on a few drilling drivers before starting drilling operations. The proposed system developed and compared a total of 11 ensemble machine learning (EML) based on collection 65,377 observations, the data was pre-processed, cleaned, and normalized to be filtered using factor analysis. For each generated algorithm, the proposed system performed Bayesian optimization to acquire the best possible results. As a result, the optimized random forests (RF) model algorithm was the optimal model for classification at 100% classification accuracy based on testing data set. Mitigation optimization model based on genetic algorithm has been incorporated to convert high severe classes into acceptable classes of lost circulation. The system classifies the LCS into 5 classes where the classes from 2 to 4 are converted to be class 0 or 1 to minimize lost circulation severity by optimizing the input parameters. Therefore, the proposed model is reliable to predict and mitigate lost circulation during drilling operations. The main drivers that served as LCS inputs were explained using the SHapley Additive exPlanations (SHAP) approach.

Agreement of Different Drug-Drug Interaction Checkers for Proton Pump Inhibitors

ImportanceProton pump inhibitors (PPIs) are a widely prescribed class of drugs, potentially interacting with a large number of medicines, especially among older patients with multimorbidity and polypharmacy. Beyond summary of product characteristics (SPCs), interaction checkers (ICs) are routinely used tools to help clinicians in medication review interventions.ObjectiveTo assess the consistency of information on drugs potentially interacting with PPIs as reported in their SPCs and different ICs.Design, Setting, and ParticipantsThis cross-sectional study was conducted using data from SPCs for 5 PPIs (omeprazole, esomeprazole, lansoprazole, pantoprazole, and rabeprazole) and 5 ICs (ie, INTERCheck WEB, Micromedex, Lexicomp, Epocrates, and drugs.com). Information from the SPCs and the ICs were extracted between July 15 and 30, 2023.Main Outcomes and MeasuresThe main outcome was the level of agreement among SPCs and the 5 ICs in identifying drugs potentially interacting with PPIs and attributing drug-drug interaction (DDI) severity categories. The level of agreement was computed using Gwet AC1 statistic on the 5 ICs and by comparing 4-sets and 2-sets of ICs. As a sensitivity analysis, the level of agreement in listing PPI-related DDIs was evaluated using Cohen κ and Fleiss κ coefficients.ResultsConsidering SPCs and the 5 ICs, a total of 518 potentially interacting drugs with omeprazole were reported, 455 for esomeprazole, 433 for lansoprazole, 421 for pantoprazole, and 405 for rabeprazole. As compared with the ICs, the SPCs reported a much smaller number of drugs potentially interacting with PPIs, with proportions ranging from 2.7% (11 potentially interacting drugs) for rabeprazole to 7.6% (33 potentially interacting drugs) for lansoprazole of the total identified drugs at risk of interaction with a PPI. The overall level of agreement among the 5 ICs for identifying potential interactions was poor (from 0.23 [95% CI, 0.21-0.25] for omeprazole to 0.27 [95% CI, 0.24-0.29] for pantoprazole and 0.27 [95% CI, 0.25-0.29] for rabeprazole). Similarly, the level of agreement was low in 4-set and 2-set analyses as well as when restricting the analysis to the potential DDIs identified as severe (range, 0.30-0.32).Conclusions and RelevanceThis cross-sectional study found significant disagreement among different ICs and SPCs, highlighting the need to focus on standardizing DDI databases. Therefore, to ensure evaluation and prevention of clinically relevant DDIs, it is recommended to revise multiple ICs and consult with specialists, such as clinical pharmacologists, particularly for patients with complex medical conditions.

Two-Stage Classification of Future Knee Osteoarthritis Severity After 8 Years Using MRI: Data from the Osteoarthritis Initiative.

Currently, there are no methods or tools available in clinical practice for classifying future knee osteoarthritis (KOA). In this study, we aimed to fill this gap by classifying future KOA into three severity grades: KL01 (healthy), KL2 (moderate), and KL34 (severe) based on the Kellgren-Lawrance scale. Due to the complex nature of multiclass classification, we used a two-stage method, which separates the classification task into two binary classifications (KL01 vs. KL234 in the first stage and KL2 vs. KL34 in the second stage). Our machine learning (ML) model used two Balanced Random Forest algorithms and was trained with gender, age, height, weight, and quantitative knee morphology obtained from magnetic resonance imaging. Our training dataset comprised longitudinal 8-year follow-up data of 1213 knees from the Osteoarthritis Initiative. Through extensive experimentation with various feature combinations, we identified KL baseline and weight as the most essential features, while gender surprisingly proved to be one of the least influential feature. Our best classification model generated a weighted F1 score of 79.0% and a balanced accuracy of 65.9%. The area under the receiver operating characteristic curve was 83.0% for healthy (KL01) versus moderate (KL2) or severe (KL34) KOA patients and 86.6% for moderate (KL2) versus severe (KL34) KOA patients. We found a statistically significant difference in performance between our two-stage classification model and the traditional single-stage classification model. These findings demonstrate the encouraging results of our two-stage classification model for multiclass KOA severity classification, suggesting its potential application in clinical settings in future.

Subtotal laparoscopic cholecystectomy versus conversion to open as a bailout procedure: a cohort study

BackgroundThe aim of this study is to evaluate morbidity and mortality in patients taken to conversion to open procedure (CO) and subtotal laparoscopic cholecystectomy (SLC) as bailout procedures when performing difficult laparoscopic cholecystectomy.MethodThis observational cohort study retrospectively analyzed patients taken to SLC or CO as bailout surgery during difficult laparoscopic cholecystectomy between 2014 and 2022. Univariable and multivariable logistic regression models were used to identify prognostic factors for morbimortality.ResultsA total of 675 patients were included. Of the 675 patients (mean [SD] age 63.85 ± 16.00 years; 390 [57.7%] male) included in the analysis, 452 (67%) underwent CO and 223 (33%) underwent SLC. Overall, neither procedure had an increased risk of major complications (89 [19.69%] vs 35 [15.69%] P.207). However, CO had an increased risk of bile duct injury (18 [3.98] vs 1 [0.44] P.009), bleeding (mean [SD] 165.43 ± 368.57 vs 43.25 ± 123.42 P < .001), intestinal injury (20 [4.42%] vs 0 [0.00] P.001), and wound infection (18 [3.98%] vs 2 [0.89%] P.026), while SLC had a higher risk of bile leak (15 [3.31] vs 16 [7.17] P.024). On the multivariable analysis, Charlson comorbidity index (odds ratio [OR], 1.20; CI95%, 1.01–1.42), use of anticoagulant agents (OR, 2.56; CI95%, 1.21–5.44), classification of severity of cholecystitis grade III (OR, 2.96; CI95%, 1.48–5.94), and emergency admission (OR, 6.07; CI95%, 1.33–27.74) were associated with presenting major complications.ConclusionsSLC was less associated with complications; however, there is scant evidence on its long-term outcomes. Further research is needed on SLC to establish if it is the safest in the long-term as a bailout procedure.

Pain Management in Patients With Lumbar Pain Syndrome Through Combined Physical and Kinetic Therapies

The aim of this study is to highlight the importance of associating various means of electrotherapy and kinesiotherapy in patients with low back pain. The study included 30 patients, diagnosed with low back pain syndrome, who underwent a complex treatment of physical therapy and kinesiotherapy associated with drug treatment. A questionnaire based on a series of standard questions, marked with 0-5 points, ranging from no pain and reduced functionality to intense pain and functional impairment, was used to quantify the intensity of pain and limitation of functionality due to pain. Data centralization, analysis, and comparison allowed us to identify the dynamic evolution of the monitored parameters so that after the application of associated therapies we found that if at the beginning of the treatment the majority of patients 70% (21 cases) were in severity class I established on the basis of the scores obtained by the patients in the questionnaire applied, at the end of the treatment only 13.3% (4 cases) remained in severity class I. In conclusion, the use of electrotherapy ensures the rapid achievement of optimal local conditions in the affected structures, removing the algesic component and thus preparing them for the kinetic program, thus creating a perfect symbiosis between the two methods of recovery of patients with low back pain syndrome.

Associations between mental health & substance use treatment and alcohol use progression and recovery among US women drinkers.

Alcohol use has profound public health impact on women; however, modifiable factors that may influence alcohol use progression/recovery, including health service utilization, are understudied in women. To investigate the association between mental health (MH) and substance use (SU) treatment with alcohol use progression and recovery among women who currently use alcohol or have in the past. This study is a secondary data analysis of prospective data from waves 1 (2001-2002) and 2 (2004-2005) of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC; a US-nationally representative sample of adults). The analytic sample was limited to women who reported past or current alcohol use at wave 1 (N = 15,515). Latent transition analysis (LTA) examined whether receiving SU/MH treatment in the year prior to wave 1 was associated with transitioning between three empirically-derived stages of alcohol involvement (no, moderate, and severe problems classes), between Waves 1 and 2 adjusting for possible confounders using propensity score weight. Compared to White female drinkers, female drinkers who were from Black, Hispanic, or other races were less likely to receive SU/MH treatment (p-values ≤. 001). SU/MH treatment in the year prior to wave 1 was associated with transitioning from the moderate problems class to the no problems class between Waves 1 and 2 (p-value = .04). Receipt of SU or MH treatment among women, was associated with a higher likelihood of remission from moderate alcohol use problems to no problems over time. Future research, including investigation into treatment characteristics (e.g., frequency, duration, type) should further explore why women initially experiencing severe alcohol use problems did not experience similar remission.

Automatic diagnosis of pediatric high myopia via Attention-based Patch Residual Shrinkage network

Fundus images can be obtained non-invasively and be adopted to monitor the follow-up on various fundus diseases, such as high myopia. Therefore, the use of fundus images for the early screening of eye diseases has principal clinical significance. However, AI-based medical research continues to face two main challenges, lack of prior-knowledge guidance and complex fundus information. In this paper, we propose a pediatric high myopia classification model, the Attention-based Patch Residual Shrinkage network (APRSnet), which facilitates early clinical diagnosis by evaluating the importance of different image features. A collection of 2492 high-resolution fundus images of children, including 768 images of high myopia and 1724 images of non-high myopia, is fed to APRSnet as the training dataset. To better understand how different features impact the classification, APRSnet is tested on multiple feature-enhanced fusion datasets. The result shows that, by removing irrelevant information in fundus images, APRSnet achieves an accuracy of 0.959 and an F1 score of 0.946 and outperforms all classic image classification networks we compared with. We also testify that, in fundus images, gradient information benefits severity classification the most since it helps with model convergence more than luminance and texture information.

Smart decision support system for keratoconus severity staging using corneal curvature and thinnest pachymetry indices

BackgroundThis study proposes a decision support system created in collaboration with machine learning experts and ophthalmologists for detecting keratoconus (KC) severity. The system employs an ensemble machine model and minimal corneal measurements.MethodsA clinical dataset is initially obtained from Pentacam corneal tomography imaging devices, which undergoes pre-processing and addresses imbalanced sampling through the application of an oversampling technique for minority classes. Subsequently, a combination of statistical methods, visual analysis, and expert input is employed to identify Pentacam indices most correlated with severity class labels. These selected features are then utilized to develop and validate three distinct machine learning models. The model exhibiting the most effective classification performance is integrated into a real-world web-based application and deployed on a web application server. This deployment facilitates evaluation of the proposed system, incorporating new data and considering relevant human factors related to the user experience.ResultsThe performance of the developed system is experimentally evaluated, and the results revealed an overall accuracy of 98.62%, precision of 98.70%, recall of 98.62%, F1-score of 98.66%, and F2-score of 98.64%. The application's deployment also demonstrated precise and smooth end-to-end functionality.ConclusionThe developed decision support system establishes a robust basis for subsequent assessment by ophthalmologists before potential deployment as a screening tool for keratoconus severity detection in a clinical setting.

Efficient deep learning-based automated diagnosis from echocardiography with contrastive self-supervised learning

BackgroundAdvances in self-supervised learning (SSL) have enabled state-of-the-art automated medical image diagnosis from small, labeled datasets. This label efficiency is often desirable, given the difficulty of obtaining expert labels for medical image recognition tasks. However, most efforts toward SSL in medical imaging are not adapted to video-based modalities, such as echocardiography.MethodsWe developed a self-supervised contrastive learning approach, EchoCLR, for echocardiogram videos with the goal of learning strong representations for efficient fine-tuning on downstream cardiac disease diagnosis. EchoCLR pretraining involves (i) contrastive learning, where the model is trained to identify distinct videos of the same patient, and (ii) frame reordering, where the model is trained to predict the correct of video frames after being randomly shuffled.ResultsWhen fine-tuned on small portions of labeled data, EchoCLR pretraining significantly improves classification performance for left ventricular hypertrophy (LVH) and aortic stenosis (AS) over other transfer learning and SSL approaches across internal and external test sets. When fine-tuning on 10% of available training data (519 studies), an EchoCLR-pretrained model achieves 0.72 AUROC (95% CI: [0.69, 0.75]) on LVH classification, compared to 0.61 AUROC (95% CI: [0.57, 0.64]) with a standard transfer learning approach. Similarly, using 1% of available training data (53 studies), EchoCLR pretraining achieves 0.82 AUROC (95% CI: [0.79, 0.84]) on severe AS classification, compared to 0.61 AUROC (95% CI: [0.58, 0.65]) with transfer learning.ConclusionsEchoCLR is unique in its ability to learn representations of echocardiogram videos and demonstrates that SSL can enable label-efficient disease classification from small amounts of labeled data.

A novel framework integrating ensemble transfer learning and Ant Colony Optimization for Knee Osteoarthritis severity classification

A novel framework integrating ensemble transfer learning and Ant Colony Optimization for Knee Osteoarthritis severity classification

Analyzing the impacts of watershed management interventions and scenarios on sediment export reduction in yezat watershed, north west Ethiopia using InVEST model

ABSTRACT This study aimed at assessing the long term impacts of watershed management interventions and scenarios on sediment export reduction in the data-scarce region of the Yezat watershed in north-west Ethiopia. The modified ecosystem services valuation model (InVEST) was employed to evaluate the amount of sediment export in the watershed. Different watershed management intervention scenarios were applied to evaluate the effectiveness of watershed management interventions on sediment export reduction. The results of the study showed that the mean annual sediment yield was 35 tons ha yr in 20000, due to the expansion of cultivated land and built-up area. Conversely, it decreased from 35 ton ha yr in 2000 -18.4 ton ha yr in 2021 due to watershed management interventions. On the contrary, moderate, severe, and very severe categories of sediment export intensity were high in 2000. But it decreased in 2021 due to watershed management interventions. The mean annual sediment export in the baseline scenario (BS) was 35 tons per year, which was reduced to 13.39 tons per year, 3.46 tons per year, and 1.17 tons per year in the reforestation scenario, soil/stone bund, and integrated scenarios, respectively. The result revealed that areas covered by high sediment severity classes were reduced.