Search Results (28)

Background/Objectives: The clinical management of adolescent idiopathic scoliosis (AIS) is hindered by the inability to accurately predict curve progression. Although skeletal maturity and the initial Cobb angle are established predictors of progression, their combined predictive accuracy remains limited. This study aimed to develop a robust and interpretable artificial intelligence (AI) system using deep learning (DL) models to predict the progression of scoliosis using only standing frontal radiographs. Methods: We conducted a multicenter study involving 542 patients with AIS. After excluding 52 borderline progression cases (6–9° progression in the Cobb angle), 294 and 196 patients were assigned to progression (≥10° increase) and non-progression (≤5° increase) groups, respectively, considering a 2-year follow-up. Frontal whole spinal radiographs were preprocessed using histogram equalization and divided into two regions of interest (ROIs) (ROI 1, skull base–femoral head; ROI 2, C7–iliac crest). Six pretrained DL models, including convolutional neural networks (CNNs) and transformer-based models, were trained on the radiograph images. Gradient-weighted class activation mapping (Grad-CAM) was further performed for model interpretation. Results: Ensemble models outperformed individual ones, with the average ensemble model achieving area under the curve (AUC) values of 0.769 for ROI 1 and 0.755 for ROI 2. Grad-CAM revealed that the CNNs tended to focus on the local curve apex, whereas the transformer-based models demonstrated global attention across the spine, ribs, and pelvis. Models trained on ROI 2 performed comparably with respect to those using ROI 1, supporting the feasibility of image standardization without a loss of accuracy. Conclusions: This study establishes the clinical potential of transformer-based DL models for predicting the progression of scoliosis using only plain radiographs. Our multicenter approach, high AUC values, and interpretable architectures support the integration of AI into clinical decision-making for the early treatment of AIS. Full article

Intraoperative radiotherapy (IORT) is a radiation technique that allows for the delivery of a high radiation dose to the target while preserving the surrounding structures, which can be displaced during the surgical procedure. An important limitation of this technique is the lack of real-time image guidance, which is one of the main achievements of modern radiation therapy because it allows for treatment optimization. IORT can be delivered by low-energy X-rays or by accelerated electrons. The present review describes the most relevant clinical applications for IORT and discusses the potential advantages of using artificial intelligence (AI) to overcome some of the current limitations of IORT. In recent decades, IORT has proven to be an effective treatment in several cancer types. In breast cancer, IORT can be used to deliver a single dose of radiation (partial breast irradiation) or as a boost in high-risk patients. In locally advanced rectal cancer, a single dose to the tumor bed can improve local control and prevent pelvic relapse in primary and recurrent tumors. In sarcomas, IORT enables the delivery of high doses, achieving good functional outcomes with low toxicity in tumors located in the retroperitoneum and extremities. In pancreatic cancer, IORT shows promising results in borderline resectable and unresectable cases. Ongoing technological advances are addressing current challenges in imaging and radiation planning, paving the way for personalized, image-guided IORT. Recent innovations such as CT- and MRI-equipped hybrid operating theaters allow for real-time imaging, which could be used for AI-assisted segmentation and planning. Moreover, the implementation of AI in terms of machine learning, deep learning, and radiomics can improve the interpretation of imaging, predict treatment outcomes, and optimize workflow efficiency. Full article

Background/Objectives: Human epidermal growth factor 2 (HER2) scoring is critical for modern breast cancer therapies, especially with emerging indications of antibody–drug conjugates for HER2-low tumors. However, inter-observer agreement remains limited in borderline cases. Automatic artificial intelligence-based scoring has the potential to improve diagnostic consistency and scalability. This study aimed to develop two transformer-based models for HER2 scoring of breast cancer whole-slide images (WSIs) and compare their performance. Methods: We adapted a large-scale foundation model (Virchow) and a lightweight model (TinyViT). Both were trained using patch-level annotations and integrated into a WSI scoring pipeline. Performance was evaluated on a clinical test set (n = 66), including clinical decision tasks and inference efficiency. Results: Both models achieved substantial agreement with pathologist reports (linear weighted kappa: 0.860 for Virchow, 0.825 for TinyViT). Virchow showed slightly higher WSI-level accuracy than TinyViT, whereas TinyViT reduced inference times by 60%. In three binary clinical tasks, both models demonstrated a diagnostic performance comparable to pathologists, particularly in identifying HER2-low tumors for antibody–drug conjugate (ADC) therapy. A continuous scoring framework demonstrated a strong correlation between the two models (Pearson’s r = 0.995) and aligned with human assessments. Conclusions: Both transformer-based artificial intelligence models achieved human-level accuracy for automated HER2 scoring with interpretable outputs. While the foundation model offers marginally higher accuracy, the lightweight model provides practical advantages for clinical deployment. In addition, continuous scoring may provide a more granular HER2 quantification, especially in borderline cases. This could support a new interpretive paradigm for HER2 assessment aligned with the evolving indications of ADC. Full article

This study introduces the concept of symmetry as a fundamental theoretical perspective for understanding the linguistic structure of cyberbullying texts. It posits that such texts often exhibit symmetry breaking between surface-level language forms and underlying semantic intent. This structural-semantic asymmetry increases the complexity of the recognition task and places higher demands on the semantic modeling capabilities of detection systems. With the rapid growth of social media, the covert and harmful nature of cyberbullying speech has become increasingly prominent, posing serious challenges to public opinion management and public safety. While mainstream approaches to cyberbullying detection—typically based on traditional deep learning models or pre-trained language models—have achieved some progress, they still struggle with low accuracy, poor generalization, and weak interpretability when handling implicit, semantically complex, or borderline expressions. To address these challenges, this paper proposes a cyberbullying detection method that combines LoRA-based fine-tuning with Small-Scale Hard-Sample Adaptive Training (S-HAT), leveraging a large language model framework based on Meta-Llama-3-8B-Instruct. The method employs prompt-based techniques to identify inference failures and integrates model-generated reasoning paths for lightweight fine-tuning. This enhances the model’s ability to capture and represent semantic asymmetry in cyberbullying texts. Experiments conducted on the ToxiCN dataset demonstrate that the S-HAT approach achieves a precision of 84.1% using only 24 hard samples—significantly outperforming baseline models such as BERT and RoBERTa. The proposed method not only improves recognition accuracy but also enhances model interpretability and deployment efficiency, offering a practical and intelligent solution for cyberbullying mitigation. Full article

Introduction: Previous neuroimaging studies have reported structural brain alterations and local network hyperexcitability in terms of increased slow-wave electroencephalography (EEG) activity in patients with borderline personality disorder (BPD). In particular, intermittent rhythmic delta and theta activity (IRDA/IRTA) has drawn attention in mental health contexts due to its links with metabolic imbalances, neuronal stress, and emotional dysregulation—processes that are highly pertinent to BPD. These functional disturbances may be reflected in corresponding structural brain changes. The current study investigated cortical thickness and subcortical volumes in BPD and examined their associations with IRDA/IRTA events per minute, symptom severity, and neuropsychological measures. Methods: Seventy female BPD patients and 36 age-matched female healthy controls (HC) were included (for clinical EEG comparisons even 72 patients were available). IRDA/IRTA rates were assessed using an automatic independent component analyses (ICA) approach. T1-weighted MRI data were obtained using a MAGNETOM Prisma 3T system and analyzed with FreeSurfer (version 7.2) for subcortical structures and CAT12 for cortical thickness and global volume measurements. Psychometric assessments included questionnaires such as Borderline Symptom List (BSL-23) and Inventory of Personality Organization (IPO). Neuropsychological performance was evaluated with the Test for Attentional Performance (TAP), Culture Fair Intelligence Test (CFT-20-R), and Verbal Learning and Memory Test (VLMT). Results: Between-group comparisons exhibited no significant increase in IRDA/IRTA rates or structural abnormalities between the BPD and HC group. However, within the BPD group, cortical thickness of the right isthmus of the cingulate gyrus negatively correlated with the IRDA/IRTA difference (after minus before hyperventilation, HV; p < 0.001). Furthermore, BPD symptom severity (BSL-23) and IPO scores positively correlated with the thickness of the right rostral anterior cingulate cortex (p < 0.001), and IPO scores were associated with the thickness of the right temporal pole (p < 0.001). Intrinsic alertness (TAP) significantly correlated with relative cerebellar volume (p = 0.01). Discussion: While no group-level structural abnormalities were observed, correlations between EEG slowing, BPD symptom severity, and alertness with cortical thickness and/or subcortical volumes suggest a potential role of the anterior cingulate cortex, temporal pole, and cerebellum in emotion regulation and cognitive functioning in BPD. Future research employing multimodal EEG-MRI approaches may provide deeper insights into the neural mechanisms underlying BPD and guide personalized therapeutic strategies. Full article

Borderline intellectual functioning (BIF) is characterized by an IQ typically ranging from 70 to 85, combined with deficits in adaptive functioning. Despite its prevalence, individuals with BIF are often excluded from diagnostic and support systems, which traditionally focus on strictly defined intellectual disabilities. This article presents a systematic review conducted across the ProQuest, WoS, SCOPUS, and EBSCOhost databases, aiming to develop a profile of the adaptive functioning in individuals with BIF. A total of 64 documents published from 2012 to the present were included, all of them addressing BIF and adaptive functioning skills, and quality was assessed using the SSAHS tool. The findings presented are synthesized according to conceptual, social, and practical domains and reveal that individuals with BIF experience widespread difficulties across the conceptual, social, and practical domains. Additionally, societal barriers, primarily limiting access to support services, persist. However, there are emerging resources aimed at supporting this population, such as legislative efforts to facilitate their integration into the labor market. The implications and limitations of the findings are discussed, highlighting the need to consider the adaptive functioning skills of individuals with BIF. Full article

Background/Objectives: Observational studies suggest a link between vascular calcification and dementia or cognitive decline, but the evidence is conflicting, and the underlying mechanisms are unclear. Here, we investigate the shared genetic and causal relationships of vascular calcification—coronary artery calcification (CAC) and abdominal aortic calcification (AAC)—with Alzheimer’s disease (AD), and five cognitive traits. Methods: We analyse large-scale genome-wide association studies (GWAS) summary statistics, using well-regarded methods, including linkage disequilibrium score regression (LDSC), Mendelian randomisation (MR), pairwise GWAS (GWAS-PW), and gene-based association analysis. Results: Our findings reveal a nominally significant positive genome-wide genetic correlation between CAC and AD, which becomes non-significant after excluding the APOE region. CAC and AAC demonstrate significant negative correlations with cognitive performance and educational attainment. MR found no causal association between CAC or AAC and AD or cognitive traits, except for a bidirectional borderline-significant association between AAC and fluid intelligence scores. Pairwise-GWAS analysis identifies no shared causal SNPs (posterior probability of association [PPA]3 < 0.5). However, we find pleiotropic loci (PPA4 > 0.9), particularly on chromosome 19, with gene association analyses revealing significant genes in shared regions, including APOE, TOMM40, NECTIN2, and APOC1. Moreover, we identify suggestively significant loci (PPA4 > 0.5) on chromosomes 1, 6, 7, 9 and 19, implicating pleiotropic genes, including NAV1, IPO9, PHACTR1, UFL1, FHL5, and FOCAD. Conclusions: Current findings reveal limited genetic correlation and no significant causal associations of CAC and AAC with AD or cognitive traits. However, significant pleiotropic loci, particularly at the APOE region, highlight the complex interplay between vascular calcification and neurodegenerative processes. Given APOE’s roles in lipid metabolism, neuroinflammation, and vascular integrity, its involvement may link vascular and neurodegenerative disorders, pointing to potential targets for further investigation. Full article

Background/Objectives: This study aimed to evaluate an artificial intelligence (AI)-enhanced electrocardiogram (ECG) screening protocol for improved accuracy, efficiency, and risk stratification across six sports: handball, football, athletics, weightlifting, judo, and karate. Methods: For each of the six sports, resting 12-lead ECGs from healthy children and junior athletes were analyzed using AI algorithms trained on annotated datasets. Parameters included the QTc intervals, PR intervals, and QRS duration. Statistical methods were used to examine each sport’s specific cardiovascular adaptations and classify cardiovascular risk predictions as low, moderate, or high risk. Results: The accuracy, sensitivity, specificity, and precision of the AI system were 97.87%, 75%, 98.3%, and 98%, respectively. Among the athletes, 94.54% were classified as low risk and 5.46% as moderate risk with AI because of borderline abnormalities like QTc prolongation or mild T-wave inversions. Sport-specific trends included increased QRS duration in weightlifters and low QTc intervals in endurance athletes. Conclusions: The statistical analyses and the AI-ECG screening protocol showed high precision and scalability for the proposed athlete cardiovascular health risk status stratification. Additional early detection research should be conducted further for diverse cohorts of individuals engaged in sports and explore other diagnostic methods that can help increase the effectiveness of screening. Full article

Borderline Ovarian Tumours (BOTs) are a heterogenous group of ovarian neoplasms which have increased mitotic activity but lack stromal invasion. We performed a narrative review of the literature, aiming to identify prognostic molecular biomarkers that can potentially be used for treatment personalisation. We identified and discussed BRAF/KRAS, Cancer Antigen 125 (Ca 125), Calprotectin, p16ink4a, and Microsatellite instability (MSI) as the most studied biomarkers related to BOTs. Overall, BRAF and KRAS mutations are associated with earlier-stage and favourable prognosis; KRASmt may indicate extraovarian disease in serous BOT (sBOT). Ca125, the only currently clinically used biomarker, can be assessed pre-operatively and has an established role in post-operative surveillance, especially when it is raised pre-operatively or a high potential for malignant transformation is suspected post-operatively. p16ink4a expression trends could also indicate the malignant transformation of the tumour. Calprotectin has an inferior specificity to Ca125 and is not yet established as a biomarker, whilst there is very limited evidence available for MSI. As new evidence is coming along with artificial intelligence platforms, these biomarkers can be integrated and used towards the development of a precision model for treatment stratification and counselling in women diagnosed with BOTs. Full article

Introduction: The concept of clinical readiness for practice among nursing students is yet to be analyzed, and there is a lack of empirical evidence on its usage among academics and clinicians. Methods: This concept analysis is anchored on a systematic literature review that adhered to the PRISMA guidelines and incorporated the eight iterative steps of Walker and Avant’s concept analysis method. This concept analysis method involved: (1) choosing a concept; (2) determining the objectives of the analysis; (3) identifying usages of the concept; (4) determining the defining attributes; (5) identifying a model case; (6) identifying other cases, including borderline, contrary, and related cases; (7) identifying antecedents and consequences; and (8) defining empirical references. The integrative thematic data synthesis method was adopted. Results: The concept of nursing students’ clinical readiness for practice is said to have four interrelated attributes. These attributes included (1) professional skills, (2) communication skills, (3) self-management skills, and (4) self-confidence. The two antecedents for nursing students’ clinical readiness to practice are (1) personal factors, including demographic characteristics, prior healthcare experience, income, and emotional intelligence; and (2) educational factors, including the clinical learning environment, clinical internship program, learning resource, and learning strategy. The consequence of clinical readiness for the practice of nursing students includes obtaining practice skills that can lead to more personal and job-related satisfactory outcomes. Conclusions: clinical readiness for practice in nursing encompasses the acquisition and integration of professional knowledge, skills, effective communication abilities, and self-management capabilities and the application of these competencies with confidence toward the provision of high-quality care to patients. Clinical Relevance: Understanding the components of clinical readiness is crucial for nursing educators, preceptors, and healthcare institutions to ensure that nursing students are adequately prepared for the challenges they will face in clinical practice. By recognizing the importance of professional knowledge, skills, communication, and self-management in clinical readiness, educators and training institutions can tailor their curricula, programs, and support systems to better prepare nursing students for the demands of real-world healthcare settings. This focus on clinical readiness ultimately delivers safe, effective, and compassionate patient care. Full article

Identifying and reducing customer churn have become a priority for financial institutions seeking to retain clients. Our research focuses on customer churn rate analysis using advanced machine learning (ML) techniques, leveraging a synthetic dataset sourced from the Kaggle platform. The dataset undergoes a preprocessing phase to select variables directly impacting customer churn behavior. SMOTETomek, a hybrid technique that combines oversampling of the minority class (churn) with SMOTE and the removal of noisy or borderline instances through Tomek links, is applied to balance the dataset and improve class separability. Two cutting-edge ML models are applied—random forest (RF) and the Light Gradient-Boosting Machine (LGBM) Classifier. To evaluate the effectiveness of these models, several key performance metrics are utilized, including precision, sensitivity, F1 score, accuracy, and Brier score, which helps assess the calibration of the predicted probabilities. A particular contribution of our research is on calibrating classification probabilities, as many ML models tend to produce uncalibrated probabilities due to the complexity of their internal mechanisms. Probability calibration techniques are employed to adjust the predicted probabilities, enhancing their reliability and interpretability. Furthermore, the Shapley Additive Explanations (SHAP) method, an explainable artificial intelligence (XAI) technique, is further implemented to increase the transparency and credibility of the model’s decision-making process. SHAP provides insights into the importance of individual features in predicting churn, providing knowledge to banking institutions for the development of personalized customer retention strategies. Full article

This project employs artificial intelligence, including machine learning and deep learning, to assess COVID-19 readmission risk in Malaysia. It offers tools to mitigate healthcare resource strain and enhance patient outcomes. This study outlines a methodology for classifying COVID-19 readmissions. It starts with dataset description and pre-processing, while the data balancing was computed through Random Oversampling, Borderline SMOTE, and Adaptive Synthetic Sampling. Nine machine learning and ten deep learning techniques are applied, with five-fold cross-validation for evaluation. Optuna is used for hyperparameter selection, while the consistency in training hyperparameters is maintained. Evaluation metrics encompass accuracy, AUC, and training/inference times. Results were based on stratified five-fold cross-validation and different data-balancing methods. Notably, CatBoost consistently excelled in accuracy and AUC across all tables. Using ROS, CatBoost achieved the highest accuracy (0.9882 ± 0.0020) with an AUC of 1.0000 ± 0.0000. CatBoost maintained its superiority in BSMOTE and ADASYN as well. Deep learning approaches performed well, with SAINT leading in ROS and TabNet leading in BSMOTE and ADASYN. Decision Tree ensembles like Random Forest and XGBoost consistently showed strong performance. Full article

Young people with borderline intelligence functioning (BIF) have intellectual functioning at the border between intellectual disability and those considered neurotypical. This population group is often underrepresented in social research, which makes it difficult to understand their experiences and needs. The research aims to understand the daily lives of young people with BIF to identify needs that society might not be aware of. The study was conducted with a sample of 30 young people. The ethnomethodological design was appropriate for the study of the routines and daily dynamics of these young people, which allowed the researchers to understand the experiences and meanings of the participants from their own perspective. The analysis was carried out in the context of the subject of Qualitative Research Tools in Social Work with fourth-year students, through participant observation, semi-structured interviews, and field diaries. Data analysis was performed using the Atlas.ti23 qualitative content analysis program. The findings suggest a strong dependence on family and social support; a daily life marked by challenges; and a search for autonomy, among many other aspects. Collaboration with the participants allowed the researchers to better understand their experiences and needs from reflexivity. Full article

Background: Neither radiological phenotypic characteristics nor reconstruction CT scan has been used to study the early anatomical disruption of the cranial bone in children with the so-called idiopathic type of West syndrome. Material and Methods: The basic diagnostic measures and the classical antiepileptic treatments were applied to these children in accordance with the conventional protocol of investigations and treatment for children with West syndrome. Boys from three unrelated families were given the diagnosis of the idiopathic type of West syndrome, aged 7, 10 and 12 years old. Parents underwent extensive clinical examinations. Three parents (age range of 28–41 year) were included in this study. All children showed a history of intellectual disabilities, cryptogenic epileptic spasms and fragmented hypsarrhythmia. These children and their parents were referred to our orthopedic departments because of variable skeletal deformities. Variable forms of skeletal deformities were the motive for the families to seek orthopedic advice. A constellation of flat foot, torticollis and early-onset osteoarthritis were observed by the family doctor. Apparently, and from the first clinical session in our practice, we felt that all these children are manifesting variable forms of abnormal craniofacial contour. Thereby, we immediately performed detailed cranial radiological phenotypic characterization of every affected child, as well as the siblings and parents, and all were enrolled in this study. All affected children underwent whole-exome sequence analysis. Results: The craniofacial phenotype of all children revealed apparent developmental anatomical disruption of the cranial bones. Palpation of the skull bones showed unusual palpable bony ridges along different sutural locations. A 7-year-old child showed abnormal bulging over the sagittal suture, associated with bilateral bony ridges over the squamosal sutures. AP skull radiograph of a 7-year-old boy with West syndrome showed facial asymmetry with early closure of the metopic suture, and other sutures seemed ill-defined. A 3D reconstruction CT scan of the skull showed early closure of the metopic suture. Another 3D reconstruction CT scan of the skull while the patient was in flexion showed early closure of the squamosal sutures, pressing the brain contents upward, causing the development of a prominent bulge at the top of the mid-sagittal suture. A reformatted 3D reconstruction CT scan confirmed the bilateral closure of the squamosal suture. Examination of the parents revealed a similar skull radiographic abnormality in his mother. A 3D reformatted frontal cranial CT of a 35-year-old mother showed early closure of the metopic and sagittal sutures, causing a mid-sagittal bony bulge. A 10-year-old boy showed an extremely narrow frontal area, facial asymmetry and a well palpable ridge over the lambdoid sutures. A 3D axial reconstruction CT scan of a 10-year-old boy with West syndrome illustrated the asymmetry of the posterior cranial bones along the lambdoid sutures. Interestingly, his 28-year-old mother has been a client at the department of spine surgery since she was 14 years old. A 3D reconstruction CT scan of the mother showed a noticeable bony ridge extending from the metopic suture upwards to involve the sagittal suture (red arrow heads). The black arrow shows a well demarcated bony ridge over the squamosal suture. A 3D reconstruction CT scan of the skull and spine showed the thick bony ridge of the metopic and the anterior sagittal as well as bilateral involvement of the squamosal, causing apparent anterior narrowing of the craniofacial contour. Note the lumbar scoliosis. A 12-year-old boy showed brachycephaly. A lateral skull radiograph of a 12-year-old boy with West syndrome showed premature sutural fusion, begetting an abnormal growth pattern, resulting in cranial deformity. The nature of the deformity depends on which sutures are involved, the time of onset and the sequence in which individual sutures fuse. In this child, brachycephalic secondary to craniosynostosis, which occurred because of bilateral early ossification of the coronal sutures, led to bi-coronal craniosynostosis. Thickened frontal bones and an ossified interclinoid ligament of the sella turcica were encountered. The lateral skull radiograph of a 38-year-old mother with a history of poor schooling achievements showed a very similar cranial contour of brachycephaly, thickening of the frontal bones and massive ossification of the clinoid ligament of the sella turcica. Maternal history revealed a history of multiple spontaneous miscarriages in the first trimester of more than five times. Investigating his parents revealed a brachycephalic mother with borderline intelligence. We affirm that the pattern of inheritance in the three boys was compatible with the X-linked recessive pattern of inheritance. Whole-exome sequencing showed non-definite phenotype/genotype correlation. Conclusions: The aim of this study was sixfold: firstly, to refute the common usage of the term idiopathic; secondly, we feel that it could be possible that West syndrome is a symptom complex rather than a separate diagnostic entity; thirdly, to further detect the genetic carrier, we explored the connection between the cranial bones in children with West syndrome with what has been clinically observed in their parents; fourthly, the early life anatomical disruptions of the cranial bones among these children seem to be heterogeneous; fifthly, it shows that the progressive deceleration in the development of this group of children is highly connected to the progressive closure of the cranial sutures; sixthly, we affirm that our findings are novel. Full article

(1) Background: Although the diagnostic criteria for massive hemorrhage with organ dysfunction, such as disseminated intravascular coagulation associated with delivery, have been empirically established based on clinical findings, strict logic has yet to be used to establish numerical criteria. (2) Methods: A dataset of 107 deliveries with >2000 mL of blood loss, among 13,368 deliveries, was obtained from nine national perinatal centers in Japan between 2020 and 2023. Twenty-three patients had fibrinogen levels <170 mg/dL, which is the initiation of coagulation system failure, according to our previous reports. Three of these patients had hematuria. We used six machine learning methods to identify the borderline criteria dividing the fibrinogen/fibrin/fibrinogen degradation product (FDP) planes, using 15 coagulation fibrinolytic factors. (3) Results: The boundaries of hematuria development on a two-dimensional plane of fibrinogen and FDP were obtained. A positive FDP–fibrinogen/3–60 (mg/dL) value indicates hematuria; otherwise, the case is nonhematuria, as demonstrated by the support vector machine method that seemed the most appropriate. (4) Conclusions: Using artificial intelligence, the borderline criterion was obtained, which divides the fibrinogen/FDP plane for patients with hematuria that could be considered organ dysfunction in massive hemorrhage during delivery; this method appears to be useful. Full article