Science Progress最新文献

ObjectiveIn older adults with Type 2 diabetes mellitus (T2DM), the risk of delirium is significantly increased, driven by neuropathological alterations stemming from chronic insulin resistance. We utilized artificial intelligence and geriatric electronic health records to create an interpretable online machine-learning algorithm for predicting delirium risk. This tool facilitates prompt identification of high-risk elderly T2DM patients, enabling optimized interventions and improved clinical outcomes.MethodsThis retrospective cohort study identified older adults with T2DM using International Classification of Diseases (ICD) codes, with delirium defined by the Confusion Assessment Method for the intensive care unit (CAM-ICU). We extracted baseline demographics, vital signs, laboratory measurements, comorbidities and clinical severity scores. Candidate predictors for eight machine-learning algorithms were selected using least absolute shrinkage and selection operator regression and the Boruta method. Discrimination was assessed using accuracy, sensitivity, specificity and the F1 score. The final model was interpreted using SHapley Additive exPlanations (SHAP) and deployed as an online risk calculator.ResultsIntegrating dual feature selection methods identified 14 key predictors and the gradient boosting machine (GBM) model accurately predicted delirium risk in elderly patients with T2DM, demonstrating strong discriminatory performance with robust calibration in both internal and external validation. SHAP analysis highlighted the Glasgow Coma Scale, ICU length of stay and Sequential Organ Failure Assessment score as the predominant contributors to model predictions. The model was successfully deployed as an accessible online tool and the accompanying web-based calculator enables rapid, personalized risk assessment to support early intervention in ICU settings.ConclusionsThe GBM model showed strong performance in predicting delirium risk among elderly patients with T2DM, supporting clinically meaningful risk stratification. The accompanying web-based calculator enables rapid, individualized bedside assessment and may facilitate early identification of high-risk patients and timely intervention in ICU settings.

ObjectiveDespite advances in prevention, cervical cancer remains a serious global health issue. Concurrent chemoradiation is the standard treatment for locally advanced squamous cell carcinoma, yet 20-30% of patients develop persistent cervical cancer due to incomplete response, resulting in poor outcomes. This study aims to develop a predictive model for persistent cervical cancer in patients with locally advanced cervical squamous cell carcinoma following concurrent chemoradiation therapy, leveraging pretreatment multisequence magnetic resonance imaging data and advanced deep learning techniques.MethodsThis retrospective study included 259 patients with locally advanced cervical squamous cell carcinoma who underwent concurrent chemoradiation therapy at two centres. Four magnetic resonance imaging sequences were used to generate 2.5D data. A deep learning model incorporating Crossformer was developed and compared with radiomics and clinical models. Model performance was evaluated using receiver operating characteristic curves, calibration curves, and decision curve analysis.ResultsCrossFormer model outperformed the traditional convolutional neural network models in slice-level analysis across all cohorts, achieving an area under the curve of 0.775 in the test cohorts. The deep learning model achieved high predictive accuracy, with area under the curves of 0.884, 0.833, and 0.814 in the training, validation, and test cohorts, respectively, outperforming both the clinical and radiomics models. Combining clinical features with the deep learning model further improved performance, yielding area under the curves of 0.914, 0.868, and 0.839 in the respective cohorts.ConclusionThe developed model, utilizing 2.5D multi-sequence magnetic resonance imaging data and the deep learning technology that incorporated Crossformer, demonstrated strong predictive performance for persistent cervical cancer in patients with locally advanced cervical squamous cell carcinoma following concurrent chemoradiation therapy. This approach offers a promising and clinically applicable tool for treatment decision-making.

Waardenburg syndrome (WS) is a rare genetic disorder characterized by congenital sensorineural hearing loss and pigmentary abnormalities, accounting for 2-5% of congenital deafness. While molecular testing is the diagnostic gold standard, clinical recognition remains crucial in low-resource or conflict-affected environments where specialized services are unavailable. We report a Syrian male in his early 20s who presented to the otorhinolaryngology clinic seeking exemption from military service, citing long-standing right-sided hearing loss. The patient and his family had never pursued medical evaluation for his pigmentary features or hearing problem. Examination revealed a white forelock, heterochromia iridis, synophrys, broad nasal root, and dystopia canthorum (W Index 2.2). Pure-tone audiometry demonstrated severe unilateral sensorineural hearing loss. Systematic neurological, ophthalmological, and musculoskeletal assessments were normal. Due to the lack of access to genetic testing, a clinical diagnosis of WS type I was made, and the patient was referred for genetic counseling. This case highlights diagnostic challenges in conflict-affected, resource-limited settings. Despite striking phenotypic features, the patient remained undiagnosed until adulthood. Missed opportunities included the absence of childhood hearing screening, delayed recognition of pigmentary signs, and a lack of educational or psychosocial support. Literature indicates that phenotypic diagnosis is reliable when multiple major criteria are present, yet diagnostic delays significantly affect quality of life. This report underscores the importance of timely recognition of WS in low-resource contexts. Strengthening primary care awareness, implementing basic audiological and pigmentary screening, and integrating psychosocial support may help mitigate diagnostic delays and improve long-term outcomes for patients with rare genetic disorders.

Guillain-Barré syndrome (GBS) is a life-threatening acute paralytic neuropathy characterized by rapidly progressive limb weakness and bilateral cranial nerve involvement. We report an early 70s male with no relevant medical history diagnosed with anti-GM3 IgG-positive GBS, who developed unilateral oculomotor nerve palsy and autonomic dysfunction during intravenous immunoglobulin (IVIG) therapy, with spontaneous symptom resolution following IVIG completion. Six-month follow-up assessments confirmed complete remission. This case reinforces the therapeutic value of IVIG in GBS patients with specific autoantibody profiles, even when complicated by transient neurological deterioration during treatment.

Current guidelines recommend VATS as the first-line treatment for stage II empyema. Patients with concomitant cardiorespiratory failure are considered ineligible for guideline-recommended management. Alternative treatment options are needed. This case report pertains to an obese man in his early 60 s with respiratory failure and decompensated heart failure. He developed an emphyema, which was refractory to the standard antibiotic therapy, thoracocentesis and other chest drainage. As patient was disqualified from surgical treatment, new approach was implemented. Minimally invasive approach of introducing a single-use bronchoscope through an existing chest drain (FOB-VAMT) was performed with support of non-invasive ventilation and analgosedation. Patient's condition improved significantly after the procedure. He remains under regular follow-up at the Pulmonology Outpatient Clinic 2 years after implemented treatment. This case reports that FOB-VAMT may provide an effective therapeutic option for patients disqualified from standard VATS and endotracheal intubation.

ObjectiveThis study aims to systematically review and synthesize the studies on the application of machine learning for classifying infant cry types, identifying pathological cries, and evaluating the accuracy of infant cry recognition.MethodsThis review followed the PRISMA guidelines and was registered in PROSPERO (CRD42024600969). The literature search was conducted on four data sources: PubMed, CINAHL, Embase, and IEEE Xplore. The included studies focused on machine learning-based classification of infants' needs cries or pathological cries. These were published in English between January 1, 2014 and October 31, 2024. Study quality was assessed using the QUADAS-2 tool.ResultsOf 919 studies were identified, 17 were included in the final synthesis. Machine learning can classify infant cries into two main types: infant needs' cries and pathological cries, with some studies addressing both. Needs-related cries comprised nine subtypes, while pathological cries included six subtypes. Classification accuracy varied by machine learning classifier and the features used, ranging from 44.5% to 99.82%. The highest accuracy for infant needs' cries was hunger and pain cries at 99.82% using a Gaussian mixture model (GMM) classifier with constant-Q cepstral coefficients features. For pathological cries, the highest accuracy was for detecting deafness (99.42% to 99.82%), using a genetic selection of Fuzzy Model and a GMM classifier.ConclusionsMachine learning shows strong potential for accurately classifying infant cries and detecting pathologies. Future research should prioritize developing diverse cry datasets to improve model generalizability, evaluating performance in real-world settings, and integrating cry analysis with physiological signals to enhance diagnostic accuracy.

ObjectiveThis study aimed to evaluate the associations of the Chinese Visceral Adiposity Index (CVAI), age, and clinic systolic blood pressure (SBP) with left ventricular hypertrophy (LVH) in postmenopausal women with primary hypertension.MethodsWe conducted a retrospective case-control study including 501 postmenopausal women hospitalized with primary hypertension between January and December 2023 at the Department of Hypertension at the Fifth Affiliated Hospital of Xinjiang Medical University. Participants were divided into an LVH group (cases, n=86) and a non-LVH group (controls, n=415) based on the left ventricular mass index (LVMI). Clinical data, biochemical parameters, echocardiographic results, and the CVAI were collected. Logistic regression and receiver operating characteristic (ROC) curve analyses were performed to determine the association between LVH and the outcome.ResultsCompared to the non-LVH group, the LVH group exhibited significantly higher age, clinical SBP, proportion of coronary artery disease, blood urea nitrogen (BUN) levels, CVAI, and cystatin C levels (all P<0.05). Left ventricular geometric patterns also differed significantly between the two groups (P < 0.001). Multivariate logistic regression analysis identified CVAI (odds ratio [OR] =1.025, 95% confidence interval [CI] = 1.013-1.038), age (OR=1.045, 95% CI = 1.013-1.079), and clinical SBP (OR=1.020, 95% CI = 1.007-1.033) as independent risk factors for LVH. Multinomial logistic regression showed that CVAI was associated with concentric (OR = 1.026, 95% CI = 1.008-1.044) and eccentric hypertrophy (OR = 1.026, 95% CI = 1.011-1.041). The receiver operating characteristic (ROC) curve showed an area under the curve (AUC) of 0.702 for CVAI alone, with a sensitivity of 83.7% and a specificity of 48%. The discriminative performance of CVAI was significantly better than that of the ventricular artery index (VAI; 0.551) and left anterior portal (LAP; 0.575). The combined discrimination using all three factors (CVAI, age, and clinical SBP) yielded an AUC of 0.751.ConclusionCVAI, age, and clinical SBP are independent risk factors for LVH in postmenopausal women with primary hypertension.

ObjectiveTo evaluate the prognostic value of the combined neutrophil-to-lymphocyte ratio (NLR) and monocyte-to-lymphocyte ratio (MLR) for adverse outcomes in Taiwanese patients with acute coronary syndrome (ACS).MethodsThis retrospective, single-center, observational cohort study analyzed 653 ACS patients from the Tri-Service General Hospital-coronary artery disease registry (Jan 2016-Aug 2018). NLR and MLR were calculated at presentation. The primary endpoint was a composite of in-hospital mortality and stroke. Optimal cut-off values were determined using receiver operating characteristic curve analysis (NLR > 5.22, MLR > 0.76). Univariable and multivariable logistic regression were used to assess the association of both elevated NLR/MLR with the primary endpoint.ResultsPatients with both elevated NLR and MLR (NLR > 5.22 and MLR > 0.76) demonstrated a significantly higher risk of the primary endpoint compared to the group with no elevation (OR = 7.16) in univariable analysis. This group also experienced longer hospital stays and higher non-cardiovascular mortality. However, this association became non-significant after multivariable adjustment for confounders.ConclusionsConcurrent elevation of both NLR and MLR at presentation is associated with increased in-hospital morbidity and prolonged hospitalization in ACS patients, although this did not remain an independent predictor in multivariable analysis. Ongoing statin therapy demonstrated a protective association in this cohort. These readily available markers may facilitate early risk stratification.

ObjectiveThe Agatston score quantifies coronary artery calcium (CAC) but shows limited specificity for obstructive coronary artery disease (CAD) with extensive calcification. The Naples Prognostic Score (NPS) integrates inflammatory and nutritional parameters. This study evaluated NPS performance alongside Agatston scoring.MethodsThis is a retrospective, single-center study that analyzed 426 stable angina patients undergoing coronary CT angiography and invasive angiography at İzmir City Hospital in 2024. Agatston and NPS scores were calculated from routine parameters. Obstructive CAD was defined as significant lesions. Performance was assessed using ROC analysis and multivariate regression.ResultsAmong 426 patients (age 61.7 ± 9.5 years, 63.4% male), 277 (65.0%) had obstructive disease. Obstructive CAD patients showed lower HDL (P = .005), higher glucose (P = .008), and elevated HbA1c (P < .001). Agatston score showed moderate accuracy (area under the curve (AUC) 0.644), NPS limited performance (AUC 0.554). In Agatston ≥1000 subgroup (n = 43), NPS achieved AUC 0.680, combined model AUC 0.725. NPS was an independent predictor (OR 1.22, P = .006). Correlation was negligible (ρ=0.048).ConclusionsMetabolic markers (HDL, glucose, HbA1c) are more discriminative than inflammatory markers for obstructive CAD. NPS provides complementary value in extensive calcification (Agatston ≥1000), capturing distinct pathophysiologic domains.

ObjectiveDietary fiber and physical activity are independent determinants of blood pressure. However, their interaction regarding hypertension remains poorly understood. This study examined the association between dietary fiber intake and hypertension. It also aimed to determine if this relationship is modified by physical activity in a representative sample of U.S. adults.MethodsThis cross-sectional study used data from 26,556 U.S. adults (National Health and Nutrition Examination Survey 2007-2018). Dietary fiber intake was assessed via two 24-h recalls and categorized into quartiles. Hypertension was defined by measured blood pressure, self-reported diagnosis, or antihypertensive medication use. Physical activity was quantified and divided into three levels. Missing covariate data were addressed using multiple imputations. Survey-weighted multivariable logistic regression was used to examine the association. A multiplicative interaction term was introduced to assess effect modification by physical activity. An XGBoost was also developed and interpreted using SHapley Additive exPlanation to explore nonlinear associations.ResultsIn the fully adjusted model, lower dietary fiber intake was associated with increased odds of hypertension (OR for Q3 vs. Q4: 1.25; 95% CI [1.08, 1.44]). This relationship showed a potential interaction with physical activity (p for interaction = .067). Stratified analysis showed no association in the low or moderate physical activity groups. Among individuals with high physical activity, lower fiber intake was associated with significantly higher odds of hypertension (OR for Q1 vs. Q4: 1.50; 95% CI [1.10, 2.04]). The machine learning model confirmed that higher fiber intake was protectively associated with hypertension, an effect most prominent in physically active individuals.ConclusionThe protective association of dietary fiber with hypertension was observed primarily in highly active adults, but not in less active individuals. These findings highlight the importance of the diet-activity interplay for hypertension prevention.