Predictive Scoring Model Research Articles

Rheumatoid Arthritis: Automated Scoring of Radiographic Joint Damage

Rheumatoid arthritis is an autoimmune disease that causes joint damage due to inflammation in the soft tissue lining the joints known as the synovium. It is vital to identify joint damage as soon as possible to provide necessary treatment early and prevent further damage to the bone structures. Radiographs are often used to assess the extent of the joint damage. Currently, the scoring of joint damage from the radiograph takes expertise, effort, and time. Joint damage associated with rheumatoid arthritis is also not quantitated in clinical practice and subjective descriptors are used. In this work, a description of a pipeline of deep learning models to automatically identify and score rheumatoid arthritic joint damage from a radiographic image is provided. An automatic tool was built to produce scores with extremely high balanced accuracy within a couple of minutes and utilizing this would remove the subjectivity of the scores between human reviewers. Using a joint segmentation approach and training joint score prediction models with ordinal class encoding, under-sampling, and transfer learning, the joint wised ±1 balanced accuracies ranging from 91.51% to 97.30% were achieved. The ±1 balanced accuracy of the 4 models showed great potential in achieving industry-standard reliability.

An Observational Study to Develop a Predictive Model for Bacterial Pneumonia Diagnosis in Severe COVID-19 Patients-C19-PNEUMOSCORE.

In COVID-19 patients, antibiotics overuse is still an issue. A predictive scoring model for the diagnosis of bacterial pneumonia at intensive care unit (ICU) admission would be a useful stewardship tool. We performed a multicenter observational study including 331 COVID-19 patients requiring invasive mechanical ventilation at ICU admission; 179 patients with bacterial pneumonia; and 152 displaying negative lower-respiratory samplings. A multivariable logistic regression model was built to identify predictors of pulmonary co-infections, and a composite risk score was developed using β-coefficients. We identified seven variables as predictors of bacterial pneumonia: vaccination status (OR 7.01; 95% CI, 1.73-28.39); chronic kidney disease (OR 3.16; 95% CI, 1.15-8.71); pre-ICU hospital length of stay ≥ 5 days (OR 1.94; 95% CI, 1.11-3.4); neutrophils ≥ 9.41 × 109/L (OR 1.96; 95% CI, 1.16-3.30); procalcitonin ≥ 0.2 ng/mL (OR 5.09; 95% CI, 2.93-8.84); C-reactive protein ≥ 107.6 mg/L (OR 1.99; 95% CI, 1.15-3.46); and Brixia chest X-ray score ≥ 9 (OR 2.03; 95% CI, 1.19-3.45). A predictive score (C19-PNEUMOSCORE), ranging from 0 to 9, was obtained by assigning one point to each variable, except from procalcitonin and vaccine status, which gained two points each. At a cut-off of ≥3, the model exhibited a sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of 84.9%, 55.9%, 69.4%, 75.9%, and 71.6%, respectively. C19-PNEUMOSCORE may be an easy-to-use bedside composite tool for the early identification of severe COVID-19 patients with pulmonary bacterial co-infection at ICU admission. Its implementation may help clinicians to optimize antibiotics administration in this setting.

Pediatric risk to orthotopic heart transplant (PRO) score: Insights from United Network for Organ Sharing (UNOS) waitlist mortality findings.

Pediatric heart transplant candidates on the waitlist have the highest mortality rate among all solid organ transplants. A risk score incorporating a candidate's individual risk factors may better predict mortality on the waitlist and optimize organ allocation to the sickest of those awaiting transplant. Using the United Network for Organ Sharing (UNOS) database, we evaluated a total of 5542 patients aged 0-18 years old on the waitlist for a single, first time, heart transplant from January 2010 to June 2019. We performed a univariate analysis on two-thirds (N = 3705) of these patients to derive the factors most associated with waitlist mortality or delisting secondary to deterioration within 1 year. Those with a p <0.2 underwent a multivariate analysis and the resulting factors were used to build a prediction model using the Fine-Grey model analysis. This predictive scoring model was then validated on the remaining one-third of the patients (N = 1852). The Pediatric Risk to OHT (PRO) scoring model utilizes the following unique patient variables: blood type, diagnosis of congenital heart disease, weight, presence of ventilator support, presence of inotropic support, extracorporeal membrane oxygenation (ecmo) status, creatinine level, and region. A higher score indicates an increased risk of mortality. The PRO score had a predictive strength of 0.762 as measured by area under the ROC curve at 1 year. The PRO score is an improved predictive model with the potential to better assess mortality for patients awaiting heart transplant.

Predicting multi-vascular diseases in patients with coronary artery disease

Background: Because of its systemic nature, the occurrence of atherosclerosis in the coronary arteries can also indicate a risk for other vascular diseases. However, screening program targeted for all patients with coronary artery disease (CAD) is highly ineffective and no studies have assessed the risk factors for developing multi-vascular diseases in general. This study constructed a predictive model and scoring system to enable targeted screening for multi-vascular diseases in CAD patients. Methods: This cross-sectional study includes patients with CAD, as diagnosed during coronary angiography or percutaneous coronary intervention from March 2021 to December 2021. Coronary artery stenosis (CAS) and abdominal aortic aneurysm (AAA) were diagnosed using Doppler ultrasound while peripheral artery disease (PAD) was diagnosed based on ABI score. Multivariate logistic regression was conducted to construct the predictive model and risk scores. Validation was conducted using ROC analysis and Hosmer-Lemeshow test. Results: Multivariate analysis showed that ages of &gt;60 years (OR [95% CI] = 1.579 [1.153-2.164]), diabetes mellitus (OR = 1.412 [1.036-1.924]), cerebrovascular disease (OR = 3.656 [2.326-5.747]), and CAD3VD (OR = 1.960 [1.250-3.073]) increased the odds for multi-vascular disease. The model demonstrated good predictive capability (AUC = 0.659) and was well-calibrated (Hosmer-Lemeshow p = 0.379). Targeted screening for high-risk patients reduced the number needed to screen (NNS) from 6 in the general population to 3 and has a high specificity of 96.5% Conclusions: Targeted screening using clinical risk scores was able to decrease NNS with good predictive capability and high specificity

A new dyslipidemia-based scoring model to predict transplant-free survival in patients with hepatitis E-triggered acute-on-chronic liver failure

Background/AimsHepatitis E virus (HEV)-triggered acute-on-chronic liver failure (ACLF) has unacceptably high short-term mortality. However, it is unclear whether the existing predictive scoring models are applicable to evaluate the prognosis of HEV-triggered ACLF.MethodsWe screened datasets of patients with HEV-triggered ACLF from a regional tertiary hospital for infectious diseases in Shanghai, China, between January 2011 and January 2021. Clinical and laboratory parameters were recorded and compared to determine a variety of short-term mortality risk factors, which were used to develop and validate a new prognostic scoring model.ResultsOut of 4952 HEV-infected patients, 817 patients with underlying chronic liver disease were enrolled in this study. Among these, 371 patients with HEV-triggered ACLF were identified and allocated to the training set (n = 254) and test set (n = 117). The analysis revealed that hepatic encephalopathy (HE), ascites, triacylglycerol and apolipoprotein A (apoA) were associated with 90-day mortality (P < 0.05). Based on these significant indicators, we designed and calculated a new prognostic score = 0.632 × (ascites: no, 1 point; mild to moderate, 2 points; severe, 3 points) + 0.865 × (HE: no, 1 point; grade 1–2, 2 points; grade 3–4, 3 points) − 0.413 × triacylglycerol (mmol/L) − 2.171 × apoA (g/L). Compared to four well-known prognostic models (MELD score, CTP score, CLIF-C OFs and CLIF-C ACLFs), the new scoring model is more accurate, with the highest auROCs of 0.878 and 0.896, respectively, to predict 28- and 90-day transplantation-free survival from HEV-triggered ACLF. When our model was compared to COSSH ACLF IIs, there was no significant difference. The test data also demonstrated good concordance.ConclusionsThis study is one of the first to address the correlation between hepatitis E and serum lipids and provides a new simple and efficient prognostic scoring model for HEV-triggered ACLF.

Systematic Review of &lt;14 weeks risk score prediction models using maternal characteristics with and without biomarkers for the prediction of GDM

Systematic Review of &lt;14 weeks risk score prediction models using maternal characteristics with and without biomarkers for the prediction of GDM

#4803 ANCA RENAL RISK SCORE 2023: THE UPDATED AND REVISED ARRS

Abstract Background and Aims Reliable prediction tools are needed to improve prognostication and personalisation of treatment in anti-neutrophil cytoplasmic antibody (ANCA) glomerulonephritides (GN). We aimed to validate and update the ANCA Renal Risk Score (ARRS) prediction model. Method The ARRS working group collated a retrospective multicentre international longitudinal cohort from referral centres and registries across the globe to revise the ARRS in a validation and recalibration study. The primary endpoint was end stage kidney disease (ESKD) and patients were censored at last follow-up. Cox proportional hazards models were used to reweight risk factors and develop a modified scoring system. Kaplan-Meier estimates, Harrell's C statistics and calibration plots were used to assess model performance. Results Of a total of 1591 patients, 1439 were included in the final analyses (959 in the development cohort, 52% male, median age 64 years). The ARRS demonstrated a discrimination of C = 0.800, comparable to the original cohort. Updating the model found an additional useful cut-off for kidney function (K), and serum creatinine replaced glomerular filtration rate which provided higher reliability (K0: &amp;lt;250 μmol/l = 0 points, K1: 250-450 μmol/l = 4 points, K2: &amp;gt;450 μmol/l = 11 points). The risk points for the percentage of normal glomeruli (N) and interstitial fibrosis and tubular atrophy (T) were also reweighted (N0: &amp;gt;25% = 0 points, N1: 10-25% = 4, N2: &amp;lt; 10% = 7, T0: none, mild or &amp;lt; 25% = 0 points, T1: ≥ mild-moderate or ≥ 25% = 3 points). We created four risk groups based on the sum of points: low (0 – 4 points), moderate (5 – 11), high (12 – 18) and an additional very high-risk (21). The model discrimination was C = 0.831 and a supplemental continuous model was developed to supply a patient-specific annual risk. Three-year kidney survival was 96%, 79%, 54%, and 19%, respectively. The ARRS23 performed similarly well in the validation cohort with excellent calibration. Conclusion We demonstrated the out-of-sample validity of the ARRS and present here the modified and improved score to optimise prognostication and risk stratification for clinical practice and trials.

Does RSClin provide additional information over classic clinico-pathologic scores (PREDICT 2.1, Influence, CTS-5)? A TEAM Pathology substudy.

533 Background: Gene-expression profiling tests (e.g. Oncotype, Mammaprint, Prosignia, Breast Cancer Index and EndoPredict) are widely used in the care of patients with early-stage hormone positive breast cancer (node negative or 1-3 lymph nodes). However, these tests are resource intensive, and few studies have compared their value with either free and widely available clinico-pathologic risk calculators (e.g. PREDICT 2.1, INFLUENCE 2.0, and CTS-5) or tools that combine genomic testing with clinico-pathologic data (e.g. RSClin). The TEAM pathology substudy population was used to compare these different predicted model scores with outcomes. Methods: The TEAM pathology study consists of 3284 postmenopausal hormone positive breast cancer patients treated with either exemestane or tamoxifen followed by exemestane. Accrual was from 2001 to 2006. Genes comprising the multi-parametric Oncotype Dx were used to train signatures to create true assay results. Patient data was then used to calculate recurrence scores through various tools. This included clinico-pathologic models and their respective endpoints PREDICT 2.1 (overall survival at 5 years), INFLUENCE 2.0 (distant metastasis at 5 years), CTS-5 (distant recurrence risk at year 5-10), the purely genomic Oncotype Dx trained results (distant recurrence risk at 9 years), as well as the new clinico-pathologic RSClin (distant recurrence risk at 10 years). We compared the level of association between these predictive model scores using Spearman correlation coefficients. The prognostic ability of each model was contrasted for each outcome using Harrell’s C-statistic. Results: Results were available for CTS-5 (3022 patients), INFLUENCE 2.0 (3485 patients), ODx-trained (3825 patients), and RSClin (3029 patients). Correlation coefficients showed low correlation between Influence ( r= 0.25) and CTS-5 ( r= 0.17) with Oncotype-Dx trained results, and high correlation between RSClin ( r = 0.84) and Oncotype-Dx trained results. The concordance index was similar (0.65 to 0.68) for all models with distant metastasis-free survival as the outcome. Analysis is ongoing and further results will be available at the time of presentation. Conclusions: Other clinico-pathologic tools such as Influence 2.0 and CTS-5 have good prognostic ability when compared to Oncotype Dx-trained results and RSClin. [Table: see text]

The Addition of the Geriatric Nutritional Risk Index to the Prognostic Scoring Systems Did Not Improve Mortality Prediction in Trauma Patients in the Intensive Care Unit.

Malnutrition is prevalent among critically ill patients and has been associated with a poor prognosis. This study sought to determine whether the addition of a nutritional indicator to the various variables of prognostic scoring models can improve the prediction of mortality among trauma patients in the intensive care unit (ICU). This study's cohort included 1,126 trauma patients hospitalized in the ICU between January 1, 2018, and December 31, 2021. Two nutritional indicators, the prognostic nutrition index (PNI), a calculation based on the serum albumin concentration and peripheral blood lymphocyte count, and the geriatric nutritional risk index (GNRI), a calculation based on the serum albumin concentration and the ratio of current body weight to ideal body weight, were examined for their association with the mortality outcome. The significant nutritional indicator was served as an additional variable in prognostic scoring models of the Trauma and Injury Severity Score (TRISS), the Acute Physiology and Chronic Health Evaluation (APACHE II), and the mortality prediction models (MPM II) at admission, 24, 48, and 72 h in the mortality outcome prediction. The predictive performance was determined by the area under the receiver operating characteristic curve. Multivariate logistic regression revealed that GNRI (OR, 0.97; 95% CI, 0.96-0.99; p=0.007), but not PNI (OR, 0.99; 95% CI, 0.97-1.02; p=0.518), was independent risk factor for mortality. However, none of these predictive scoring models showed a significant improvement in prediction when the GNRI variable is incorporated. The addition of GNRI as a variable to the prognostic scoring models did not significantly enhance the performance of the predictors.

An Intelligent Boosting and Decision-Tree-Regression-Based Score Prediction (BDTR-SP) Method in the Reform of Tertiary Education Teaching

The reform of tertiary education teaching promotes teachers to adjust timely teaching plans based on students’ learning feedback in order to improve teaching performance. Thefore, learning score prediction is a key issue in process of the reform of tertiary education teaching. With the development of information and management technologies, a lot of teaching data are generated as the scale of online and offline education expands. However, a teacher or educator does not have a comprehensive dataset in practice, which challenges his/her ability to predict the students’ learning performance from the individual’s viewpoint. How to overcome the drawbacks of small samples is an open issue. To this end, it is desirable that an effective artificial intelligent tool is designed to help teachers or educators predict students’ scores well. We propose a boosting and decision-tree-regression-based score prediction (BDTR-SP) model, which relies on an ensemble learning structure with base learners of decision tree regression (DTR) to improve the prediction accuracy. Experiments on small samples are conducted to examine the important features that affect students’ scores. The results show that the proposed model has advantages over its peer in terms of prediction correctness. Moreover, the predicted results are consistent with the actual facts implied in the original dataset. The proposed BDTR-SP method aids teachers and students to predict students’ performance in the on-going courses in order to adjust the teaching and learning strategies, plans and practices in advance, enhancing the teaching and learning quality. Therefore, the integration of information technology and artificial intelligence into teaching and learning practices is able to push forward the reform of tertiary education teaching.

Immune-related lncRNAs pairs prognostic score model for prediction of survival in acute myeloid leukemia patients

Acute myeloid leukemia (AML) is one of the most common malignant and aggressive hematologic tumors, and risk stratification is indispensable to ensure proper treatment. But immune-related long noncoding RNAs (ir-lncRNAs) pairs prognostic risk models used to stratify AML have yet to be reported. In this study, we established a prognostic risk model based on eight ir-lncRNAs pairs using LASSO-penalized Cox regression analysis and successfully validated the model in an independent cohort. According to risk scores, patients were divided into a high-risk group and a low-risk group. High-risk patients presented more tumor mutation frequency and higher expression of human leukocyte antigen (HLA)-related genes and immune checkpoint molecules. Gene Set Enrichment Analysis (GSEA) indicated that the transforming growth factors β (TGFβ) pathway was activated in the high-risk group; meanwhile, we found that TGFβ1 mRNA levels were significantly elevated in AML patients and correlated with poor prognosis, which is closely related to drug resistance. Consistently, in vitro studies found that exogenous TGFβ1 can protect AML cells from chemotherapy-induced apoptosis. Collectively, we developed an ir-lncRNA prognostic model that helps predict the prognosis of AML patients and provides valuable information about their response to immune checkpoint inhibitors, and we found that increased TGFβ1 levels resulting in chemoresistance may be one of the leading causes of treatment failure in high-risk AML patients.

Development and validation of a CT radiomics and clinical feature model to predict omental metastases for locally advanced gastric cancer

“”We employed radiomics and clinical features to develop and validate a preoperative prediction model to estimate the omental metastases status of locally advanced gastric cancer (LAGC). A total of 460 patients (training cohort, n = 250; test cohort, n = 106; validation cohort, n = 104) with LAGC who were confirmed T3/T4 stage by postoperative pathology were continuously collected retrospectively, including clinical data and preoperative arterial phase computed tomography images (APCT). Dedicated radiomics prototype software was used to segment the lesions and extract features from the preoperative APCT images. The least absolute shrinkage and selection operator (LASSO) regression was used to select the extracted radiomics features, and a radiomics score model was constructed. Finally, a prediction model of omental metastases status and a nomogram were constructed combining the radiomics scores and selected clinical features. An area under the curve (AUC) of the receiver operating characteristic curve (ROC) was used to validate the capability of the prediction model and nomogram in the training cohort. Calibration curves and decision curve analysis (DCA) were used to evaluate the prediction model and nomogram. The prediction model was internally validated by the test cohort. In addition, 104 patients from another hospital's clinical and imaging data were gathered for external validation. In the training cohort, the combined prediction (CP) model (AUC 0.871, 95% CI 0.798–0.945) of the radiomics scores combined with the clinical features, compared with clinical features prediction (CFP) model (AUC 0.795, 95% CI 0.710–0.879) and radiomics scores prediction (RSP) model (AUC 0.805, 95% CI 0.730–0.879), had the better predictive ability. The Hosmer–Lemeshow test of the CP model showed that the prediction model did not deviate from the perfect fitting (p = 0.893). In the DCA, the clinical net benefit of the CP model was higher than that of the CFP model and RSP model. In the test and validation cohorts, the AUC values of the CP model were 0.836 (95% CI 0.726–0.945) and 0.779 (95% CI 0.634–0.923), respectively. The preoperative APCT-based clinical-radiomics nomogram showed good performance in predicting omental metastases status in LAGC, which may contribute to clinical decision-making.

Comparison of SCORE and SCORE 2 model for cardiovascular risk stratification in men and women from high risk European country

Abstract Funding Acknowledgements Type of funding sources: None. European Society of Cardiology has recently introduced new SCORE 2 algorhythm for cardiovascular risk assessment in apparently healthy individuals. However, there are limited data from high risk countries on CV risk stratification and potential gender differences in risk estimation using SCORE 2 model. The aim of the current study was to compare risk prediction using SCORE 2 and previous SCORE model and to assess the potential gender differences in risk estimation. Methods Our study included 1361 apparently healthy individuals (mean age 56±9 years, 61% women) without known CV disease or risk factors aged 40 to 70 years old. Data on CV risk factors were prospectively collected in 20 primary care centers throughout Serbia from January 2020 to December 2020. CV risk was assessed both by SCORE and SCORE 2 model. Based on CV risk profile individuals were stratified in three categories: low to intermediate, high and very high risk category. Results In overall population using SCORE model 57.6% of individuals were in low-intermediate group, 20.9% were classified as high risk group and 21.5% as very high risk group. By the use of SCORE 2 model 1046/1361 (77%) individuals were shifted to higher risk category (Figure). In comparison to men women had lower both systolic (139 ± 19 mmHg vs 144 ± 18 mmHg in men p &amp;lt; 0.001) and diastolic (84 ± 9 mmHg vs 88 ± 10mmHg in men, p &amp;lt; 0.001 )blood pressure, lower BMI (26.4 ± 0.5 kg/m2 vs 27.4 ± 5.0 kg/m2 in men, p= 0.036) and higher values of HDL cholesterol (1.5 ± 0.5 mmol/L vs. 1.3 ± 0.4 mmol/L in men, p &amp;lt; 0.001). Smoking was more prevalent in men (49%) than in women (42%), p=0.04. There were no gender differences in age (56 ± 9 years for both genders) and non-HDL cholesterol ( 4.86 ± 1.24 mmol in men vs. 4.75 ± 1.23 mmol/l in women, p= 0.100). More women (85.9%) than men (62.6%) were shifted in higher risk category using SCORE 2 in comparison to SCORE (p &amp;lt; 0.001) (Figure). Conclusion By the use of SCORE 2 instead of SCORE prediction model for CV risk stratification in contemporary high risk European population significant proportion of apparently healthy individuals are shifted to higher risk category. In spite of having better CV risk factors profile, more women than men were shifted in higher risk category.

Survival benefit of postoperative radiotherapy for pediatric patients with primary intracranial atypical teratoid/rhabdoid tumors: Propensity score analysis and prediction model construction in a multi-registry based cohort

BackgroundThe role of postoperative radiotherapy (PORT) in children with primary intracranial atypical teratoid/rhabdoid tumor (AT/RT) remains controversial, and real-world data with large sample size are still lacking. This study aims to estimate the survival benefit of PORT in pediatric patients with resected AT/RT. MethodsUsing Seer database, we collected 246 eligible intracranial AT/RT patients diagnosed between 2000 and 2016 for our analysis. Propensity score matching (PSM) analysis was employed to minimize selection bias for evaluation of the PORT efficacy. Multivariate Cox regression was conducted to determine the factors related to the outcome. Interaction tests were further performed between PORT and the prognostic variables. After identifying the significant prognostic factors, we further developed a novel prediction model to predict the life expectancy of these patients, as well as the potential benefit from PORT. ResultsWe found that PORT was significantly related to the improved survival after adjusting for other prognosticators in both the entire and PSM-matched cohort. Significant interactions of PORT with age at diagnosis and tumor extension were also observed. On basis of the prognostic indictors identified by L1-penalized lasso Cox regression analysis, a novel nomogram model was successfully established and externally validated. ConclusionOur study indicated that PORT was significantly associated with the improved survival in pediatric AT/RT patients, and the greater survival benefit from PORT could be achieved in patients <3 years old or with locoregional tumors. The novel prediction model was developed to provide help in clinical practice and in the design of related trials.

The Gluteus-Score-7 Predicts the Likelihood of Both Clinical Success and Failure Following Surgical Repair of the Hip Gluteus Medius and/or Minimus

To identify patient preoperative history, examination, and imaging characteristics that increase the risk of postoperative failure of gluteus medius/minimus repair, and to develop a decision-making aid predictive of clinical outcomes for patients undergoing gluteus medius/minimus repair. Patients from 2012 to 2020 at a single institution undergoing gluteus medius/minimus repair with minimum 2-year follow-up were identified. MRIs were graded according to the "three-grade" classification system: grade 1: partial-thickness tear, grade 2: full-thickness tears with <2 cm of retraction, grade 3: full-thickness tears with ≥2 cm retraction. Failure was defined as undergoing revision within 2 years postoperatively or not achieving both a cohort-calculated minimal clinically important difference (MCID) and responding "no" to patient acceptable symptom state (PASS). Inversely, success was defined as reaching both an MCID and responding "yes" to PASS. Predictors of failure were verified on logistic regression and a predictive scoring model, the Gluteus-Score-7, was generated to guide treatment-decision making. In total, 30 of 142 patients (21.1%) were clinical failures at mean ± SD follow-up of 27.0 ± 5.2 months. Preoperative smoking (odds ratio [OR], 3.0; 95% confidence interval [CI], 1.0-8.4; P= .041), lower back pain (OR, 2.8; 95% CI, 1.1-7.3; P= .038), presence of a limp or Trendelenburg gait (OR, 3.8; 95% CI, 1.5-10.2; P= .006), history of psychiatric diagnosis (OR, 3.7; 95% CI, 1.3-10.8; P= .014), and increased MRI classification grades (P ≤ .042) were independent predictors of failure. The Gluteus-Score-7 was generated with each history/examination predictor assigned 1 point and MRI classes assigned corresponding 1-3 points (min 1, max 7 score). A score of ≥4/7 points was associated with risk of failure and a score ≤2/7 points was associated with clinical success. Independent risk factors for revision or not achieving either MCID or PASS after gluteus medius and/or minimus tendon repair include smoking, preoperative lower back pain, psychiatric history, Trendelenburg gait, and full-thickness tears, especially tears with ≥2 cm retraction. The Gluteus-Score-7 tool incorporating these factors can identify patients at risk of both surgical treatment failure and success, which may be useful for clinical decision-making. Level IV, prognostic case series.

Validation of the SACOV-19 score for identifying patients at risk of complicated or more severe COVID-19: a prospective study

PurposeIdentification of patients at risk of complicated or more severe COVID-19 is of pivotal importance, since these patients might require monitoring, antiviral treatment, and hospitalization. In this study, we prospectively evaluated the SACOV-19 score for its ability to predict complicated or more severe COVID-19.MethodsIn this prospective multicenter study, we included 124 adult patients with acute COVID-19 in three German hospitals, who were diagnosed in an early, uncomplicated stage of COVID-19 within 72 h of inclusion. We determined the SACOV-19 score at baseline and performed a follow-up at 30 days.ResultsThe SACOV-19 score’s AUC was 0.816. At a cutoff of > 3, it predicted deterioration to complicated or more severe COVID-19 with a sensitivity of 94% and a specificity of 55%. It performed significantly better in predicting complicated COVID-19 than the random tree-based SACOV-19 predictive model, the CURB-65, 4C mortality, or qCSI scores.ConclusionThe SACOV-19 score is a feasible tool to aid decision making in acute COVID-19.

Identification and in vitro and in vivo validation of the key role of GSDME in pyroptosis-related genes signature in hepatocellular carcinoma

We used pyroptosis-related genes to establish a risk–score model for prognostic prediction of liver hepatocellular carcinoma (LIHC) patients. A total of 52 pyroptosis-associated genes were identified. Then, data for 374 LIHC patients and 50 normal individuals were acquired from the TCGA database. Through gene expression analyses, differentially expressed genes (DEGs) were determined. The 13 pyroptosis-related genes (PRGs) confirmed as potential prognostic factors through univariate Cox regression analysis were entered into Lasso and multivariate Cox regression to build a PRGs prognostic signature, containing four PRGs (BAK1, GSDME, NLRP6, and NOD2) determined as independent prognostic factors. mRNA levels were evaluated by qRT-PCR, while overall survival (OS) rates were assessed by the Kaplan–Meier method. Enrichment analyses were done to establish the mechanisms associated with differential survival status of LIHC patients from a tumor immunology perspective. Additionally, a risk score determined by the prognostic model could divide LIHC patients into low- and high-risk groups using median risk score as cut-off. A prognostic nomogram, derived from the prognostic model and integrating clinical characteristics of patients, was constructed. The prognostic function of the model was also validated using GEO, ICGC cohorts, and online databases Kaplan–Meier Plotter. Small interfering RNA-mediated knockdown of GSDME, as well as lentivirus-mediated GSDME knockdown, were performed to validate that knockdown of GSDME markedly suppressed growth of HCC cells both in vivo and in vitro. Collectively, our study demonstrated a PRGs prognostic signature that had great clinical value in prognosis assessment.

Two Clinical Prediction Tools to Inform Rapid Tuberculosis Treatment Decision-making in Children.

In the absence of bacteriologic confirmation to diagnose tuberculosis (TB) in children, it is suggested that treatment should be initiated when sufficient clinical evidence of disease is available. However, it is unclear what clinical evidence is sufficient to make this decision. To identify children who would benefit from rapid initiation of TB treatment, we developed 2 clinical prediction tools. We conducted a secondary analysis of a prospective intensified TB patient-finding intervention conducted in Pakistan in 2014-2016. TB disease was determined through either bacteriologic confirmation or a clinical diagnosis. We derived 2 tools: 1 uses classification and regression tree (CART) analysis to develop decision trees, while the second uses multivariable logistic regression to calculate a risk score. Of the 5162 and 5074 children included in the CART and prediction score, respectively, 1417 (27.5%) and 1365 (26.9%) were eligible for TB treatment. CART identified abnormal chest radiographs and family history of TB as the most important predictors (area under the receiver operating characteristic curve [AUC], 0.949). The final prediction score model included age group (0-4, 5-9, 10-14), weight <5th percentile, cough, fever, weight loss, chest radiograph suggestive of TB disease, and family history of TB; the identified best cutoff score was 9 (AUC, 0.985%). Use of clinical evidence was sufficient to accurately identify children who would benefit from treatment initiation. Our tools performed well compared with existing algorithms, though these results need to be externally validated before operationalization.

Development of a Bispectral index score prediction model based on an interpretable deep learning algorithm

BackgroundProper maintenance of hypnosis is crucial for ensuring the safety of patients undergoing surgery. Accordingly, indicators, such as the Bispectral index (BIS), have been developed to monitor hypnotic levels. However, the black-box nature of the algorithm coupled with the hardware makes it challenging to understand the underlying mechanisms of the algorithms and integrate them with other monitoring systems, thereby limiting their use. ObjectiveWe propose an interpretable deep learning model that forecasts BIS values 25 s in advance using 30 s electroencephalogram (EEG) data. Material and methodsThe proposed model utilized EEG data as a predictor, which is then decomposed into amplitude and phase components using fast Fourier Transform. An attention mechanism was applied to interpret the importance of these components in predicting BIS. The predictability of the model was evaluated on both regression and binary classification tasks, where the former involved predicting a continuous BIS value, and the latter involved classifying a dichotomous status at a BIS value of 60. To evaluate the interpretability of the model, we analyzed the attention values expressed in the amplitude and phase components according to five ranges of BIS values. The proposed model was trained and evaluated using datasets collected from two separate medical institutions. Results and conclusionThe proposed model achieved excellent performance on both the internal and external validation datasets. The model achieved a root-mean-square error of 6.614 for the regression task, and an area under the receiver operating characteristic curve of 0.937 for the binary classification task. Interpretability analysis provided insight into the relationship between EEG frequency components and BIS values. Specifically, the attention mechanism revealed that higher BIS values were associated with increased amplitude attention values in high-frequency bands and increased phase attention values in various frequency bands. This finding is expected to facilitate a more profound understanding of the BIS prediction mechanism, thereby contributing to the advancement of anesthesia technologies.

Development and validation of early death risk score model for emergency status prediction in very severe aplastic anemia.

This study developed and validated the Early Death Risk Score Model for early identification of emergency patients with very severe aplastic anemia (VSAA). All 377 patients with VSAA receiving first-line immunosuppressive therapy (IST) were categorized into training (n=252) and validation (n=125) cohorts. In the training cohort, age >24 years, absolute neutrophil count ≤0.015×109/L, serum ferritin >900ng/mL and times of fever before IST >1 time were significantly associated with early death. Covariates were assigned scores and categorized as: low (score 0-4), medium (score 5-7) and high (score ≥8) risk. Early death rate was significantly different between risk groups and the validation cohort results were consistent with those of the training cohort. The area under the receiver operating characteristic curve for the model was 0.835 (0.734,0.936) in the training cohort and 0.862 (0.730,0.994) in the validation cohort. The calibration plots showed high agreement, and decision curve analysis showed good benefit in clinical applications. The VSAA Early Death Risk Score Model can help with early identification of emergency VSAA and optimize treatment strategies. Emergency VSAA with high risk is associated with high early death rate, and alternative donor hematopoietic stem cell transplantation could be a better treatment than IST even without HLA-matching.