A COVID-19 specific multiparametric and ECG-based score for the prediction of in-hospital mortality: ELCOVID score

We aimed to develop and validate a COVID-19 specific scoring system, also including some ECG features, to predict all-cause in-hospital mortality at admission. Patients were retrieved from the ELCOVID study (ClinicalTrials.gov identifier: NCT04367129), a prospective, multicenter Italian study enrolling COVID-19 patients between May to September 2020. For the model validation, we randomly selected two-thirds of participants to create a derivation dataset and we used the remaining one-third of participants as the validation set. Over the study period, 1014 hospitalized COVID-19 patients (mean age 74 years, 61% males) met the inclusion criteria and were included in this analysis. During a median follow-up of 12 (IQR 7-22) days, 359 (35%) patients died. Age (HR 2.25 [95%CI 1.72-2.94], p < 0.001), delirium (HR 2.03 [2.14-3.61], p = 0.012), platelets (HR 0.91 [0.83-0.98], p = 0.018), D-dimer level (HR 1.18 [1.01-1.31], p = 0.002), signs of right ventricular strain (RVS) (HR 1.47 [1.02-2.13], p = 0.039) and ECG signs of previous myocardial necrosis (HR 2.28 [1.23-4.21], p = 0.009) were independently associated to in-hospital all-cause mortality. The derived risk-scoring system, namely EL COVID score, showed a moderate discriminatory capacity and good calibration. A cut-off score of ≥ 4 had a sensitivity of 78.4% and 65.2% specificity in predicting all-cause in-hospital mortality. ELCOVID score represents a valid, reliable, sensitive, and inexpensive scoring system that can be used for the prognostication of COVID-19 patients at admission and may allow the earlier identification of patients having a higher mortality risk who may be benefit from more aggressive treatments and closer monitoring.

Incidence and risk factors for febrile neutropenia of patients with diffuse large B-cell lymphoma receiving R-CHOP-21 in China

OBJECTIVE

Febrile neutropenia (FN) is a serious complication of patients with diffuse large B-cell lymphoma (DLBCL) receiving R-CHOP-21. The prophylactic use of granulocyte colony-stimulating factors (G-CSFs) can significantly reduce the risk of FN. International guidelines recommend G-CSFs for patients receiving chemotherapy with FN risk of 20% or 10 to 20% with defined risk factors. However, there are few studies on the incidence and risk factors of FN in patients with DLBCL receiving R-CHOP-21, especially in patients without primary G-CSF prophylaxis.

METHODS

We conducted a retrospective analysis for the clinical data of 103 patients with DLBCL who underwent first R-CHOP-21 without primary G-CSF prophylaxis. The objective of the assessment was the incidence and risk factors of FN after the first chemotherapy cycle.

RESULTS

After the first chemotherapy cycle, the incidence of FN was 20.4%. Multivariate analysis showed that age ≥ 65 years, bone marrow involvement, albumin < 35 g/L, and average relative dose intensity ≥ 80% were independent risk factors for FN. According to risk factors, we created a risk score system. The incidence of FN in the low-, intermediate- and high-risk groups was 5.6%, 17.2%, and 61.9%, respectively.

CONCLUSION

Our data indicated that R-CHOP-21 itself is associated with a high-risk regiment for FN. We recommend that intermediate/high-risk patients should actively consider primary G-CSF prophylaxis to reduce the incidence of FN after chemotherapy.

Construction of Immune Infiltration-Related LncRNA Signatures Based on Machine Learning for the Prognosis in Colon Cancer

Colon cancer is one of the malignant tumors with high morbidity, lethality, and prevalence across global human health. Molecular biomarkers play key roles in its prognosis. In particular, immune-related lncRNAs (IRL) have attracted enormous interest in diagnosis and treatment, but less is known about their potential functions. We aimed to investigate dysfunctional IRL and construct a risk model for improving the outcomes of patients. Nineteen immune cell types were collected for identifying house-keeping lncRNAs (HKLncRNA). GSE39582 and TCGA-COAD were treated as the discovery and validation datasets, respectively. Four machine learning algorithms (LASSO, Random Forest, Boruta, and Xgboost) and a Gaussian mixture model were utilized to mine the optimal combination of lncRNAs. Univariate and multivariate Cox regression was utilized to construct the risk score model. We distinguished the functional difference in an immune perspective between low- and high-risk cohorts calculated by this scoring system. Finally, we provided a nomogram. By leveraging the microarray, sequencing, and clinical data for immune cells and colon cancer patients, we identified the 221 HKLncRNAs with a low cell type-specificity index. Eighty-seven lncRNAs were up-regulated in the immune compared to cancer cells. Twelve lncRNAs were beneficial in improving performance. A risk score model with three lncRNAs (CYB561D2, LINC00638, and DANCR) was proposed with robust ROC performance on an independent dataset. According to immune-related analysis, the risk score is strongly associated with the tumor immune microenvironment. Our results emphasized IRL has the potential to be a powerful and effective therapy for enhancing the prognostic of colon cancer.

Development of a clinical risk score system for peritoneal dialysis-associated peritonitis treatment failure

OBJECTIVE

This study aimed to construct a clinical risk score system for peritoneal dialysis-associated peritonitis (PDAP) treatment failure to provide a theoretical basis for clinical workers.

METHODS

A total of 161 PDAP individuals admitted to our hospital were included, among whom 70 cases were in the treatment-improved group and 87 cases were in the treatment failure group. We compared the general condition, clinical manifestations, and laboratory examination indicators of the two groups of individuals, used multivariate logistic regression analysis to identify the factors influencing PDAP treatment failure, and developed a clinical risk score system. The diagnostic performance of the risk score system was evaluated utilizing the receiver operating characteristic (ROC) curve.

RESULTS

Significant differences (P < 0.05) were observed between the two groups in terms of contamination, peritoneal fluid culture results, blood urea nitrogen (BUN) level, C-reactive protein (CRP) level, B-type natriuretic peptide (BNP) level, average residual urine (RU) volume, and urea clearance rate (UCR). Multivariate logistic regression analysis showed that BUN level, CRP level, BNP level, average RU volume, and UCR were independent risk factors affecting PDAP patient treatment outcomes (P < 0.05). The ROC curve analysis of the risk score system for predicting treatment failure in PDAP individuals showed an area under the curve of 0.895 [95% confidence interval (0.847-0.943)]. The optimal cut-off point was 2.5 points, with corresponding sensitivity and specificity of 88.5% and 74.3%, separately.

CONCLUSION

BUN level, CRP level, BNP level, average RU volume, and UCR are independent risk factors for PDAP treatment failure. The clinical risk score system based on these five independent risk factors can accurately predict the risk of treatment failure in PDAP individuals.

Who would avoid severe adverse events from nasointestinal tube in small bowel obstruction? A matched case-control study

Background

Nasointestinal tubes (NITs) have been increasingly used in patients with small bowel obstruction (SBO); However, severe adverse events (SAEs) of NITs might threaten the lives of patients. The indications of NITs need to be identified. This study was designed to explore the indications for the insertion of NITs in patients with SBO and to suggest the optimal strategies for individuals based on the outcomes of SAEs.

Methods

After propensity score matching, 68 pairs were included (Success group and failure group). The occurrence of SAEs and the clinical parameters were compared between the SAE group and the non-SAE group. Independent risk factors were evaluated among the subgroups. A novel scoring system was established to detect the subgroups that would benefit from NITs insertion.

Results

Successful implementation of NITs could avoid hypochloremia (p = 0.010), SAEs (p = 0.001), pneumonia (p = 0.006). SAEs occurred in 13 of 136 (9.6%) patients who accepted NITs insertion treatment. Risk factors for SAEs included tumors (p = 0.002), reduced BMI (p = 0.048), reduced hemoglobin (p = 0.001), abnormal activated partial thromboplastin time (p = 0.015) and elevated white blood cells (p = 0.002). A novel risk scoring system consists of hemoglobin before NITs insertion (95% CI 0.685, 0.893) and bowel obstruction symptoms relieved after NITs insertion (95% CI 0.575, 0.900) had the highest area under curve for predicting the occurrence of SAEs. We divided the risk score system into 3 grades, with the increasing grades, the rates of SAEs surged from 1.3% (1/74) to (6/11) 54.5%.

Conclusion

NITs successfully insertion could avoid SAEs occurrence in SBO conservative treatment. SBO patients without anemia and could be relieved after NITs insertion could be the potential benefit group for this therapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-022-02405-8.

Competing Nomogram for Late-Period Breast Cancer-Specific Death in Patients with Early-Stage Hormone Receptor-Positive Breast Cancer

BACKGROUND

More than half of early breast cancer recurrences occur after 5 years from the initial diagnosis. An individualized estimate of the risk of late-period breast cancer-specific death (LP-BCSD) after 5 years of endocrine therapy (ET) can improve decision-making for extended endocrine therapy (EET).

MATERIALS AND METHODS

A total of 147,059 eligible patients with breast cancer who survived 5+ years after diagnosis between 1990 and 2007 were identified from the Surveillance, Epidemiology, and End Results (SEER) database. Univariate and multivariate analyses based on the competing risk regression model were used to evaluate predictive factors for high risk of LP-BCSD or late-period non-breast cancer-specific death (LP-non-BCSD). Significant factors were used to build a nomogram to individualize estimates of LP-BCSD or LP-non-BCSD.

RESULTS

The 5- and 10-year LP-BCSD rates were 5.7% and 10.1%, respectively, and the 5- and 10-year LP-non-BCSD rates were 6.7% and 15.5%, respectively. Young age, black race, single marital status, poor differentiation, large tumor size, lymph node metastasis, and estrogen receptor-positive/progesterone receptor-negative (ER+/PR-) status were independent predictive factors for high risk of LP-BCSD. Age was the most important factor for predicting high risk of LP-non-BCSD. The nomograms, which were based on significant factors identified by the competing risk regression model. A risk score system based on the competing risk nomogram was established to describe the relative risk of LP-BCSD and LP-non-BCSD.

CONCLUSION

This study explored the novel endpoint of LP-BCSD for further clinical trials. The risk score system might be highly useful for patient counseling, especially in discussing EET options with elderly or comorbid patients.

A novel risk score for predicting left atrial and left atrial appendage thrombogenic milieu in patients with non-valvular atrial fibrillation

BACKGROUND

Although the CHADS2 and CHA2DS2-VASc scoring systems are commonly used as measures of thromboembolic risk in patients with nonvalvular atrial fibrillation (NVAF), data are inconsistent as to their value in predicting the presence of left atrial (LA) and/or left atrial appendage (LAA) thrombogenic milieu (TM). The present study aimed to establish a novel risk score to assess the risk of LA and/or LAATM in NVAF patients.

METHODS

This is a retrospective case-control study that included 125 consecutive patients with NVAF plus TM, as evidenced by transesophageal echocardiography (TEE) during a period from1 January 2010 to 1 February 2017. The controls were 1098 NVAF patients without TM during the same period. Risk factors for LA and/or LAATM were identified using univariable analysis and multivariable logistic regression. The risk score model was developed based on 10-fold validation and multiple regression. Risk model performance was evaluated using receiver operating characteristic (ROC) curves. Net reclassification improvement (NRI) was used for the comparison of C-statistics. The AUCs were compared using the Z test.

RESULTS

Among all 1223 NVAF patients, 125 (10.22%) patients had LA and/or LAATM. A score system (0-12) was developed based on the following 6 independent variables identified by 10-fold validation with sequential methods. Different points were assigned for each variable, according to multivariable regression using relative coefficients (coefficient of the index risk factor divided by the lowest coefficient among the 6 risk factors; rounded to closest integer): 1 for blood type A and N-terminal B-type natriuretic peptide (NT-proBNP) ≥864.85 pg/mL; 2 for LAD ≥43.5 mm and age ≥ 73.5 years old; 3 for previous heart failure and previous stroke or TIA. The present risk score system had a sensitivity of 58.3%, specificity of 91.4 and accuracy of 81.6%. The area under the ROC curve (AUC) was 0.832, (95% CI: 0.784-0.881; P < 0.001). The negative predictive value (NPV) was 92% when we set the cut-off point at 4; when the cut-off point was set at 8, the positive predictive value (PPV) was 85.7%. Compared with CHADS2 and CHA2DS2-VASc score, the present novel risk score has better predictive power [net reclassification improvement (NRI) +96.3% and +66.2%, respectively; all P < 0.001].

CONCLUSION

This study developed a novel risk score to help predicting LA and/or LAATM in NVAF patients, which had higher accuracy than CHADS2 and CHA2DS2-VASc score system.

Data science and machine learning in anesthesiology

Machine learning (ML) is revolutionizing anesthesiology research. Unlike classical research methods that are largely inference-based, ML is geared more towards making accurate predictions. ML is a field of artificial intelligence concerned with developing algorithms and models to perform prediction tasks in the absence of explicit instructions. Most ML applications, despite being highly variable in the topics that they deal with, generally follow a common workflow. For classification tasks, a researcher typically tests various ML models and compares the predictive performance with the reference logistic regression model. The main advantage of ML lies in its ability to deal with many features with complex interactions and its specific focus on maximizing predictive performance. However, emphasis on data-driven prediction can sometimes neglect mechanistic understanding. This article mainly focuses on the application of supervised ML to electronic health record (EHR) data. The main limitation of EHR-based studies is in the difficulty of establishing causal relationships. However, the associated low cost and rich information content provide great potential to uncover hitherto unknown correlations. In this review, the basic concepts of ML are introduced along with important terms that any ML researcher should know. Practical tips regarding the choice of software and computing devices are also provided. Towards the end, several examples of successful ML applications in anesthesiology are discussed. The goal of this article is to provide a basic roadmap to novice ML researchers working in the field of anesthesiology.

A 13-gene risk score system and a nomogram survival model for predicting the prognosis of clear cell renal cell carcinoma

BACKGROUND

Renal cell carcinoma (RCC) is the second common malignant tumor in the urinary system, and 85% of RCC cases are clear cell RCC (ccRCC). This study is designed to build a risk score system for ccRCC.

METHODS

The gene methylation and expression data of ccRCC samples were downloaded from The Cancer Genome Atlas database (training set) and ArrayExpress database (validation set). The differentially methylated genes (DMGs) and differentially expressed genes (DEGs) were identified by limma package, and their intersecting genes with negative Pearson correlation coefficients were remained using cor.test function. Prognosis-associated genes were identified by survival package, and the optimal DMGs were obtained using penalized package. After risk score system was built, nomogram survival model was constructed using rms package. Additionally, pathways were enriched for the DEGs between high- and low-risk groups using Gene Set Enrichment Analysis.

RESULTS

There were 3,638 DMGs and 2,702 DEGs between tumor and normal samples. Among the 312 intersecting genes, 43 prognosis-associated genes were identified. A total of 13 optimal DMGs (BTBD19, ADAM8, BGLAP, TNFRSF13C, JPH4, BEST1, GNRH2, UBE2QL1, CHODL, GDF9, UPB1, KCNH3; and ADAMTSL4) were obtained for building the risk score system. After pathological M, pathological T, platelet qualitative, and RS status were revealed to be independent prognostic factors, a nomogram survival model was constructed. For the 920 DEGs between the high- and low-risk samples, 6 significant pathways were enriched.

CONCLUSION

The 13-gene risk score system and the nomogram survival model might be used for prognostic prediction of ccRCC patients.

A risk score system to timely manage treatment in Crohn's disease: a cohort study

Background

Clinical severity and intestinal lesions of Crohn’s disease (CD) usually progress over time and require a step up adjustment of the therapy either to prevent or to treat complications. The aim of the study was to  develop a simple risk scoring system to assess in individual CD patients the risk of disease progression and the need for more intensive treatment and monitoring.

Methods

Prospective cohort study (January 2002–September 2014) including 160 CD patients (93 female, median age 31 years; disease behavior (B)1 25%, B2 55.6%, B3 19.4%; location (L)1 61%, L3 31.9%, L2 6%; L4 0.6%; perianal disease 28.8%) seen at 6–12-month interval. Median follow-up 7.9 years (IQR: 4.3–10.5 years). Poisson models were used to evaluate predictors, at each clinical assessment, of having the following outcomes at the subsequent clinical assessment a) use of steroids; b) start of azathioprine; c) start of anti-TNF-α drugs; d) need of surgery. For each outcome 32 variables, including demographic and clinical characteristics of patients and assessment of CD intestinal lesions and complications, were evaluated as potential predictors. The predictors included in the model were chosen by a backward selection. Risk scores were calculated taking for each predictor the integer part of the Poisson model parameter.

Results

Considering 1464 clinical assessments 12 independent risk factors were identified, CD lesions, age at diagnosis < 40 years, stricturing behavior (B2), specific intestinal symptoms, female gender, BMI < 21, CDAI> 50, presence of inflammatory markers, no previous surgery or presence of termino-terminal anastomosis, current use of corticosteroid, no corticosteroid at first flare-up. Six of these predicted steroids use (score 0–9), three to start azathioprine (score 0–4); three to start anti-TNF-α drugs (score 0–4); six need of surgery (score 0–11). The predicted percentage risk to be treated with surgery within one year since the referral assessment varied from 1 to 28%; with azathioprine from 3 to 13%; with anti-TNF-α drugs from 2 to 15%.

Conclusions

These scores may provide a useful clinical tool for clinicians in the prognostic assessment and treatment adjustment of Crohn’s disease in any individual patient.