Abstract
Aims
The aim of this study was to derive and validate the SCORE2-Older Persons (SCORE2-OP) risk model to estimate 5- and 10-year risk of cardiovascular disease (CVD) in individuals aged ...over 70 years in four geographical risk regions.
Methods and results
Sex-specific competing risk-adjusted models for estimating CVD risk (CVD mortality, myocardial infarction, or stroke) were derived in individuals aged over 65 without pre-existing atherosclerotic CVD from the Cohort of Norway (28 503 individuals, 10 089 CVD events). Models included age, smoking status, diabetes, systolic blood pressure, and total- and high-density lipoprotein cholesterol. Four geographical risk regions were defined based on country-specific CVD mortality rates. Models were recalibrated to each region using region-specific estimated CVD incidence rates and risk factor distributions. For external validation, we analysed data from 6 additional study populations {338 615 individuals, 33 219 CVD validation cohorts, C-indices ranged between 0.63 95% confidence interval (CI) 0.61–0.65 and 0.67 (0.64–0.69)}. Regional calibration of expected-vs.-observed risks was satisfactory. For given risk factor profiles, there was substantial variation across the four risk regions in the estimated 10-year CVD event risk.
Conclusions
The competing risk-adjusted SCORE2-OP model was derived, recalibrated, and externally validated to estimate 5- and 10-year CVD risk in older adults (aged 70 years or older) in four geographical risk regions. These models can be used for communicating the risk of CVD and potential benefit from risk factor treatment and may facilitate shared decision-making between clinicians and patients in CVD risk management in older persons.
Graphical Abstract
Development process, risk regions and illustrative example for the SCORE2-OP algorithm.
The area under the receiver operating characteristic (ROC) curve (AUC) is commonly used for assessing the discriminative ability of prediction models even though the measure is criticized for being ...clinically irrelevant and lacking an intuitive interpretation. Every tutorial explains how the coordinates of the ROC curve are obtained from the risk distributions of diseased and non-diseased individuals, but it has not become common sense that therewith the ROC plot is just another way of presenting these risk distributions. We show how the ROC curve is an alternative way to present risk distributions of diseased and non-diseased individuals and how the shape of the ROC curve informs about the overlap of the risk distributions. For example, ROC curves are rounded when the prediction model included variables with similar effect on disease risk and have an angle when, for example, one binary risk factor has a stronger effect; and ROC curves are stepped rather than smooth when the sample size or incidence is low, when the prediction model is based on a relatively small set of categorical predictors. This alternative perspective on the ROC plot invalidates most purported limitations of the AUC and attributes others to the underlying risk distributions. AUC is a measure of the discriminative ability of prediction models. The assessment of prediction models should be supplemented with other metrics to assess their clinical utility.
Introduction
Breast cancer is a significant contributor to female mortality, exerting a public health burden worldwide, especially in China, where risk‐prediction models with good discriminating ...accuracy for breast cancer are still scarce.
Methods
A multicenter screening cohort study was conducted as part of the Cancer Screening Program in Urban China. Dwellers aged 40–74 years were recruited between 2014 and 2019 and prospectively followed up until June 30, 2021. The entire data set was divided by year of enrollment to develop a prediction model and validate it internally. Multivariate Cox regression was used to ascertain predictors and develop a risk‐prediction model. Model performance at 1, 3, and 5 years was evaluated using the area under the curve, nomogram, and calibration curves and subsequently validated internally. The prediction model incorporates selected factors that are assigned appropriate weights to establish a risk‐scoring algorithm. Guided by the risk score, participants were categorized into low‐, medium‐, and high‐risk groups for breast cancer. The cutoff values were chosen using X‐tile plots. Sensitivity analysis was conducted by categorizing breast cancer risk into the low‐ and high‐risk groups. A decision curve analysis was used to assess the clinical utility of the model.
Results
Of the 70,520 women enrolled, 447 were diagnosed with breast cancer (median follow‐up, 6.43 interquartile range, 3.99–7.12 years). The final prediction model included age and education level (high, hazard ratio HR, 2.01 95% CI, 1.31–3.09), menopausal age (≥50 years, 1.34 1.03–1.75), previous benign breast disease (1.42 1.09–1.83), and reproductive surgery (1.28 0.97–1.69). The 1‐year area under the curve was 0.607 in the development set and 0.643 in the validation set. Moderate predictive discrimination and satisfactory calibration were observed for the validation set. The risk predictions demonstrated statistically significant differences between the low‐, medium‐, and high‐risk groups (p < .001). Compared with the low‐risk group, women in the high‐ and medium‐risk groups posed a 2.17‐fold and 1.62‐fold elevated risk of breast cancer, respectively. Similar results were obtained in the sensitivity analyses. A web‐based calculator was developed to estimate risk stratification for women.
Conclusions
This study developed and internally validated a risk‐adapted and user‐friendly risk‐prediction model by incorporating easily accessible variables and female factors. The personalized model demonstrated reliable calibration and moderate discriminative ability. Risk‐stratified screening strategies contribute to precisely distinguishing high‐risk individuals from asymptomatic individuals and prioritizing breast cancer screening.
Plain Language Summary
Breast cancer remains a burden in China. To enhance breast cancer screening, we need to incorporate population stratification in screening.
Accurate risk‐prediction models for breast cancer remain scarce in China. We established and validated a risk‐adapted and user‐friendly risk‐prediction model by incorporating routinely available variables along with female factors.
Using this risk‐stratified model helps accurately identify high‐risk individuals, which is of significant importance when considering integrating individual risk assessments into mass screening programs for breast cancer.
Current clinical breast cancer screening lacks a constructive clinical pathway and guiding recommendations. Our findings can better guide clinicians and health care providers.
This study developed and internally validated a risk‐adapted and user‐friendly risk‐prediction model by incorporating easily accessible variables and female factors. Risk‐stratified screening strategies contribute to precisely distinguishing high‐risk individuals from asymptomatic individuals and prioritizing breast cancer screening.
Display omitted
•A risk score was developed that predicted HCC in Asian patients with chronic hepatitis B under antiviral therapy.•Modified PAGE-B scores required easily available information on age, ...gender, platelet counts, and serum albumin levels.•Modified PAGE-B scores significantly differentiated the 5-year HCC risk: low ≤8 and high ≥13.
Recently, the PAGE-B score and Toronto HCC risk index (THRI) have been developed to predict the risk of hepatocellular carcinoma (HCC) in Caucasian patients with chronic hepatitis B (CHB). We aimed to validate PAGE-B scores and THRI in Asian patients with CHB and suggested modified PAGE-B scores to improve the predictive performance.
From 2007 to 2017, we examined 3,001 Asian patients with CHB receiving entecavir/tenofovir therapy. We assessed the performances of PAGE-B, THRI, CU-HCC, GAG-HCC, and REACH-B for HCC development. A modified PAGE-B score (mPAGE-B) was developed (derivation set, n = 2,001) based on multivariable Cox models. Bootstrap for internal validation and external validation (validation set, n = 1,000) were performed.
The five-year cumulative HCC incidence rates were 6.6% and 7.2% in the derivation and validation datasets after entecavir/tenofovir onset. In the derivation dataset, age, gender, serum albumin levels, and platelet counts were independently associated with HCC. The mPAGE-B score was developed based on age, gender, platelet counts, and serum albumin levels (time-dependent area under receiver operating characteristic curves AUROC = 0.82). In the validation set, the PAGE-B and THRI showed similar AUROCs to CU-HCC, GAG-HCC, and REACH-B at five years (0.72 and 0.73 vs. 0.70, 0.71, and 0.61 respectively; all p >0.05 except REACH-B), whereas the AUROC of mPAGE-B at five years was 0.82, significantly higher than the five other models (all p <0.01). HCC incidence rates after initiation of entecavir/tenofovir therapy in patients with CHB were significantly decreased in all risk groups in long-term follow-up periods.
Although PAGE-B and THRI are applicable in Asian patients with CHB receiving entecavir/tenofovir therapy, mPAGE-B scores including additional serum albumin levels showed better predictive performance than the PAGE-B score.
PAGE-B scores and Toronto HCC risk index were developed to predict the risk of hepatocellular carcinoma (HCC) in Caucasian patients with CHB under potent antiviral therapy. This study validated these two scores in Asian patients with CHB and suggested that modified PAGE-B scores could improve the predictive performance. A modified PAGE-B score, which is based only on a patient’s age, gender, baseline platelet counts, and serum albumin levels at treatment initiation, represents a reliable and easily available risk score to predict HCC development during the first five years of antiviral treatment in Asian patients with CHB. With a scoring range from 0 to 21 points, a modified PAGE-B score differentiates the HCC risk. A modified PAGE-B score significantly differentiates the five-year HCC risk: low ≤8 points and high ≥13 points.
•A method is designed to extract driving behaviour features and predict risk levels.•Massive driving behaviour features are extracted from real vehicle trajectory data.•Key features are identified by ...feature importance ranking and recursive elimination.•XGBoost can achieve satisfactory results of behaviour-based crash risk prediction.
This study designs a framework of feature extraction and selection, to assess vehicle driving and predict risk levels. The framework integrates learning-based feature selection, unsupervised risk rating, and imbalanced data resampling. For each vehicle, about 1300 driving behaviour features are extracted from trajectory data, which produce in-depth and multi-view measures on behaviours. To estimate the risk potentials of vehicles in driving, unsupervised data labelling is proposed. Based on extracted risk indicator features, vehicles are clustered into various groups labelled with graded risk levels. Data under-sampling of the safe group is performed to reduce the risk-safe class imbalance. Afterwards, the linkages between behaviour features and corresponding risk levels are built using XGBoost, and key features are identified according to feature importance ranking and recursive elimination. The risk levels of vehicles in driving are predicted based on key features selected. As a case study, NGSIM trajectory data are used in which four risk levels are clustered by Fuzzy C-means, 64 key behaviour features are identified, and an overall accuracy of 89% is achieved for behaviour-based risk prediction. Findings show that this approach is effective and reliable to identify important features for driving assessment, and achieve an accurate prediction of risk levels.
Urologists regularly develop clinical risk prediction models to support clinical decisions. In contrast to traditional performance measures, decision curve analysis (DCA) can assess the utility of ...models for decision making. DCA plots net benefit (NB) at a range of clinically reasonable risk thresholds.
To provide recommendations on interpreting and reporting DCA when evaluating prediction models.
We informally reviewed the urological literature to determine investigators’ understanding of DCA. To illustrate, we use data from 3616 patients to develop risk models for high-grade prostate cancer (n=313, 9%) to decide who should undergo a biopsy. The baseline model includes prostate-specific antigen and digital rectal examination; the extended model adds two predictors based on transrectal ultrasound (TRUS).
We explain risk thresholds, NB, default strategies (treat all, treat no one), and test tradeoff. To use DCA, first determine whether a model is superior to all other strategies across the range of reasonable risk thresholds. If so, that model appears to improve decisions irrespective of threshold. Second, consider if there are important extra costs to using the model. If so, obtain the test tradeoff to check whether the increase in NB versus the best other strategy is worth the additional cost. In our case study, addition of TRUS improved NB by 0.0114, equivalent to 1.1 more detected high-grade prostate cancers per 100 patients. Hence, adding TRUS would be worthwhile if we accept subjecting 88 patients to TRUS to find one additional high-grade prostate cancer or, alternatively, subjecting 10 patients to TRUS to avoid one unnecessary biopsy.
The proposed guidelines can help researchers understand DCA and improve application and reporting.
Decision curve analysis can identify risk models that can help us make better clinical decisions. We illustrate appropriate reporting and interpretation of decision curve analysis.
Decision curve analysis (DCA) is a method to evaluate whether risk prediction models can have utility for supporting treatment decisions. This guide for researchers explains what DCA is, how to interpret it, and how to report its results.
This position statement from the Heart Failure Association of the European Society of Cardiology Cardio‐Oncology Study Group in collaboration with the International Cardio‐Oncology Society presents ...practical, easy‐to‐use and evidence‐based risk stratification tools for oncologists, haemato‐oncologists and cardiologists to use in their clinical practice to risk stratify oncology patients prior to receiving cancer therapies known to cause heart failure or other serious cardiovascular toxicities. Baseline risk stratification proformas are presented for oncology patients prior to receiving the following cancer therapies: anthracycline chemotherapy, HER2‐targeted therapies such as trastuzumab, vascular endothelial growth factor inhibitors, second and third generation multi‐targeted kinase inhibitors for chronic myeloid leukaemia targeting BCR‐ABL, multiple myeloma therapies (proteasome inhibitors and immunomodulatory drugs), RAF and MEK inhibitors or androgen deprivation therapies. Applying these risk stratification proformas will allow clinicians to stratify cancer patients into low, medium, high and very high risk of cardiovascular complications prior to starting treatment, with the aim of improving personalised approaches to minimise the risk of cardiovascular toxicity from cancer therapies.
The dual antiplatelet therapy (DAPT) score guides decisions on DAPT duration after coronary stenting by simultaneously predicting ischemic and bleeding risk.
This study sought to assess the ...performance of the DAPT score in a nationwide real-world population.
The study used register data in Sweden (2006 to 2014) and followed 41,101 patients who had undergone 12 months of event-free DAPT, from months 12 to 30 after stenting. Risk of myocardial infarction (MI) or stent thrombosis, major adverse cardiovascular and cerebrovascular events (MACCE) (MI, stroke, and all-cause death), and fatal or major bleeding were compared according to DAPT score.
The score had a discrimination of 0.58 (95% confidence interval CI: 0.56 to 0.60) for MI or stent thrombosis, 0.54 (95% CI: 0.53 to 0.55) for MACCE, and 0.49 (95% CI: 0.45 to 0.53) for fatal or major bleeding. Risk of MI or stent thrombosis was significantly increased at scores of ≥3 while MACCE risk followed a J-shaped pattern and increased at scores of ≥4. Absolute differences in fatal or major bleeding risk were small between scores. Event rates of ischemic and bleeding outcomes in patients with high (≥2) and low (<2) scores differed compared to the DAPT Study from which the score was derived; fatal or major bleeding rates were approximately one-half of those in the placebo arm of the DAPT Study.
In a nationwide population, the DAPT score did not adequately discriminate ischemic and bleeding risk, the relationship between score and ischemic risk did not correspond to the suggested decision rule for extended DAPT, and risk of bleeding was lower compared with the DAPT Study. The score and its decision rule may not be generalizable to real-world populations.
Display omitted
An Expert Panel of the National Lipid Association reviewed the evidence related to the use of coronary artery calcium (CAC) scoring in clinical practice for adults seen for primary prevention of ...atherosclerotic cardiovascular disease. Recommendations for optimal use of this test in adults of various races/ethnicities, ages and multiple domains of primary prevention, including those with a 10-year ASCVD risk <20%, those with diabetes or the metabolic syndrome, and those with severe hypercholesterolemia were provided. Recommendations were also made on optimal timing for repeat calcium scoring after an initial test, use of CAC scoring in those taking statins, and its role in informing the clinician patient discussion on the benefit of aspirin and anti-hypertensive drug therapy. Finally, a vision is provided for the future of coronary calcium scoring.
•CAC scoring strongly informs ASCVD risk discrimination and reclassification.•CAC scoring aids in ASCVD risk prediction, regardless of race, gender or ethnicity.•CAC scoring aids the clinician to allocate statin therapy based on ASCVD risk.•Very high CAC scores may inform decision-making about add-on therapies to statins.•CAC scoring aids decision-making about aspirin and anti-hypertensive therapy.