The added value of genetic information in colorectal cancer risk prediction models: development and evaluation in the UK Biobank prospective cohort study.
Ontology highlight
ABSTRACT:
SUBMITTER: Smith T
PROVIDER: S-EPMC6203780 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BackgroundThis study aimed to compare and externally validate risk scores developed to predict incident colorectal cancer (CRC) that include variables routinely available or easily obtainable via self-completed questionnaire.MethodsExternal validation of fourteen risk models from a previous systematic review in 373 112 men and women within the UK Biobank cohort with 5-year follow-up, no prior history of CRC and data for incidence of CRC through linkage to national cancer registries.ResultsThere were 1719 (0.46%) cases of incident CRC. The performance of the risk models varied substantially. In men, the QCancer10 model and models by Tao, Driver and Ma all had an area under the receiver operating characteristic curve (AUC) between 0.67 and 0.70. Discrimination was lower in women: the QCancer10, Wells, Tao, Guesmi and Ma models were the best performing with AUCs between 0.63 and 0.66. Assessment of calibration was possible for six models in men and women. All would require country-specific recalibration if estimates of absolute risks were to be given to individuals.ConclusionsSeveral risk models based on easily obtainable data have relatively good discrimination in a UK population. Modelling studies are now required to estimate the potential health benefits and cost-effectiveness of implementing stratified risk-based CRC screening.
| S-EPMC5846069 | biostudies-literature
Project description:An amendment to this paper has been published and can be accessed via a link at the top of the paper.
| S-EPMC7217758 | biostudies-literature
Project description:BackgroundMost of the previous studies on diet and colorectal cancer were based on diets consumed during the 1990s.MethodsWe used Cox-regression models to estimate adjusted hazard ratios for colorectal cancer by dietary factors in the UK Biobank study. Men and women aged 40-69?years at recruitment (2006-10) reported their diet on a short food-frequency questionnaire (n?=?475 581). Dietary intakes were re-measured in a large sub-sample (n?=?175?402) who completed an online 24-hour dietary assessment during follow-up. Trends in risk across the baseline categories were calculated by assigning re-measured intakes to allow for measurement error and changes in intake over time.ResultsDuring an average of 5.7?years of follow-up, 2609 cases of colorectal cancer occurred. Participants who reported consuming an average of 76?g/day of red and processed meat compared with 21?g/day had a 20% [95% confidence interval (CI): 4-37] higher risk of colorectal cancer. Participants in the highest fifth of intake of fibre from bread and breakfast cereals had a 14% (95% CI: 2-24) lower risk of colorectal cancer. Alcohol was associated with an 8% (95% CI: 4-12) higher risk per 10?g/day higher intake. Fish, poultry, cheese, fruit, vegetables, tea and coffee were not associated with colorectal-cancer risk.ConclusionsConsumption of red and processed meat at an average level of 76 g/d that meets the current UK government recommendation (?90?g/day) was associated with an increased risk of colorectal cancer. Alcohol was also associated with an increased risk of colorectal cancer, whereas fibre from bread and breakfast cereals was associated with a reduced risk.
| S-EPMC7124508 | biostudies-literature
Project description:BackgroundIt remains unclear whether serum lipids influence colorectal cancer (CRC) risk.MethodsWe conducted a prospective cohort study of 380,087 adults aged 40-69 years in the UK Biobank. Serum high-density cholesterol, low-density cholesterol, total cholesterol, triglycerides, and apolipoprotein A and B were measured. We used Cox proportional hazard models to estimate the multivariable hazard ratios (HRs) of CRC according to one standard deviation (SD) increment in serum lipids. We conducted subgroup analysis by tumour anatomical subsites.ResultsDuring a median of 10.3 years of follow-up, we documented 2667 incident CRC cases. None of the lipid biomarkers was associated with the risk of CRC after adjusting for potential confounding factors, including body mass index and waist circumference. When assessed by cancer subsites, serum triglycerides was associated with an increased risk of cancer in the caecum and transverse colon, with the HR of 1.12 (95% CI, 1.00-1.25) and 1.29 (95% CI, 1.09-1.53), respectively; and apolipoprotein A was associated with a lower risk of hepatic flexure cancer (HR, 0.73, 95% CI, 0.56-0.96).ConclusionsSerum lipid profiles were not associated with colorectal cancer risk after adjusting for obesity indicators. The potential subsite-specific effects of triglycerides and apolipoprotein A require further confirmation.
| S-EPMC7851156 | biostudies-literature
Project description:BackgroundBoth genetic and lifestyle factors play an etiologic role in colorectal cancer (CRC).ObjectivesWe evaluated potential gene-environment interactions in CRC risk.MethodsWe used data from 346,297 participants in the UK Biobank cohort. Healthy lifestyle scores (HLSs) were constructed using 8 lifestyle factors, primarily according to the American Cancer Society guidelines, and were categorized into unhealthy, intermediate, and healthy groups. A polygenic risk score (PRS) was created using 95 genetic risk variants identified by genome-wide association studies of CRC and was categorized by tertile. Cox models were used to estimate the HRs and 95% CIs of CRC risk associated with the HLS and PRS.ResultsDuring a median follow-up of 5.8 y, 2066 incident cases of CRC were identified. Healthier HLSs were associated with reduced risk of CRC in a dose-response manner. The risk reduction was more apparent among those with high PRS (HRhealthy vs. unhealthy HLS1: 0.58; 95% CI: 0.43, 0.79 for men and 0.71; 0.58, 0.85 for men and women combined) than those with low PRS. Although no multiplicative interactions were identified, the HLS1 and PRS showed a significant additive interaction (P = 0.02 for all participants combined, 0.04 for men). In analyses including all participants, the adjusted CRC cumulative risk from age 40 to 75 y was 6.40% for those with high PRS/unhealthy HLS1, with a relative excess risk due to interaction of 0.58 (95% CI: 0.06, 1.10), compared with 2.09% among those with low PRS/healthy HLS1. This pattern was more apparent among those who reported not having received any bowel screening before baseline.ConclusionsAlthough the observational nature of the study precludes proof of causality, our findings suggest that individuals with a high genetic susceptibility could benefit more substantially than those with a low genetic risk from lifestyle modification in reducing CRC risk.
| S-EPMC8023827 | biostudies-literature
Project description:The aim of this study was to compare and externally validate risk scores developed to predict incident colorectal cancer that include common genetic variants (SNPs), with or without established lifestyle/environmental (questionnaire-based/classical/phenotypic) risk factors. We externally validated 23 risk models from a previous systematic review in 443,888 participants ages 37 to 73 from the UK Biobank cohort who had 6-year prospective follow-up, no prior history of colorectal cancer, and data for incidence of colorectal cancer through linkage to national cancer registries. There were 2,679 (0.6%) cases of incident colorectal cancer. We assessed model discrimination using the area under the operating characteristic curve (AUC) and relative risk calibration. The AUC of models including only SNPs increased with the number of included SNPs and was similar in men and women: the model by Huyghe with 120 SNPs had the highest AUC of 0.62 [95% confidence interval (CI), 0.59-0.64] in women and 0.64 (95% CI, 0.61-0.66) in men. Adding phenotypic risk factors without age improved discrimination in men but not in women. Adding phenotypic risk factors and age increased discrimination in all cases (P < 0.05), with the best performing models including SNPs, phenotypic risk factors, and age having AUCs between 0.64 and 0.67 in women and 0.67 and 0.71 in men. Relative risk calibration varied substantially across the models. Among middle-aged people in the UK, existing polygenic risk scores discriminate moderately well between those who do and do not develop colorectal cancer over 6 years. Consideration should be given to exploring the feasibility of incorporating genetic and lifestyle/environmental information in any future stratified colorectal cancer screening program.
| S-EPMC7610623 | biostudies-literature
Project description:The discrimination of a risk prediction model measures that model's ability to distinguish between subjects with and without events. The area under the receiver operating characteristic curve (AUC) is a popular measure of discrimination. However, the AUC has recently been criticized for its insensitivity in model comparisons in which the baseline model has performed well. Thus, 2 other measures have been proposed to capture improvement in discrimination for nested models: the integrated discrimination improvement and the continuous net reclassification improvement. In the present study, the authors use mathematical relations and numerical simulations to quantify the improvement in discrimination offered by candidate markers of different strengths as measured by their effect sizes. They demonstrate that the increase in the AUC depends on the strength of the baseline model, which is true to a lesser degree for the integrated discrimination improvement. On the other hand, the continuous net reclassification improvement depends only on the effect size of the candidate variable and its correlation with other predictors. These measures are illustrated using the Framingham model for incident atrial fibrillation. The authors conclude that the increase in the AUC, integrated discrimination improvement, and net reclassification improvement offer complementary information and thus recommend reporting all 3 alongside measures characterizing the performance of the final model.
| S-EPMC3530349 | biostudies-literature
Project description:BackgroundComponents of social connection are associated with mortality, but research examining their independent and combined effects in the same dataset is lacking. This study aimed to examine the independent and combined associations between functional and structural components of social connection and mortality.MethodsAnalysis of 458,146 participants with full data from the UK Biobank cohort linked to mortality registers. Social connection was assessed using two functional (frequency of ability to confide in someone close and often feeling lonely) and three structural (frequency of friends/family visits, weekly group activities, and living alone) component measures. Cox proportional hazard models were used to examine the associations with all-cause and cardiovascular disease (CVD) mortality.ResultsOver a median of 12.6 years (IQR 11.9-13.3) follow-up, 33,135 (7.2%) participants died, including 5112 (1.1%) CVD deaths. All social connection measures were independently associated with both outcomes. Friends/family visit frequencies < monthly were associated with a higher risk of mortality indicating a threshold effect. There were interactions between living alone and friends/family visits and between living alone and weekly group activity. For example, compared with daily friends/family visits-not living alone, there was higher all-cause mortality for daily visits-living alone (HR 1.19 [95% CI 1.12-1.26]), for never having visits-not living alone (1.33 [1.22-1.46]), and for never having visits-living alone (1.77 [1.61-1.95]). Never having friends/family visits whilst living alone potentially counteracted benefits from other components as mortality risks were highest for those reporting both never having visits and living alone regardless of weekly group activity or functional components. When all measures were combined into overall functional and structural components, there was an interaction between components: compared with participants defined as not isolated by both components, those considered isolated by both components had higher CVD mortality (HR 1.63 [1.51-1.76]) than each component alone (functional isolation 1.17 [1.06-1.29]; structural isolation 1.27 [1.18-1.36]).ConclusionsThis work suggests (1) a potential threshold effect for friends/family visits, (2) that those who live alone with additional concurrent markers of structural isolation may represent a high-risk population, (3) that beneficial associations for some types of social connection might not be felt when other types of social connection are absent, and (4) considering both functional and structural components of social connection may help to identify the most isolated in society.
| S-EPMC10637015 | biostudies-literature
Project description:BackgroundRegular stair climbing has the potential to lower the risk of premature death, but current evidence is scarce. We aimed to examine whether daily stair climbing is associated with lower risk of all-cause, cancer, and cardiovascular disease (CVD) mortality.MethodsUsing the UK Biobank cohort, we extracted information of self-reported daily flights of stairs climbed at home, categorized as none, 1 to 5, 6 to 10, 11 to 15, and ≥16 flights per day. Associations between flights of stair climbed per day and mortality were examined as hazard ratios (HRs) from Cox proportional hazards models adjusted for demographic, clinical, and behavioural covariates including time spent in other physical activities. We calculated the restricted mean survival time as an absolute measure of association. The risk of residual confounding was examined using propensity score matching and by using lung cancer as negative control outcome. Participants were followed from baseline (2006-2010) through 31 March 2020.ResultsA total of 280 423 participants (median follow-up 11.1 years, during which 9445 deaths occurred) were included. Compared with not climbing any stairs, climbing more than five flights of stairs at home per day was associated with lower risk of premature mortality. The lowest risk was found for those climbing 6-10 flights per day: 0.91; 95% confidence interval (CI): 0.85, 0.98, translated to approximately 44 to 55 days of additional survival. A similar pattern was found after applying propensity score matching and for cancer mortality (6-10 flights per day HR: 0.88; 95% CI: 0.80, 0.97), but not for CVD mortality (6-10 flights per day HR: 1.08; 95% CI: 0.91, 1.29). The association between stair climbing and lung cancer was similar to that of all-cause mortality.ConclusionsClimbing more than five flights of stairs at home per day was associated with a lower risk of all-cause and cancer mortality, but not CVD mortality, compared with those who did not take the stairs. The magnitude of the association was small and appeared susceptible to residual confounding. It is unlikely that at-home stair climbing is sufficient physical activity stimuli to lower the risk of premature mortality.
| S-EPMC8061405 | biostudies-literature
Project description:ObjectiveAtrial fibrillation (AF) is the most common cardiac arrythmia, and it is associated with increased risk for ischemic stroke, which is underestimated, as AF can be asymptomatic. The aim of this study was to develop optimal ML models for prediction of AF in the population, and secondly for ischemic stroke in AF patients.MethodsTo develop ML models for prediction of 1) AF in the general population and 2) ischemic stroke in patients with AF we constructed XGBoost, LightGBM, Random Forest, Deep Neural Network, Support Vector Machine and Lasso penalised logistic regression models using UK-Biobank's extensive real-world clinical data, questionnaires, as well as biochemical and genetic data, and their predictive performances were compared. Ranking and contribution of the different features was assessed by SHapley Additive exPlanations (SHAP) analysis. The clinical tool CHA2DS2-VASc for prediction of ischemic stroke among AF patients, was used for comparison to the best performing ML model.FindingsThe best performing model for AF prediction was LightGBM, with an area-under-the-roc-curve (AUROC) of 0.729 (95% confidence intervals (CI): 0.719, 0.738). The best performing model for ischemic stroke prediction in AF patients was XGBoost with AUROC of 0.631 (95% CI: 0.604, 0.657). The improved AUROC in the XGBoost model compared to CHA2DS2-VASc was statistically significant based on DeLong's test (p-value = 2.20E-06). In addition, the SHAP analysis showed that several peripheral blood biomarkers (e.g. creatinine, glycated haemoglobin, monocytes) were associated with ischemic stroke, which are not considered by CHA2DS2-VASc.ImplicationsThe best performing ML models presented have the potential for clinical use, but further validation in independent studies is required. Our results endorse the incorporation of some routinely measured blood biomarkers for ischemic stroke prediction in AF patients.
| S-EPMC10987914 | biostudies-literature