Showing papers by "Ralph B. D'Agostino published in 2017"

Goodness-of-Fit-Techniques

Abstract: "a comprehensive book, well presented and clearly written." The New Zealand Statistician "Chapters contain some theory but they are mainly directed toward application with some numerical illustrations which generally use simulated sets to be found in the AppendixThe level of exposition is generally clear and directthe book should prove useful to practising statisticians." Short Book Reviews (International Statistical Institute) "the authors have carried out the helpful task of bringing together into a useful reference volume a large amount of material on an interesting topic which has previously been scattered throughout the literature." Royal Statistical Society

Association of Type 1 Diabetes vs Type 2 Diabetes Diagnosed During Childhood and Adolescence With Complications During Teenage Years and Young Adulthood.

Abstract: Importance The burden and determinants of complications and comorbidities in contemporary youth-onset diabetes are unknown. Objective To determine the prevalence of and risk factors for complications related to type 1 diabetes vs type 2 diabetes among teenagers and young adults who had been diagnosed with diabetes during childhood and adolescence. Design, Setting, and Participants Observational study from 2002 to 2015 in 5 US locations, including 2018 participants with type 1 and type 2 diabetes diagnosed at younger than 20 years, with single outcome measures between 2011 and 2015. Exposures Type 1 and type 2 diabetes and established risk factors (hemoglobin A 1c level, body mass index, waist-height ratio, and mean arterial blood pressure). Main Outcomes and Measures Diabetic kidney disease, retinopathy, peripheral neuropathy, cardiovascular autonomic neuropathy, arterial stiffness, and hypertension. Results Of 2018 participants, 1746 had type 1 diabetes (mean age, 17.9 years [SD, 4.1]; 1327 non-Hispanic white [76.0%]; 867 female patients [49.7%]), and 272 had type 2 (mean age, 22.1 years [SD, 3.5]; 72 non-Hispanic white [26.5%]; 181 female patients [66.5%]). Mean diabetes duration was 7.9 years (both groups). Patients with type 2 diabetes vs those with type 1 had higher age-adjusted prevalence of diabetic kidney disease (19.9% vs 5.8%; absolute difference [AD], 14.0%; 95% CI, 9.1%-19.9%; P P = .02), peripheral neuropathy (17.7% vs 8.5%; AD, 9.2%; 95% CI, 4.8%-14.4%; P P P P = .62). After adjustment for established risk factors measured over time, participants with type 2 diabetes vs those with type 1 had significantly higher odds of diabetic kidney disease (odds ratio [OR], 2.58; 95% CI, 1.39-4.81; P =.003), retinopathy (OR, 2.24; 95% CI, 1.11-4.50; P = .02), and peripheral neuropathy (OR, 2.52; 95% CI, 1.43-4.43; P = .001), but no significant difference in the odds of arterial stiffness (OR, 1.07; 95% CI, 0.63-1.84; P = .80) and hypertension (OR, 0.85; 95% CI, 0.50-1.45; P = .55). Conclusions and Relevance Among teenagers and young adults who had been diagnosed with diabetes during childhood or adolescence, the prevalence of complications and comorbidities was higher among those with type 2 diabetes compared with type 1, but frequent in both groups. These findings support early monitoring of youth with diabetes for development of complications.

Tutorial in biostatistics: data-driven subgroup identification and analysis in clinical trials

Abstract: It is well known that both the direction and magnitude of the treatment effect in clinical trials are often affected by baseline patient characteristics (generally referred to as biomarkers). Characterization of treatment effect heterogeneity plays a central role in the field of personalized medicine and facilitates the development of tailored therapies. This tutorial focuses on a general class of problems arising in data-driven subgroup analysis, namely, identification of biomarkers with strong predictive properties and patient subgroups with desirable characteristics such as improved benefit and/or safety. Limitations of ad-hoc approaches to biomarker exploration and subgroup identification in clinical trials are discussed, and the ad-hoc approaches are contrasted with principled approaches to exploratory subgroup analysis based on recent advances in machine learning and data mining. A general framework for evaluating predictive biomarkers and identification of associated subgroups is introduced. The tutorial provides a review of a broad class of statistical methods used in subgroup discovery, including global outcome modeling methods, global treatment effect modeling methods, optimal treatment regimes, and local modeling methods. Commonly used subgroup identification methods are illustrated using two case studies based on clinical trials with binary and survival endpoints. Copyright © 2016 John Wiley & Sons, Ltd.

Challenges in the Design and Interpretation of Noninferiority Trials

Abstract: In this review, the authors consider the characteristics and challenges of noninferiority trial designs, including seven essential features of noninferiority trials.

Safety and tolerability of the first-in-class agent CPI-613 in combination with modified FOLFIRINOX in patients with metastatic pancreatic cancer: a single-centre, open-label, dose-escalation, phase 1 trial

Abstract: Summary Background Pancreatic cancer statistics are dismal, with a 5-year survival of less than 10%, and more than 50% of patients presenting with metastatic disease. Metabolic reprogramming is an emerging hallmark of pancreatic adenocarcinoma. CPI-613 is a novel anticancer agent that selectively targets the altered form of mitochondrial energy metabolism in tumour cells, causing changes in mitochondrial enzyme activities and redox status that lead to apoptosis, necrosis, and autophagy of tumour cells. We aimed to establish the maximum tolerated dose of CPI-613 when used in combination with modified FOLFIRINOX chemotherapy (comprising oxaliplatin, leucovorin, irinotecan, and fluorouracil) in patients with metastatic pancreatic cancer. Methods In this single-centre, open-label, dose-escalation phase 1 trial, we recruited adult patients (aged ≥18 years) with newly diagnosed metastatic pancreatic adenocarcinoma from the Comprehensive Cancer Center of Wake Forest Baptist Medical Center (Winston-Salem, NC, USA). Patients had good bone marrow, liver and kidney function, and good performance status (Eastern Cooperative Oncology Group [ECOG] performance status 0–1). We studied CPI-613 in combination with modified FOLFIRINOX (oxaliplatin at 65 mg/m 2 , leucovorin at 400 mg/m 2 , irinotecan at 140 mg/m 2 , and fluorouracil 400 mg/m 2 bolus followed by 2400 mg/m 2 over 46 h). We applied a two-stage dose-escalation scheme (single patient and traditional 3+3 design). In the single-patient stage, one patient was accrued per dose level. The starting dose of CPI-613 was 500 mg/m 2 per day; the dose level was then escalated by doubling the previous dose if there were no adverse events worse than grade 2 within 4 weeks attributed as probably or definitely related to CPI-613. The traditional 3+3 dose-escalation stage was triggered if toxic effects attributed as probably or definitely related to CPI-613 were grade 2 or worse. The dose level for CPI-613 for the first cohort in the traditional dose-escalation stage was the same as that used in the last cohort of the single-patient dose-escalation stage. The primary objective was to establish the maximum tolerated dose of CPI-613 (as assessed by dose-limiting toxicities). This trial is registered with ClinicalTrials.gov, number NCT01835041, and is closed to recruitment. Findings Between April 22, 2013, and Jan 8, 2016, we enrolled 20 patients. The maximum tolerated dose of CPI-613 was 500 mg/m 2 . The median number of treatment cycles given at the maximum tolerated dose was 11 (IQR 4–19). Median follow-up of the 18 patients treated at the maximum tolerated dose was 378 days (IQR 250–602). Two patients enrolled at a higher dose of 1000 mg/m 2 , and both had a dose-limiting toxicity. Two unexpected serious adverse events occurred, both for the first patient enrolled. Expected serious adverse events were: thrombocytopenia, anaemia, and lymphopenia (all for patient number 2; anaemia and lymphopenia were dose-limiting toxicities); hyperglycaemia (in patient number 7); hypokalaemia, hypoalbuminaemia, and sepsis (patient number 11); and neutropenia (patient number 20). No deaths due to adverse events were reported. For the 18 patients given the maximum tolerated dose, the most common grade 3–4 non-haematological adverse events were hyperglycaemia (ten [55%] patients), hypokalaemia (six [33%]), peripheral sensory neuropathy (five [28%]), diarrhoea (five [28%]), and abdominal pain (four [22%]). The most common grade 3–4 haematological adverse events were neutropenia (five [28%] of 18 patients), lymphopenia (five [28%]), anaemia (four [22%], and thrombocytopenia in three [17%]). Sensory neuropathy (all grade 1–3) was recorded in 17 (94%) of the 18 patients and was managed with dose de-escalation or discontinuation per standard of care. No patients died while on active treatment; 11 study participants died, with cause of death as terminal pancreatic cancer. Of the 18 patients given the maximum tolerated dose, 11 (61%) achieved an objective (complete or partial) response. Interpretation A maximum tolerated dose of CPI-613 was established at 500 mg/m 2 when used in combination with modified FOLFIRINOX in patients with metastatic pancreatic cancer. The findings of clinical activity will require validation in a phase 2 trial. Funding Comprehensive Cancer Center of Wake Forest Baptist Medical Center.

Association Between Use of a Scalp Cooling Device and Alopecia After Chemotherapy for Breast Cancer

Abstract: Importance Chemotherapy-induced alopecia is a common and distressing adverse effect. In previous studies of scalp cooling to prevent chemotherapy-induced alopecia, conclusions have been limited. Objectives To evaluate whether use of a scalp cooling system is associated with a lower amount of hair loss among women receiving specific chemotherapy regimens for early-stage breast cancer and to assess related changes in quality of life. Design, Setting, and Participants A prospective cohort study conducted at 5 US medical centers of women with stage I or II breast cancer receiving adjuvant or neoadjuvant chemotherapy regimens excluding sequential or combination anthracycline and taxane (106 patients in the scalp cooling group and 16 in the control group; 14 matched by both age and chemotherapy regimen). The study was conducted between August 2013 and October 2014 with ongoing annual follow-up for 5 years. Exposures Use of a scalp cooling system. Scalp cooling was initiated 30 minutes prior to each chemotherapy cycle, with scalp temperature maintained at 3°C (37°F) throughout chemotherapy and for 90 minutes to 120 minutes afterward. Main Outcomes and Measures Self-estimated hair loss using the Dean scale was assessed 4 weeks after the last dose of chemotherapy by unblinded patient review of 5 photographs. A Dean scale score of 0 to 2 (≤50% hair loss) was defined as treatment success. A positive association between scalp cooling and reduced risk of hair loss would be demonstrated if 50% or more of patients in the scalp cooling group achieved treatment success, with the lower bound of the 95% CI greater than 40% of the success proportion. Quality of life was assessed at baseline, at the start of the last chemotherapy cycle, and 1 month later. Median follow-up was 29.5 months. Results Among the 122 patients in the study, the mean age was 53 years (range, 28-77 years); 77.0% were white, 9.0% were black, and 10.7% were Asian; and the mean duration of chemotherapy was 2.3 months (median, 2.1 months). No participants in the scalp cooling group received anthracyclines. Hair loss of 50% or less (Dean score of 0-2) was seen in 67 of 101 patients (66.3%; 95% CI, 56.2%-75.4%) evaluable for alopecia in the scalp cooling group vs 0 of 16 patients (0%) in the control group ( P P = .02). Of the 106 patients in the scalp cooling group, 4 (3.8%) experienced the adverse event of mild headache and 3 (2.8%) discontinued scalp cooling due to feeling cold. Conclusions and Relevance Among women undergoing non–anthracycline-based adjuvant chemotherapy for early-stage breast cancer, the use of scalp cooling vs no scalp cooling was associated with less hair loss at 4 weeks after the last dose of chemotherapy. Further research is needed to assess outcomes after patients receive anthracycline regimens, longer-term measures of alopecia, and adverse effects. Trial Registration clinicaltrials.gov Identifier:NCT01831024

Revised Framingham Stroke Risk Profile to Reflect Temporal Trends.

Abstract: Background: Age-adjusted stroke incidence has decreased over the past 50 years, likely as a result of changes in the prevalence and impact of various stroke risk factors. An updated version of the Framingham Stroke Risk Profile (FSRP) might better predict current risks in the FHS (Framingham Heart Study) and other cohorts. We compared the accuracy of the standard (old) and of a revised (new) version of the FSRP in predicting the risk of all-stroke and ischemic stroke and validated this new FSRP in 2 external cohorts, the 3C (3 Cities) and REGARDS (Reasons for Geographic and Racial Differences in Stroke) studies. Methods: We computed the old FSRP as originally described and a new model that used the most recent epoch-specific risk factor prevalence and hazard ratios for individuals ≥55 years of age and for the subsample ≥65 years of age (to match the age range in REGARDS and 3C studies, respectively) and compared the efficacy of these models in predicting 5- and 10-year stroke risks. Results: The new FSRP was a better predictor of current stroke risks in all 3 samples than the old FSRP (calibration χ2 of new/old FSRP: in men: 64.0/12.1, 59.4/30.6, and 20.7/12.5; in women: 42.5/4.1, 115.4/90.3, and 9.8/6.5 in FHS, REGARDS, and 3C, respectively). In the REGARDS, the new FSRP was a better predictor among whites compared with blacks. Conclusions: A more contemporaneous, new FSRP better predicts current risks in 3 large community samples and could serve as the basis for examining geographic and racial differences in stroke risk and the incremental diagnostic utility of novel stroke risk factors. # Clinical Perspective {#article-title-44}

Tests for the Normal Distribution

Abstract: The chapter deals with the second class of use and discusses goodness-of-fit tests designed to test formally the appropriateness or adequacy of the normal distribution as a model for the underlying phenomenon from which data were generated. The single most used distribution in statistical analysis is the normal distribution. The chapter discusses tests that assume a complete random sample is available for analysis. It purposes tests for normality grouped into five categories, chi-square test, empirical distribution function tests, moment tests, regression tests, and miscellaneous tests. A number of investigators have considered extending and modifying the Shapiro-Wilk test. Spiegelhalter used the theory of most powerful location and scale invariant tests to develop tests of normality against the uniform and the double exponential distributions. Results of power studies are not the only means for judging or comparing the normality tests.

Incident Type 2 Myocardial Infarction in a Cohort of Patients Undergoing Coronary or Peripheral Arterial Angiography

Abstract: Background —Despite growing recognition of Type 2 myocardial infarction (T2MI; related to supply-demand mismatch), little is known about its risk factors or its association with outcome. Methods —A single center cohort of patients undergoing coronary or peripheral angiography with or without intervention was prospectively enrolled and followed for incident type 1 and T2MI, as well as major adverse cardiovascular events (MACE, a composite of all-cause death, non-fatal MI, heart failure, stroke, transient ischemic attack, peripheral arterial complication and cardiac arrhythmia). T2MI was adjudicated using criteria from the Third Universal Definition of MI. Baseline characteristics, blood samples and angiography information were obtained. Major end points subsequent to first MI were assessed using landmark analyses to compare the rates of first events only where everyone with a prior history of any MACE prior to MI were censored and adjusted for follow up times. Cox proportional hazard models were used for time-to-event analyses with age and sex forced into all models and additional covariates evaluated using the stepwise option for the selection. Results —1251 patients were enrolled and followed for a median of 3.4 years. 152 (12.2%) had T2MI during follow up; T2MI was frequently recurrent. Multivariable predictors of T2MI were older age, lower systolic blood pressure, history of coronary artery disease, heart failure, chronic obstructive pulmonary disease, diabetes mellitus, nitrate use, and elevated concentrations of glucose, N-terminal pro-B type natriuretic peptide and cystatin C. Patients with T2MI had higher rates of subsequent adverse events compared with those without T2MI (per 100 person/years: MACE, 53.7 vs. 21.1, p<0.001; all-cause death, 23.3 vs. 3.3, p<0.001; cardiovascular death, 17.5 vs. 2.6, p<0.001; heart failure events: 22.4 vs. 7.4 p<0.001); these rates are similar to those seen in those with Type 1 MI. Incident diagnosis of T2MI strongly predicted risk for subsequent MACE (adjusted HR 1.90, 95% CI=1.46-2.48; p<0.001), all-cause death (adjusted HR 2.96, 95% CI=2.01-4.36; p<0.001) and cardiovascular death (adjusted HR 2.16, 95% CI=1.36-3.43; p=0.001). Conclusions —T2MI is common and associated with poor prognosis. Studies evaluating treatment strategies for management of T2MI are needed. Clinical Trial Registration —https://clinicaltrials.gov/ct2/show/[NCT00842868][1], Clinical Trials.Gov # [NCT00842868][1]. [1]: /lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT00842868&atom=%2Fcirculationaha%2Fearly%2F2016%2F11%2F21%2FCIRCULATIONAHA.116.023052.atom

The Comprehensive Post-Acute Stroke Services (COMPASS) study: Design and methods for a cluster-randomized pragmatic trial

Abstract: Patients discharged home after stroke face significant challenges managing residual neurological deficits, secondary prevention, and pre-existing chronic conditions. Post-discharge care is often fragmented leading to increased healthcare costs, readmissions, and sub-optimal utilization of rehabilitation and community services. The COMprehensive Post-Acute Stroke Services (COMPASS) Study is an ongoing cluster-randomized pragmatic trial to assess the effectiveness of a comprehensive, evidence-based, post-acute care model on patient-centered outcomes. Forty-one hospitals in North Carolina were randomized (as 40 units) to either implement the COMPASS care model or continue their usual care. The recruitment goal is 6000 patients (3000 per arm). Hospital staff ascertain and enroll patients discharged home with a clinical diagnosis of stroke or transient ischemic attack. Patients discharged from intervention hospitals receive 2-day telephone follow-up; a comprehensive clinic visit within 2 weeks that includes a neurological evaluation, assessments of social and functional determinants of health, and an individualized COMPASS Care Plan™ integrated with a community-specific resource database; and additional follow-up calls at 30 and 60 days post-stroke discharge. This model is consistent with the Centers for Medicare and Medicaid Services transitional care management services provided by physicians or advanced practice providers with support from a nurse to conduct patient assessments and coordinate follow-up services. Patients discharged from usual care hospitals represent the control group and receive the standard of care in place at that hospital. Patient-centered outcomes are collected from telephone surveys administered at 90 days. The primary endpoint is patient-reported functional status as measured by the Stroke Impact Scale 16. Secondary outcomes are: caregiver strain, all-cause readmissions, mortality, healthcare utilization, and medication adherence. The study engages patients, caregivers, and other stakeholders (including policymakers, advocacy groups, payers, and local community coalitions) to advise and support the design, implementation, and sustainability of the COMPASS care model. Given the high societal and economic burden of stroke, identifying a care model to improve recovery, independence, and quality of life is critical for stroke survivors and their caregivers. The pragmatic trial design provides a real-world assessment of the COMPASS care model effectiveness and will facilitate rapid implementation into clinical practice if successful. Clinicaltrials.gov: NCT02588664 ; October 23, 2015.

Net reclassification index at event rate: properties and relationships

Abstract: The net reclassification improvement (NRI) is an attractively simple summary measure quantifying improvement in performance because of addition of new risk marker(s) to a prediction model. Originally proposed for settings with well-established classification thresholds, it quickly extended into applications with no thresholds in common use. Here we aim to explore properties of the NRI at event rate. We express this NRI as a difference in performance measures for the new versus old model and show that the quantity underlying this difference is related to several global as well as decision analytic measures of model performance. It maximizes the relative utility (standardized net benefit) across all classification thresholds and can be viewed as the Kolmogorov-Smirnov distance between the distributions of risk among events and non-events. It can be expressed as a special case of the continuous NRI, measuring reclassification from the 'null' model with no predictors. It is also a criterion based on the value of information and quantifies the reduction in expected regret for a given regret function, casting the NRI at event rate as a measure of incremental reduction in expected regret. More generally, we find it informative to present plots of standardized net benefit/relative utility for the new versus old model across the domain of classification thresholds. Then, these plots can be summarized with their maximum values, and the increment in model performance can be described by the NRI at event rate. We provide theoretical examples and a clinical application on the evaluation of prognostic biomarkers for atrial fibrillation. Copyright © 2016 John Wiley & Sons, Ltd.

Association of coronary artery calcium score vs age with cardiovascular risk in older adults: An analysis of pooled population-based studies

Abstract: Importance Besides age, other discriminators of atherosclerotic cardiovascular disease (ASCVD) risk are needed in older adults. Objectives To examine the predictive ability of coronary artery calcium (CAC) score vs age for incident ASCVD and how risk prediction changes by adding CAC score and removing only age from prediction models. Design, Setting, and Participants We conducted an analysis of pooled US population-based studies, including the Framingham Heart Study, the Multi-Ethnic Study of Atherosclerosis, and the Cardiovascular Health Study. Results were compared with 2 European cohorts, the Rotterdam Study and the Heinz Nixdorf Recall Study. Participants underwent CAC scoring between 1998 and 2006 using cardiac computed tomography. The participants included adults older than 60 years without known ASCVD at baseline. Exposures Coronary artery calcium scores. Main Outcomes and Measures Incident ASCVD events including coronary heart disease (CHD) and stroke. Results The study included 4778 participants from 3 US cohorts, with a mean age of 70.1 years; 2582 (54.0%) were women, and 2431 (50.9%) were nonwhite. Over 11 years of follow-up (44 152 person-years), 405 CHD and 228 stroke events occurred. Coronary artery calcium score (vs age) had a greater association with incident CHD (C statistic, 0.733 vs 0.690; C statistics difference, 0.043; 95% CI of difference, 0.009-0.075) and modestly improved prediction of incident stroke (C statistic, 0.695 vs 0.670; C statistics difference, 0.025; 95% CI of difference, −0.015 to 0.064). Adding CAC score to models including traditional cardiovascular risk factors, with only age being removed, provided improved discrimination for incident CHD (C statistic, 0.735 vs 0.703; C statistics difference, 0.032; 95% CI of difference, 0.002-0.062) but not for stroke. Coronary artery calcium score was more likely than age to provide higher category-free net reclassification improvement among participants who experienced an ASCVD event (0.390; 95% CI, 0.312-0.467 vs 0.08; 95% CI −0.001 to 0.181) and to result in more accurate reclassification of risk for ASCVD events among these individuals. The findings were similar in the 2 European cohorts (n = 4990). Conclusions and Relevance Coronary artery calcium may be an alternative marker besides age to better discriminate between lower and higher CHD risk in older adults. Whether CAC score can assist in guiding the decision to initiate statin treatment for primary prevention in older adults requires further investigation.

Coronary Artery Calcium Distribution Is an Independent Predictor of Incident Major Coronary Heart Disease Events: Results From the Framingham Heart Study.

Abstract: Background— The presence and extent of coronary artery calcium (CAC) are associated with increased risk for cardiovascular events. We determined whether information on the distribution of CAC and coronary dominance as detected by cardiac computed tomography were incremental to traditional Agatston score (AS) in predicting incident major coronary heart disease (CHD). Methods and Results— We assessed total AS and the presence of CAC per coronary artery, per segment, and coronary dominance by computed tomography in participants from the offspring and third-generation cohorts of the Framingham Heart Study. The primary outcome was major CHD (myocardial infarction or CHD death). We performed multivariable Cox proportional hazards analysis and calculated relative integrated discrimination improvement. In 1268 subjects (mean age, 56.2±10.3 years, 63.2% men) with AS >0 and no history of major CHD, a total of 42 major CHD events occurred during median follow-up of 7.4 years. The number of coronary arteries with CAC (hazard ratio, 1.68 per artery; 95% confidence interval, 1.10–2.57; P =0.02) and the presence of CAC in the proximal dominant coronary artery (hazard ratio, 2.59; 95% confidence interval, 1.15–5.83; P =0.02) were associated with major CHD events after multivariable adjustment for Framingham risk score and categories of AS. In addition, measures of CAC distribution improved discriminatory capacity for major CHD events (relative integrated discrimination improvement, 0.14). Conclusions— Distribution of coronary atherosclerosis, especially CAC in the proximal dominant coronary artery and an increased number of coronary arteries with CAC, predict major CHD events independently of the traditional AS in community-dwelling men and women.

Automated assessments of circumferential strain from cine CMR correlate with LVEF declines in cancer patients early after receipt of cardio-toxic chemotherapy

Abstract: In patients with cancer receiving potentially cardio-toxic chemotherapy, measurements of left ventricular (LV) circumferential or longitudinal strain are often used clinically to identify myocardial dysfunction. Using a new software algorithm, we sought to determine in individuals receiving treatment for cancer the association between automated assessments of LV mean mid-wall circumferential strain and conventional measures of LV ejection fraction (EF) both obtained from cardiovascular magnetic resonance (CMR) cine balanced steady-state free-precession (bSSFP) white-blood acquisitions. Before and 3 months after initiating treatment with potentially cardio-toxic chemotherapy, 72 individuals (aged 54 ± 14 years with breast cancer [39%], lymphoma [49%], or sarcoma [12%]) underwent serial CMR cine bSSFP assessments of LV volumes and EF, and mean mid-wall circumferential strain determined from these same cine images as well as from additional tagged CMR images. On the cine images, assessments of strain were obtained using the newly developed deformation-based segmentation algorithm. Assessments of LV volumes/EF from the cine images and strain from tagged CMR were accomplished using commercially available software. All measures were analyzed in a blinded fashion independent of one another. Acceptable measures for the automated assessments of mean mid-wall circumferential strain from the cine images were obtained in 142 of 144 visits (98.6%) with an overall analysis time averaging 6:47 ± 1:06 min. The results from these automated measures averaged −18.8 ± 2.9 at baseline and −17.6 ± 3.1 at 3 months (p = 0.001). Left ventricular EF declined slightly from 65 ± 7% at baseline to 62 ± 7% at 3 months (p = 0.0002). The correlation between strain from cine imaging and LVEF was r = −0.61 (p < 0.0001). In addition, the 3-month changes in LV strain and LVEF were correlated (r = −0.49; p < 0.0001). The correlation between cine and tagged derived assessments of strain was r = 0.23; p = 0.01. Automated measures of LV mean mid-wall circumferential strain can be obtained in 6¾ minutes from cine bSSFP LV short-axis images (used concurrently to assess LV volumes and EF) in 98.6% of patients receiving treatment for cancer with potentially cardio-toxic chemotherapy. These cine derived measures of circumferential strain correlate with early subclinical declines in LVEF.

Progressive 3-Month Increase in LV Myocardial ECV After Anthracycline-Based Chemotherapy.

Abstract: Following myocardial injury, the left ventricular (LV) myocardial extracellular matrix (ECM) can undergo abnormal expansion (due to inflammation and interstitial fibrosis) that can be identified with cardiac magnetic resonance (CMR) assessments of extracellular volume fraction (ECV) [(1,2)][1].

Tutorial on statistical considerations on subgroup analysis in confirmatory clinical trials.

Abstract: Clinical trials target patients who are expected to benefit from a new treatment under investigation. However, the magnitude of the treatment benefit, if it exists, often depends on the patient baseline characteristics. It is therefore important to investigate the consistency of the treatment effect across subgroups to ensure a proper interpretation of positive study findings in the overall population. Such assessments can provide guidance on how the treatment should be used. However, great care has to be taken when interpreting consistency results. An observed heterogeneity in treatment effect across subgroups can arise because of chance alone, whereas true heterogeneity may be difficult to detect by standard statistical tests because of their low power. This tutorial considers issues related to subgroup analyses and their impact on the interpretation of findings of completed trials that met their main objectives. In addition, we provide guidance on the design and analysis of clinical trials that account for the expected heterogeneity of treatment effects across subgroups by establishing treatment benefit in a pre-defined targeted subgroup and/or the overall population. Copyright © 2016 John Wiley & Sons, Ltd.

Early Myocardial Strain Changes During Potentially Cardiotoxic Chemotherapy May Occur as a Result of Reductions in Left Ventricular End-Diastolic Volume: The Need to Interpret Left Ventricular Strain With Volumes.

Abstract: Cardiac dysfunction and myocellular injury from cancer therapeutics are identified by reductions in left ventricular (LV) ejection fraction (LVEF) or >15% deteriorations in myocardial strain.1 Myocardial strain may deteriorate as a result of increases in LV end-systolic volume (LVESV), reductions in LV end-diastolic volume (LVEDV), or both. Decreases in LVEDV caused by hypovolemia from poor oral intake, emesis, or myocardial loss2 occur during cancer treatment. We sought to determine the frequency with which decrements in myocardial strain were mediated by decreases in LVEDV versus increases in LVESV in patients receiving potentially cardiotoxic chemotherapy. The study was approved by the local institutional review board, and participants provided witnessed, written informed consent. Cardiovascular magnetic resonance examinations were performed on a 1.5-T Siemens Avanto scanner <6 hours before chemotherapy administration both before and 3 months after the initiation of cancer treatment. LV volumes, LVEF, LV mass, relative wall thickness, and midwall eulerian circumferential strain (ECC) were calculated from a series of LV short-axis white-blood cine stacks and a midcavity short-axis grid-tagged image.3 In addition, global longitudinal strain (GLS) was assessed from high-temporal-resolution 2- and 4-chamber cine views …

Pepsin in saliva as a biomarker for oropharyngeal reflux compared with 24-hour esophageal impedance/pH monitoring in pediatric patients.

Abstract: Background Pepsin in saliva is a proposed biomarker for oropharyngeal reflux. Pepsin may be prevalent in saliva from subjects with gastro-esophageal reflux and may correlate with proximal reflux by intraluminal impedance/pH monitoring (MII/pH). Methods Patients (3 days to 17.6 years, n=90) undergoing 24-hour MII/pH monitoring and asymptomatic controls (2 months to 13.7 years, n=43) were included. Salivary pepsin was determined using a pepsin enzyme-linked immunosorbent assay. Eight saliva samples were collected from patients undergoing 24-hr MII/pH: (i) before catheter placement, (ii) before and 30 minutes after each of three meals, and (iii) upon awakening. One sample was collected from each control. Key Results In MII/pH subjects, 85.6% (77/90) had at least one pepsin-positive sample compared with 9.3% (4/43) in controls. The range of pepsin observed in individual subjects varied widely over 24 hours. The average pepsin concentration in all samples obtained within 2 hours following the most recent reflux event was 30.7±135 ng/mL, decreasing to 16.5±39.1 ng/mL in samples collected more than 2 hours later. The frequency of pepsin-positive samples correlated significantly with symptom index (rS=0.332, P=.0014), proximal (rS=0.340, P=.0010), and distal (rS=0.272, P=.0095) MII events. Conclusions & Inferences Concentration of salivary pepsin may not be an accurate measure of severity of reflux because of the wide range observed in individuals over 24 hours. Saliva samples must be obtained soon after a reflux event. Defining a regimen for optimal saliva collection may help to achieve the goal of using salivary pepsin as a biomarker for oropharyngeal reflux. Clinical Trial registry NCT01091805.

Predictors of Dyslipidemia Over Time in Youth With Type 1 Diabetes: For the SEARCH for Diabetes in Youth Study

Abstract: OBJECTIVE Understanding the risk factors associated with progression and regression of dyslipidemia in youth with type 1 diabetes may guide treatments. RESEARCH DESIGN AND METHODS We studied 1,478 youth with type 1 diabetes (age 10.8 ± 3.9 years, 50% male, 77% non-Hispanic white, not on lipid-lowering medications) at baseline and at a mean follow-up of 7.1 ± 1.9 years in the SEARCH for Diabetes in Youth (SEARCH) study. Progression to dyslipidemia was defined as normal lipid concentrations at baseline and abnormal at follow-up (non–HDL-cholesterol [C] >130 mg/dL or HDL-C RESULTS Non–HDL-C progressed, regressed, was stable normal, and stable abnormal in 19%, 5%, 69%, and 7% of youth with type 1 diabetes, respectively. Corresponding percentages for HDL-C were 3%, 3%, 94%, and 1%, respectively. Factors associated with non–HDL-C progression were higher A1C AUC and higher WHtR AUC in males. Non–HDL-C regression was associated with lower WHtR AUC, and HDL-C progression was associated with male sex and higher WHtR AUC. HDL-C regression was not modeled due to small numbers. CONCLUSIONS A1C and WHtR are modifiable risk factors associated with change in dyslipidemia over time in youth with type 1 diabetes.

Discrimination slope and integrated discrimination improvement - properties, relationships and impact of calibration.

Abstract: Discrimination slope, defined as the slope of a linear regression of predicted probabilities of event derived from a prognostic model on the binary event status, has recently gained popularity as a measure of model performance. It is as a building block for the integrated discrimination improvement that equals the difference in discrimination slopes between the two models being compared. Several authors have pointed out that it does not make sense to apply the integrated discrimination improvement and discrimination slope when working with mis-calibrated models, whereas others have raised concerns about the ability of improving discrimination slope without adding new information. In this paper, we show that under certain assumptions the discrimination slope is asymptotically related to two other R-squared measures, one of which is a rescaled version of the Brier score, known to be proper. Furthermore, we illustrate how a simple recalibration makes the slope equal to the rescaled Brier R-squared metric. We also show that the discrimination slope can be interpreted as a measure of reduction in expected regret for the Gini-Brier regret function. Using theoretical and practical examples, we illustrate how all of these metrics are affected by different levels of model mis-calibration. In particular, we demonstrate that simple recalibration ascertaining calibration in-the-large and calibration slope equal to 1 are not sufficient to correct for some forms of mis-calibration. We conclude that R-squared metrics, including the discrimination slope, offer an attractive choice for quantifying model performance as long as one accounts for their sensitivity to model calibration. Copyright © 2016 John Wiley & Sons, Ltd.

Temporal Associations Between Smoking and Cardiovascular Disease, 1971 to 2006 (from the Framingham Heart Study)

Abstract: Smoking has consistently been related to cardiovascular risk. Public health efforts have yielded reduced smoking prevalence and gains in cardiovascular disease (CVD) prevention. We hypothesized that the contribution of tobacco to CVD risk would be attenuated over prospective decades (1971 to 2006) in a community-based cohort. We evaluated 5,041 Framingham Heart Study Offspring Cohort participants (mean age 36.1 years, 52% women) without prevalent CVD. We collected prospective data on smoking status, relevant CVD risk factors, and incident CVD events across prospective decades. We used multivariable-adjusted, Cox proportional hazard models to measure the effect of smoking on incident CVD over 3 prospective 12-year follow-up periods. Our results demonstrated a consistent twofold increased risk of CVD in men who smoke compared with nonsmokers for each 12-year time period spanning from 1971 to 2006. Women who smoked had a 1.5-fold increased CVD risk. Smoking remains an important risk factor despite substantial improvements in the prevention and treatment of CVD. Significant, contemporary improvements in CVD prevention—such as gains in hypertension and cholesterol treatment—have not attenuated the strong and persistent associations between smoking and CVD observed here. In conclusion, our results highlight the importance of continued public health efforts to address smoking as a modifiable exposure that strongly contributes toward CVD risk.

Frequency of Left Ventricular End-Diastolic Volume-Mediated Declines in Ejection Fraction in Patients Receiving Potentially Cardiotoxic Cancer Treatment.

Abstract: We sought to determine the frequency by which decreases in left ventricular (LV) end-diastolic volume (LVEDV) with and without increases in end-systolic volume (LVESV) influenced early cancer treatment-associated declines in LV ejection fraction (LVEF) or LV mass. One hundred twelve consecutively recruited subjects (aged 52 ± 14 years) with cancer underwent blinded cardiovascular magnetic resonance measurements of LV volumes, mass, and LVEF before and 3 months after initiating potentially cardiotoxic chemotherapy (72% of participants received anthracyclines). Twenty-six participants developed important declines in LVEF of >10% or to values

Circulating mutational portrait of cancer: manifestation of aggressive clonal events in both early and late stages

Abstract: Solid tumors residing in tissues and organs leave footprints in circulation through circulating tumor cells (CTCs) and circulating tumor DNAs (ctDNA). Characterization of the ctDNA portraits and comparison with tumor DNA mutational portraits may reveal clinically actionable information on solid tumors that is traditionally achieved through more invasive approaches. We isolated ctDNAs from plasma of patients of 103 lung cancer and 74 other solid tumors of different tissue origins. Deep sequencing using the Guardant360 test was performed to identify mutations in 73 clinically actionable genes, and the results were associated with clinical characteristics of the patient. The mutation profiles of 37 lung cancer cases with paired ctDNA and tumor genomic DNA sequencing were used to evaluate clonal representation of tumor in circulation. Five lung cancer cases with longitudinal ctDNA sampling were monitored for cancer progression or response to treatments. Mutations in TP53, EGFR, and KRAS genes are most prevalent in our cohort. Mutation rates of ctDNA are similar in early (I and II) and late stage (III and IV) cancers. Mutation in DNA repair genes BRCA1, BRCA2, and ATM are found in 18.1% (32/177) of cases. Patients with higher mutation rates had significantly higher mortality rates. Lung cancer of never smokers exhibited significantly higher ctDNA mutation rates as well as higher EGFR and ERBB2 mutations than ever smokers. Comparative analysis of ctDNA and tumor DNA mutation data from the same patients showed that key driver mutations could be detected in plasma even when they were present at a minor clonal population in the tumor. Mutations of key genes found in the tumor tissue could remain in circulation even after frontline radiotherapy and chemotherapy suggesting these mutations represented resistance mechanisms. Longitudinal sampling of five lung cancer cases showed distinct changes in ctDNA mutation portraits that are consistent with cancer progression or response to EGFR drug treatment. This study demonstrates that ctDNA mutation rates in the key tumor-associated genes are clinical parameters relevant to smoking status and mortality. Mutations in ctDNA may serve as an early detection tool for cancer. This study quantitatively confirms the hypothesis that ctDNAs in circulation is the result of dissemination of aggressive tumor clones and survival of resistant clones. This study supports the use of ctDNA profiling as a less-invasive approach to monitor cancer progression and selection of appropriate drugs during cancer evolution.

Increased Aortic Diameters on Multidetector Computed Tomographic Scan Are Independent Predictors of Incident Adverse Cardiovascular Events: The Framingham Heart Study.

Abstract: Background—Adverse aortic remodeling, such as dilation, is associated with multiple cardiovascular disease (CVD) risk factors. We sought to determine whether measures of enlarged aortic diameters i...

Relation of Risk Factors and Abdominal Aortic Calcium to Progression of Coronary Artery Calcium (from the Framingham Heart Study)

Abstract: Coronary artery calcium (CAC) and abdominal aortic calcium (AAC) on multidetector computed tomography (MDCT) permit assessment of the presence and burden of coronary and systemic atherosclerosis. Risk factors for progression of CAC and AAC and the association of AAC with CAC progression have not been well characterized in a community-dwelling cohort. We studied 1,959 asymptomatic participants from the Framingham Heart Study who underwent serial MDCT scans with a median interval of 6.1 years. Primary outcomes were (a) the incidence of CAC and AAC (CAC >0 and AAC >0 with baseline CAC = 0 and AAC = 0) and (b) absolute progression of CAC (CAC > baseline CAC and AAC > baseline AAC). Covariates were collected at adjacent cycle examinations and included age, gender, use of antihypertensive therapy, use of lipid-lowering therapy, cigarette smoking, and total and high-density lipoprotein cholesterol. Predictors for CAC and AAC progression included baseline CAC, baseline AAC, lipid-lowering therapy, diabetes, high-density lipoprotein cholesterol, BMI, and serum creatinine. Multivariable stepwise logistic and linear regression models were used to test the association of these risk factors with CAC and AAC. Those who developed incident CAC on follow-up scanning comprised 18.8% of 1,124 participants, and 84.9% of 780 participants, with detectable baseline CAC, had further progression. Baseline AAC was a predictor of both CAC incidence and progression, independent of other risk factors. In stepwise models, addition of baseline AAC slightly improved the area under the curve from 0.72 (0.68 to 0.76) to 0.74 (0.70 to 0.78). In conclusion, standard cardiovascular disease risk factors are associated with incidence and progression of CAC and AAC, and AAC augments CAC incidence and progression above cardiovascular disease risk factors.

Association of Multiorgan Computed Tomographic Phenomap With Adverse Cardiovascular Health Outcomes: The Framingham Heart Study

Abstract: Importance Increased ability to quantify anatomical phenotypes across multiple organs provides the opportunity to assess their cumulative ability to identify individuals at greatest susceptibility for adverse outcomes. Objective To apply unsupervised machine learning to define the distribution and prognostic importance of computed tomography–based multiorgan phenotypes associated with adverse health outcomes. Design, Setting, and Participants This asymptomatic community-based cohort study included 2924 Framingham Heart Study participants between July 2002 and April 2005 undergoing computed tomographic imaging of the chest and abdomen. Participants are from the offspring and third-generation cohorts. Exposures Eleven computed tomography–based measures of valvular/vascular calcification, adiposity, and muscle attenuation. Main Outcomes and Measures All-cause mortality and cardiovascular disease (myocardial infarction, stroke, or cardiovascular death). Results The median age of the participants was 50 years (interquartile range, 43-60 years), and 1422 (48.6%) were men. Principal component analysis identified 3 major anatomic axes: (1) global calcification (defined by aortic, thoracic, coronary, and valvular calcification); (2) adiposity (defined by pericardial, visceral, hepatic, and intrathoracic fat); and (3) muscle attenuation that explained 65.7% of the population variation. Principal components showed different evolution with age (continuous increase in global calcification, decrease in muscle attenuation, and U-shaped association with adiposity) but similar patterns in men and women. Using unsupervised clustering approaches in the offspring cohort (n = 1150), we identified a cohort (n = 232; 20.2%) with an unfavorable multiorgan phenotype across all 3 anatomic axes as compared with a favorable multiorgan phenotype. Membership in the unfavorable phenotypic cluster was associated with a greater prevalence of cardiovascular disease risk factors and with increased all-cause mortality (hazard ratio, 2.61; 95% CI, 1.74-3.92; P Conclusions and Relevance This proof-of-concept analysis demonstrates that unsupervised machine learning, in an asymptomatic community cohort, identifies an unfavorable multiorgan phenotype associated with adverse health outcomes, especially in elderly American adults. Future investigations in larger populations are required not only to validate the present results, but also to harness clinical, biochemical, imaging, and genetic markers to increase our understanding of healthy cardiovascular aging.

Guideline-Based Statin Eligibility, Cancer Events, and Noncardiovascular Mortality in the Framingham Heart Study.

Abstract: Purpose Cancer and cardiovascular disease share risk factors, and there is some evidence that statins reduce cancer mortality. We sought to determine the accuracy of the 2013 American College of Cardiology/American Heart Association statin eligibility criteria to identify individuals at a higher risk of developing cancer or of dying as a result of cancer or other noncardiovascular causes. Methods We included 2,196 participants (50.5 ± 8.1 years of age; 55% female) who were statin naive and free of cancer at baseline from the offspring and third-generation cohorts of the community-based longitudinal Framingham Heart Study. Statin eligibility was determined per American College of Cardiology/American Heart Association guidelines, and subclinical coronary atherosclerosis was assessed by computed tomography. The primary outcome was incident cancer at a median of 10.0 years (interquartile range, 9.1-10.6 years) of follow-up, and secondary outcomes were cancer mortality and noncardiovascular mortality. Results The incident cancer rate was 11.2% (247 of 2,196), with 58 noncardiovascular deaths, including 39 cancer deaths (1.8%). Overall, 37% (812 of 2,196) were statin eligible. Incident cancer occurred in 125 (15%) of the 812 statin-eligible participants versus 122 (8.8%) of the 1,384 of noneligible participants (subdistribution hazard ratio [SDHR], 1.8 [1.4 to 2.3]; P < .001). Cancer mortality occurred in 34 (4.2%) of the 812 statin-eligible participants versus five (0.4%) of the 1,384 noneligible participants (SDHR, 12.1 [4.7 to 31]; P < .001). Noncardiovascular mortality occurred in 49 (6.0%) of the 812 statin-eligible participants versus nine (0.7%) of the 1,384 noneligible participants (SDHR, 10.1 [5.0 to 21]; P < .001). In stratified analyses, these findings were independent of any individual causative risk factor such as body mass index, age, or smoking status. Conclusion In this community-based primary prevention cohort, guideline-based statin eligibility accurately identified patients at a higher risk of developing cancer and cancer-related mortality. Shared risk profiles and potential benefits of statins between cancer and cardiovascular outcomes may provide a unique opportunity to improve population health.

Asymptotic distribution of ∆AUC, NRIs, and IDI based on theory of U‐statistics

Abstract: The change in area under the curve (∆AUC), the integrated discrimination improvement (IDI), and net reclassification index (NRI) are commonly used measures of risk prediction model performance. Some authors have reported good validity of associated methods of estimating their standard errors (SE) and construction of confidence intervals, whereas others have questioned their performance. To address these issues, we unite the ∆AUC, IDI, and three versions of the NRI under the umbrella of the U-statistics family. We rigorously show that the asymptotic behavior of ∆AUC, NRIs, and IDI fits the asymptotic distribution theory developed for U-statistics. We prove that the ∆AUC, NRIs, and IDI are asymptotically normal, unless they compare nested models under the null hypothesis. In the latter case, asymptotic normality and existing SE estimates cannot be applied to ∆AUC, NRIs, or IDI. In the former case, SE formulas proposed in the literature are equivalent to SE formulas obtained from U-statistics theory if we ignore adjustment for estimated parameters. We use Sukhatme-Randles-deWet condition to determine when adjustment for estimated parameters is necessary. We show that adjustment is not necessary for SEs of the ∆AUC and two versions of the NRI when added predictor variables are significant and normally distributed. The SEs of the IDI and three-category NRI should always be adjusted for estimated parameters. These results allow us to define when existing formulas for SE estimates can be used and when resampling methods such as the bootstrap should be used instead when comparing nested models. We also use the U-statistic theory to develop a new SE estimate of ∆AUC. Copyright © 2017 John Wiley & Sons, Ltd.