Resistance prediction in AML: analysis of 4601 patients from MRC/NCRI, HOVON/SAKK, SWOG and MD Anderson Cancer Center

Therapeutic resistance remains the principal problem in acute myeloid leukemia (AML). We used area under receiver-operating characteristic curves (AUCs) to quantify our ability to predict therapeutic resistance in individual patients, where AUC=1.0 denotes perfect prediction and AUC=0.5 denotes a coin flip, using data from 4601 patients with newly diagnosed AML given induction therapy with 3+7 or more intense standard regimens in UK Medical Research Council/National Cancer Research Institute, Dutch–Belgian Cooperative Trial Group for Hematology/Oncology/Swiss Group for Clinical Cancer Research, US cooperative group SWOG and MD Anderson Cancer Center studies. Age, performance status, white blood cell count, secondary disease, cytogenetic risk and FLT3-ITD/NPM1 mutation status were each independently associated with failure to achieve complete remission despite no early death (‘primary refractoriness’). However, the AUC of a bootstrap-corrected multivariable model predicting this outcome was only 0.78, indicating only fair predictive ability. Removal of FLT3-ITD and NPM1 information only slightly decreased the AUC (0.76). Prediction of resistance, defined as primary refractoriness or short relapse-free survival, was even more difficult. Our limited ability to forecast resistance based on routinely available pretreatment covariates provides a rationale for continued randomization between standard and new therapies and supports further examination of genetic and posttreatment data to optimize resistance prediction in AML.

General distress, hopelessness—suicidal ideation and worrying in adolescence: Concurrent and predictive validity of a symptom-level bifactor model for clinical diagnoses

BACKGROUND: Clinical disorders often share common symptoms and aetiological factors. Bifactor models acknowledge the role of an underlying general distress component and more specific sub-domains of psychopathology which specify the unique components of disorders over and above a general factor. METHODS: A bifactor model jointly calibrated data on subjective distress from The Mood and Feelings Questionnaire and the Revised Children's Manifest Anxiety Scale. The bifactor model encompassed a general distress factor, and specific factors for (a) hopelessness-suicidal ideation, (b) generalised worrying and (c) restlessness-fatigue at age 14 which were related to lifetime clinical diagnoses established by interviews at ages 14 (concurrent validity) and current diagnoses at 17 years (predictive validity) in a British population sample of 1159 adolescents. RESULTS: Diagnostic interviews confirmed the validity of a symptom-level bifactor model. The underlying general distress factor was a powerful but non-specific predictor of affective, anxiety and behaviour disorders. The specific factors for hopelessness-suicidal ideation and generalised worrying contributed to predictive specificity. Hopelessness-suicidal ideation predicted concurrent and future affective disorder; generalised worrying predicted concurrent and future anxiety, specifically concurrent generalised anxiety disorders. Generalised worrying was negatively associated with behaviour disorders. LIMITATIONS: The analyses of gender differences and the prediction of specific disorders was limited due to a low frequency of disorders other than depression. CONCLUSIONS: The bifactor model was able to differentiate concurrent and predict future clinical diagnoses. This can inform the development of targeted as well as non-specific interventions for prevention and treatment of different disorders.

Multi-Model Data Fusion to Improve an Early Warning System for Hypo-/Hyperglycemic Events

Correct predictions of future blood glucose levels in individuals with Type 1 Diabetes (T1D) can be used to provide early warning of upcoming hypo-/hyperglycemic events and thus to improve the patient's safety. To increase prediction accuracy and efficiency, various approaches have been proposed which combine multiple predictors to produce superior results compared to single predictors. Three methods for model fusion are presented and comparatively assessed. Data from 23 T1D subjects under sensor-augmented pump (SAP) therapy were used in two adaptive data-driven models (an autoregressive model with output correction - cARX, and a recurrent neural network - RNN). Data fusion techniques based on i) Dempster-Shafer Evidential Theory (DST), ii) Genetic Algorithms (GA), and iii) Genetic Programming (GP) were used to merge the complimentary performances of the prediction models. The fused output is used in a warning algorithm to issue alarms of upcoming hypo-/hyperglycemic events. The fusion schemes showed improved performance with lower root mean square errors, lower time lags, and higher correlation. In the warning algorithm, median daily false alarms (DFA) of 0.25%, and 100% correct alarms (CA) were obtained for both event types. The detection times (DT) before occurrence of events were 13.0 and 12.1 min respectively for hypo-/hyperglycemic events. Compared to the cARX and RNN models, and a linear fusion of the two, the proposed fusion schemes represents a significant improvement.

Experimental validation of a nonlinear μ FE model based on cohesive-frictional plasticity for trabecular bone

Trabecular bone is a porous mineralized tissue playing a major load bearing role in the human body. Prediction of age-related and disease-related fractures and the behavior of bone implant systems needs a thorough understanding of its structure-mechanical property relationships, which can be obtained using microcomputed tomography-based finite element modeling. In this study, a nonlinear model for trabecular bone as a cohesive-frictional material was implemented in a large-scale computational framework and validated by comparison of μFE simulations with experimental tests in uniaxial tension and compression. A good correspondence of stiffness and yield points between simulations and experiments was found for a wide range of bone volume fraction and degree of anisotropy in both tension and compression using a non-calibrated, average set of material parameters. These results demonstrate the ability of the model to capture the effects leading to failure of bone for three anatomical sites and several donors, which may be used to determine the apparent behavior of trabecular bone and its evolution with age, disease, and treatment in the future.

Comparison of co-located independent ground-based middle atmospheric wind and temperature measurements with numerical weather prediction models

High-resolution, ground-based and independent observations including co-located wind radiometer, lidar stations, and infrasound instruments are used to evaluate the accuracy of general circulation models and data-constrained assimilation systems in the middle atmosphere at northern hemisphere midlatitudes. Systematic comparisons between observations, the European Centre for Medium-Range Weather Forecasts (ECMWF) operational analyses including the recent Integrated Forecast System cycles 38r1 and 38r2, the NASA’s Modern-Era Retrospective Analysis for Research and Applications (MERRA) reanalyses, and the free-running climate Max Planck Institute–Earth System Model–Low Resolution (MPI-ESM-LR) are carried out in both temporal and spectral dom ains. We find that ECMWF and MERRA are broadly consistent with lidar and wind radiometer measurements up to ~40 km. For both temperature and horizontal wind components, deviations increase with altitude as the assimilated observations become sparser. Between 40 and 60 km altitude, the standard deviation of the mean difference exceeds 5 K for the temperature and 20 m/s for the zonal wind. The largest deviations are observed in winter when the variability from large-scale planetary waves dominates. Between lidar data and MPI-ESM-LR, there is an overall agreement in spectral amplitude down to 15–20 days. At shorter time scales, the variability is lacking in the model by ~10 dB. Infrasound observations indicate a general good agreement with ECWMF wind and temperature products. As such, this study demonstrates the potential of the infrastructure of the Atmospheric Dynamics Research Infrastructure in Europe project that integrates various measurements and provides a quantitative understanding of stratosphere-troposphere dynamical coupling for numerical weather prediction applications.

Clinical Validation of a Peritoneal Dialysis Prescription Model in the PatientOnLine Software

Peritoneal transport characteristics and residual renal function require regular control and subsequent adjustment of the peritoneal dialysis (PD) prescription. Prescription models shall facilitate the prediction of the outcome of such adaptations for a given patient. In the present study, the prescription model implemented in the PatientOnLine software was validated in patients requiring a prescription change. This multicenter, international prospective cohort study with the aim to validate a PD prescription model included patients treated with continuous ambulatory peritoneal dialysis. Patients were examined with the peritoneal function test (PFT) to determine the outcome of their current prescription and the necessity for a prescription change. For these patients, a new prescription was modeled using the PatientOnLine software (Fresenius Medical Care, Bad Homburg, Germany). Two to four weeks after implementation of the new PD regimen, a second PFT was performed. The validation of the prescription model included 54 patients. Predicted and measured peritoneal Kt/V were 1.52 ± 0.31 and 1.66 ± 0.35, and total (peritoneal + renal) Kt/V values were 1.96 ± 0.48 and 2.06 ± 0.44, respectively. Predicted and measured peritoneal creatinine clearances were 42.9 ± 8.6 and 43.0 ± 8.8 L/1.73 m2/week and total creatinine clearances were 65.3 ± 26.0 and 63.3 ± 21.8 L/1.73 m2/week, respectively. The analysis revealed a Pearson's correlation coefficient for peritoneal Kt/V of 0.911 and Lin's concordance coefficient of 0.829. The value of both coefficients was 0.853 for peritoneal creatinine clearance. Predicted and measured daily net ultrafiltration was 0.77 ± 0.49 and 1.16 ± 0.63 L/24 h, respectively. Pearson's correlation and Lin's concordance coefficient were 0.518 and 0.402, respectively. Predicted and measured peritoneal glucose absorption was 125.8 ± 38.8 and 79.9 ± 30.7 g/24 h, respectively, and Pearson's correlation and Lin's concordance coefficient were 0.914 and 0.477, respectively. With good predictability of peritoneal Kt/V and creatinine clearance, the present model provides support for individual dialysis prescription in clinical practice. Peritoneal glucose absorption and ultrafiltration are less predictable and are likely to be influenced by additional clinical factors to be taken into consideration.

Privacy-preserving genomic testing in the clinic: a model using HIV treatment

PURPOSE The implementation of genomic-based medicine is hindered by unresolved questions regarding data privacy and delivery of interpreted results to health-care practitioners. We used DNA-based prediction of HIV-related outcomes as a model to explore critical issues in clinical genomics. METHODS We genotyped 4,149 markers in HIV-positive individuals. Variants allowed for prediction of 17 traits relevant to HIV medical care, inference of patient ancestry, and imputation of human leukocyte antigen (HLA) types. Genetic data were processed under a privacy-preserving framework using homomorphic encryption, and clinical reports describing potentially actionable results were delivered to health-care providers. RESULTS A total of 230 patients were included in the study. We demonstrated the feasibility of encrypting a large number of genetic markers, inferring patient ancestry, computing monogenic and polygenic trait risks, and reporting results under privacy-preserving conditions. The average execution time of a multimarker test on encrypted data was 865 ms on a standard computer. The proportion of tests returning potentially actionable genetic results ranged from 0 to 54%. CONCLUSIONS The model of implementation presented herein informs on strategies to deliver genomic test results for clinical care. Data encryption to ensure privacy helps to build patient trust, a key requirement on the road to genomic-based medicine.Genet Med advance online publication 14 January 2016Genetics in Medicine (2016); doi:10.1038/gim.2015.167.

Fast Prediction of Femoral Biomechanics Using Supervised Machine Learning and Statistical Shape Modeling

Finite element (FE) analysis is an important computational tool in biomechanics. However, its adoption into clinical practice has been hampered by its computational complexity and required high technical competences for clinicians. In this paper we propose a supervised learning approach to predict the outcome of the FE analysis. We demonstrate our approach on clinical CT and X-ray femur images for FE predictions ( FEP), with features extracted, respectively, from a statistical shape model and from 2D-based morphometric and density information. Using leave-one-out experiments and sensitivity analysis, comprising a database of 89 clinical cases, our method is capable of predicting the distribution of stress values for a walking loading condition with an average correlation coefficient of 0.984 and 0.976, for CT and X-ray images, respectively. These findings suggest that supervised learning approaches have the potential to leverage the clinical integration of mechanical simulations for the treatment of musculoskeletal conditions.

A meta-analysis of trabecular bone score in fracture risk prediction and its relationship to FRAX

Trabecular bone score (TBS) is a grey-level textural index of bone microarchitecture derived from lumbar spine dual-energy X-ray absorptiometry (DXA) images. TBS is a BMD-independent predictor of fracture risk. The objective of this meta-analysis was to determine whether TBS predicted fracture risk independently of FRAX probability and to examine their combined performance by adjusting the FRAX probability for TBS. We utilized individual level data from 17,809 men and women in 14 prospective population-based cohorts. Baseline evaluation included TBS and the FRAX risk variables and outcomes during follow up (mean 6.7 years) comprised major osteoporotic fractures. The association between TBS, FRAX probabilities and the risk of fracture was examined using an extension of the Poisson regression model in each cohort and for each sex and expressed as the gradient of risk (GR; hazard ratio per 1SD change in risk variable in direction of increased risk). FRAX probabilities were adjusted for TBS using an adjustment factor derived from an independent cohort (the Manitoba Bone Density Cohort). Overall, the GR of TBS for major osteoporotic fracture was 1.44 (95% CI: 1.35-1.53) when adjusted for age and time since baseline and was similar in men and women (p > 0.10). When additionally adjusted for FRAX 10-year probability of major osteoporotic fracture, TBS remained a significant, independent predictor for fracture (GR 1.32, 95%CI: 1.24-1.41). The adjustment of FRAX probability for TBS resulted in a small increase in the GR (1.76, 95%CI: 1.65, 1.87 vs. 1.70, 95%CI: 1.60-1.81). A smaller change in GR for hip fracture was observed (FRAX hip fracture probability GR 2.25 vs. 2.22). TBS is a significant predictor of fracture risk independently of FRAX. The findings support the use of TBS as a potential adjustment for FRAX probability, though the impact of the adjustment remains to be determined in the context of clinical assessment guidelines. This article is protected by copyright. All rights reserved.

Incremental Value of the CRUSADE, ACUITY, and HAS-BLED Risk Scores for the Prediction of Hemorrhagic Events After Coronary Stent Implantation in Patients Undergoing Long or Short Duration of Dual Antiplatelet Therapy

BACKGROUND Multiple scores have been proposed to stratify bleeding risk, but their value to guide dual antiplatelet therapy duration has never been appraised. We compared the performance of the CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines), ACUITY (Acute Catheterization and Urgent Intervention Triage Strategy), and HAS-BLED (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile INR, Elderly, Drugs/Alcohol Concomitantly) scores in 1946 patients recruited in the Prolonging Dual Antiplatelet Treatment After Grading Stent-Induced Intimal Hyperplasia Study (PRODIGY) and assessed hemorrhagic and ischemic events in the 24- and 6-month dual antiplatelet therapy groups. METHODS AND RESULTS Bleeding score performance was assessed with a Cox regression model and C statistics. Discriminative and reclassification power was assessed with net reclassification improvement and integrated discrimination improvement. The C statistic was similar between the CRUSADE score (area under the curve 0.71) and ACUITY (area under the curve 0.68), and higher than HAS-BLED (area under the curve 0.63). CRUSADE, but not ACUITY, improved reclassification (net reclassification index 0.39, P=0.005) and discrimination (integrated discrimination improvement index 0.0083, P=0.021) of major bleeding compared with HAS-BLED. Major bleeding and transfusions were higher in the 24- versus 6-month dual antiplatelet therapy groups in patients with a CRUSADE score >40 (hazard ratio for bleeding 2.69, P=0.035; hazard ratio for transfusions 4.65, P=0.009) but not in those with CRUSADE score ≤40 (hazard ratio for bleeding 1.50, P=0.25; hazard ratio for transfusions 1.37, P=0.44), with positive interaction (Pint=0.05 and Pint=0.01, respectively). The number of patients with high CRUSADE scores needed to treat for harm for major bleeding and transfusion were 17 and 15, respectively, with 24-month rather than 6-month dual antiplatelet therapy; corresponding figures in the overall population were 67 and 71, respectively. CONCLUSIONS Our analysis suggests that the CRUSADE score predicts major bleeding similarly to ACUITY and better than HAS BLED in an all-comer population with percutaneous coronary intervention and potentially identifies patients at higher risk of hemorrhagic complications when treated with a long-term dual antiplatelet therapy regimen. CLINICAL TRIAL REGISTRATION URL: http://clinicaltrials.gov. Unique identifier: NCT00611286.

Prediction of improvement in skin fibrosis in diffuse cutaneous systemic sclerosis: a EUSTAR analysis

OBJECTIVES Improvement of skin fibrosis is part of the natural course of diffuse cutaneous systemic sclerosis (dcSSc). Recognising those patients most likely to improve could help tailoring clinical management and cohort enrichment for clinical trials. In this study, we aimed to identify predictors for improvement of skin fibrosis in patients with dcSSc. METHODS We performed a longitudinal analysis of the European Scleroderma Trials And Research (EUSTAR) registry including patients with dcSSc, fulfilling American College of Rheumatology criteria, baseline modified Rodnan skin score (mRSS) ≥7 and follow-up mRSS at 12±2 months. The primary outcome was skin improvement (decrease in mRSS of >5 points and ≥25%) at 1 year follow-up. A respective increase in mRSS was considered progression. Candidate predictors for skin improvement were selected by expert opinion and logistic regression with bootstrap validation was applied. RESULTS From the 919 patients included, 218 (24%) improved and 95 (10%) progressed. Eleven candidate predictors for skin improvement were analysed. The final model identified high baseline mRSS and absence of tendon friction rubs as independent predictors of skin improvement. The baseline mRSS was the strongest predictor of skin improvement, independent of disease duration. An upper threshold between 18 and 25 performed best in enriching for progressors over regressors. CONCLUSIONS Patients with advanced skin fibrosis at baseline and absence of tendon friction rubs are more likely to regress in the next year than patients with milder skin fibrosis. These evidence-based data can be implemented in clinical trial design to minimise the inclusion of patients who would regress under standard of care.

Prediction of Benthic Impact for Salmon Net-Pens Based on the Balance of Benthic Oxygen Supply and Demand

The ratio between oxygen supply and oxygen demand was examined as a predictor of benthic response to organic enrichment caused by salmon net-pen aquaculture. Oxygen supply to the benthos was calculated based on Fickian diffusion and near-bottom flow velocities. A strong linear correlation was found between measured carbon sedimentation rates and rates of benthic metabolism. This relationship allowed an estimation of oxygen demand based on sedimentation rates. Comparison of several production sites in Maine (USA) coastal waters showed that for sites where oxygen demand exceeded supply benthic impacts were high and for sites where oxygen supply exceeded demand benthic impacts were low. These findings were summarized in the form of a predictive model that should be useful in siting salmon production facilities.

Simulation study of joint transition and Weibull survival model with shared parameters

This study investigates a theoretical model where a longitudinal process, that is a stationary Markov-Chain, and a Weibull survival process share a bivariate random effect. Furthermore, a Quality-of-Life adjusted survival is calculated as the weighted sum of survival time. Theoretical values of population mean adjusted survival of the described model are computed numerically. The parameters of the bivariate random effect do significantly affect theoretical values of population mean. Maximum-Likelihood and Bayesian methods are applied on simulated data to estimate the model parameters. Based on the parameter estimates, predicated population mean adjusted survival can then be calculated numerically and compared with the theoretical values. Bayesian method and Maximum-Likelihood method provide parameter estimations and population mean prediction with comparable accuracy; however Bayesian method suffers from poor convergence due to autocorrelation and inter-variable correlation. ^

Selection bias and the cross-validation of regression models for prediction

Strategies are compared for the development of a linear regression model with stochastic (multivariate normal) regressor variables and the subsequent assessment of its predictive ability. Bias and mean squared error of four estimators of predictive performance are evaluated in simulated samples of 32 population correlation matrices. Models including all of the available predictors are compared with those obtained using selected subsets. The subset selection procedures investigated include two stopping rules, C$\sb{\rm p}$ and S$\sb{\rm p}$, each combined with an 'all possible subsets' or 'forward selection' of variables. The estimators of performance utilized include parametric (MSEP$\sb{\rm m}$) and non-parametric (PRESS) assessments in the entire sample, and two data splitting estimates restricted to a random or balanced (Snee's DUPLEX) 'validation' half sample. The simulations were performed as a designed experiment, with population correlation matrices representing a broad range of data structures.^ The techniques examined for subset selection do not generally result in improved predictions relative to the full model. Approaches using 'forward selection' result in slightly smaller prediction errors and less biased estimators of predictive accuracy than 'all possible subsets' approaches but no differences are detected between the performances of C$\sb{\rm p}$ and S$\sb{\rm p}$. In every case, prediction errors of models obtained by subset selection in either of the half splits exceed those obtained using all predictors and the entire sample.^ Only the random split estimator is conditionally (on $\\beta$) unbiased, however MSEP$\sb{\rm m}$ is unbiased on average and PRESS is nearly so in unselected (fixed form) models. When subset selection techniques are used, MSEP$\sb{\rm m}$ and PRESS always underestimate prediction errors, by as much as 27 percent (on average) in small samples. Despite their bias, the mean squared errors (MSE) of these estimators are at least 30 percent less than that of the unbiased random split estimator. The DUPLEX split estimator suffers from large MSE as well as bias, and seems of little value within the context of stochastic regressor variables.^ To maximize predictive accuracy while retaining a reliable estimate of that accuracy, it is recommended that the entire sample be used for model development, and a leave-one-out statistic (e.g. PRESS) be used for assessment. ^

Bayesian and survival model for tumor relapse status and disease-specific survival, with applications to breast cancer

Breast cancer is the most common non-skin cancer and the second leading cause of cancer-related death in women in the United States. Studies on ipsilateral breast tumor relapse (IBTR) status and disease-specific survival will help guide clinic treatment and predict patient prognosis.^ After breast conservation therapy, patients with breast cancer may experience breast tumor relapse. This relapse is classified into two distinct types: true local recurrence (TR) and new ipsilateral primary tumor (NP). However, the methods used to classify the relapse types are imperfect and are prone to misclassification. In addition, some observed survival data (e.g., time to relapse and time from relapse to death)are strongly correlated with relapse types. The first part of this dissertation presents a Bayesian approach to (1) modeling the potentially misclassified relapse status and the correlated survival information, (2) estimating the sensitivity and specificity of the diagnostic methods, and (3) quantify the covariate effects on event probabilities. A shared frailty was used to account for the within-subject correlation between survival times. The inference was conducted using a Bayesian framework via Markov Chain Monte Carlo simulation implemented in softwareWinBUGS. Simulation was used to validate the Bayesian method and assess its frequentist properties. The new model has two important innovations: (1) it utilizes the additional survival times correlated with the relapse status to improve the parameter estimation, and (2) it provides tools to address the correlation between the two diagnostic methods conditional to the true relapse types.^ Prediction of patients at highest risk for IBTR after local excision of ductal carcinoma in situ (DCIS) remains a clinical concern. The goals of the second part of this dissertation were to evaluate a published nomogram from Memorial Sloan-Kettering Cancer Center, to determine the risk of IBTR in patients with DCIS treated with local excision, and to determine whether there is a subset of patients at low risk of IBTR. Patients who had undergone local excision from 1990 through 2007 at MD Anderson Cancer Center with a final diagnosis of DCIS (n=794) were included in this part. Clinicopathologic factors and the performance of the Memorial Sloan-Kettering Cancer Center nomogram for prediction of IBTR were assessed for 734 patients with complete data. Nomogram for prediction of 5- and 10-year IBTR probabilities were found to demonstrate imperfect calibration and discrimination, with an area under the receiver operating characteristic curve of .63 and a concordance index of .63. In conclusion, predictive models for IBTR in DCIS patients treated with local excision are imperfect. Our current ability to accurately predict recurrence based on clinical parameters is limited.^ The American Joint Committee on Cancer (AJCC) staging of breast cancer is widely used to determine prognosis, yet survival within each AJCC stage shows wide variation and remains unpredictable. For the third part of this dissertation, biologic markers were hypothesized to be responsible for some of this variation, and the addition of biologic markers to current AJCC staging were examined for possibly provide improved prognostication. The initial cohort included patients treated with surgery as first intervention at MDACC from 1997 to 2006. Cox proportional hazards models were used to create prognostic scoring systems. AJCC pathologic staging parameters and biologic tumor markers were investigated to devise the scoring systems. Surveillance Epidemiology and End Results (SEER) data was used as the external cohort to validate the scoring systems. Binary indicators for pathologic stage (PS), estrogen receptor status (E), and tumor grade (G) were summed to create PS+EG scoring systems devised to predict 5-year patient outcomes. These scoring systems facilitated separation of the study population into more refined subgroups than the current AJCC staging system. The ability of the PS+EG score to stratify outcomes was confirmed in both internal and external validation cohorts. The current study proposes and validates a new staging system by incorporating tumor grade and ER status into current AJCC staging. We recommend that biologic markers be incorporating into revised versions of the AJCC staging system for patients receiving surgery as the first intervention.^ Chapter 1 focuses on developing a Bayesian method to solve misclassified relapse status and application to breast cancer data. Chapter 2 focuses on evaluation of a breast cancer nomogram for predicting risk of IBTR in patients with DCIS after local excision gives the statement of the problem in the clinical research. Chapter 3 focuses on validation of a novel staging system for disease-specific survival in patients with breast cancer treated with surgery as the first intervention. ^