Identification of Reduced ERAP2 Expression and a Novel HLA Allele as Components of a Risk Score for Susceptibility to Liver Injury Due to Amoxicillin-Clavulanate.

BACKGROUND & AIMS: Drug-induced liver injury (DILI) due to amoxicillin-clavulanate (AC) has been associated with HLA-A∗02:01, HLA-DRB1∗15:01, and rs2476601, a missense variant in PTPN22. The aim of this study was to identify novel risk factors for AC-DILI and to construct a genetic risk score (GRS). METHODS: Transcriptome-wide association study and genome-wide association study analyses were performed on 444 AC-DILI cases and 10,397 population-based controls of European descent. Associations were confirmed in a validation cohort (n = 133 cases and 17,836 population-based controls). Discovery and validation AC-DILI cases were also compared with 1358 and 403 non-AC-DILI cases. RESULTS: Transcriptome-wide association study revealed a significant association of AC-DILI risk with reduced liver expression of ERAP2 (P = 3.7 × 10-7), coding for an aminopeptidase involved in antigen presentation. The lead eQTL single nucleotide polymorphism, rs1363907 (G), was associated with AC-DILI risk in the discovery (odds ratio [OR], 1.68; 95% CI, 1.23-1.66; P = 1.7 × 10-7) and validation cohorts (OR, 1.2; 95% CI, 1.04-2.05; P = .03), following a recessive model. We also identified HLA-B∗15:18 as a novel AC-DILI risk factor in both discovery (OR, 4.19; 95% CI, 2.09-8.36; P = 4.9 × 10-5) and validation (OR, 7.78; 95% CI, 2.75-21.99; P = .0001) cohorts. GRS, incorporating rs1363907, rs2476601, HLA-B∗15:18, HLA-A∗02:01, and HLA-DRB1∗15:01, was highly predictive of AC-DILI risk when cases were analyzed against both general population and non-AC-DILI control cohorts. GRS was the most significant predictor in a regression model containing known AC-DILI clinical risk characteristics and significantly improved the predictive model. CONCLUSIONS: We identified novel associations of AC-DILI risk with ERAP2 low expression and with HLA-B∗15:18. GRS based on the 5 risk variants may assist AC-DILI causality assessment and risk management.

Value of liver biopsy in the diagnosis of drug-induced liver injury.

BACKGROUND & AIMS: The utility of liver biopsy in diagnosing or staging idiosyncratic drug-induced liver injury (DILI) is unclear. The aim of this study was to determine whether liver histology impacted causality assessment in suspected DILI using a novel simulation model. METHODS: Fifty patients enrolled in the DILI Network (DILIN) who had liver biopsies performed within 60 days of DILI onset were randomly selected. All had standard DILIN consensus causality scoring using a 5-point scale (1=definite, 2=highly likely, 3=probable, 4=possible, 5=unlikely) based on 6-month post-injury data. Three experienced hepatologists independently performed a causality assessment using redacted case records, with the biopsy and selected post-biopsy laboratory data removed. The 3 hepatologists also reviewed the liver histology with a hepatopathologist and then repeated causality assessment for each case. RESULTS: Of the 50 cases, there were 42 high causality DILI cases (1, 2 or 3) and 8 low causality cases (4 and 5). The hepatologists judged that liver biopsy was indicated in 62% of patients; after histology review, biopsy was judged to have been helpful in 70% of patients. Histology review changed the causality score in 68% of patients, with an increase in DILI likelihood in 48% and a decrease in 20%. Biopsy results changed diagnostic certainty from less certain (3 or 4) to highly certain (1, 2 or 5) in 38% of patients. CONCLUSIONS: Liver histologic findings may help clarify the diagnosis of DILI. Histology appears to be particularly helpful in cholestatic or equivocal cases of DILI (possible or probable), shifting assessment toward a greater or lower certainty of a DILI diagnosis. LAY SUMMARY: The utility of liver biopsy in diagnosing or staging idiosyncratic drug-induced liver injury (DILI) is unclear. Herein, we show that, in patients with suspected DILI, a liver biopsy can help physicians diagnose DILI or other causes of liver injury with more certainty.

Tools for causality assessment in drug-induced liver disease.

PURPOSE OF REVIEW: There are three liver-specific causality assessment tools currently available to guide clinical diagnosis of Drug-Induced Liver Injury (DILI): Roussel-Uclaf Causality Assessment Method (RUCAM), Digestive-Disease-Week Japan 2004 scale (DDW-J), and Clinical Diagnostic Scale (CDS). The purpose of this review is to assess these tools and discuss how to improve the causality assessment process as a whole. RECENT FINDINGS: Existing DILI-specific causality assessment tools are surprisingly similar and exhibit only minor differences in point allocation. But difference in threshold for likelihood of being DILI. We reviewed the literature on currently used causality assessment tools, identified areas for future improvement, and herein propose approaches for refinement. Opportunities to improve current models, as well as the assessment process, in general, include in particular provision of more precise clinical detail and to perhaps add new components to scoring systems. For example, the incorporation of drug-specific clinical signature patterns, accounting for a drug's inherent hepatotoxicity potential, and/or incorporation of other drug properties to scoring systems may allow enhancement. Further, more systemic exclusion of competing diagnoses is needed. Finally, causality assessment processes will likely benefit from a data-driven and computer-assisted approach. SUMMARY: Current tools used for DILI adjudication are imperfect. Avenues to improve these tools are described.

Association of Liver Injury From Specific Drugs, or Groups of Drugs, With Polymorphisms in HLA and Other Genes in a Genome-Wide Association Study.

BACKGROUND & AIMS: We performed a genome-wide association study (GWAS) to identify genetic risk factors for drug-induced liver injury (DILI) from licensed drugs without previously reported genetic risk factors. METHODS: We performed a GWAS of 862 persons with DILI and 10,588 population-matched controls. The first set of cases was recruited before May 2009 in Europe (n = 137) and the United States (n = 274). The second set of cases were identified from May 2009 through May 2013 from international collaborative studies performed in Europe, the United States, and South America. For the GWAS, we included only cases with patients of European ancestry associated with a particular drug (but not flucloxacillin or amoxicillin-clavulanate). We used DNA samples from all subjects to analyze HLA genes and single nucleotide polymorphisms. After the discovery analysis was concluded, we validated our findings using data from 283 European patients with diagnosis of DILI associated with various drugs. RESULTS: We associated DILI with rs114577328 (a proxy for A*33:01 a HLA class I allele; odds ratio [OR], 2.7; 95% confidence interval [CI], 1.9-3.8; P = 2.4 × 10-8) and with rs72631567 on chromosome 2 (OR, 2.0; 95% CI, 1.6-2.5; P = 9.7 × 10-9). The association with A*33:01 was mediated by large effects for terbinafine-, fenofibrate-, and ticlopidine-related DILI. The variant on chromosome 2 was associated with DILI from a variety of drugs. Further phenotypic analysis indicated that the association between DILI and A*33:01 was significant genome wide for cholestatic and mixed DILI, but not for hepatocellular DILI; the polymorphism on chromosome 2 was associated with cholestatic and mixed DILI as well as hepatocellular DILI. We identified an association between rs28521457 (within the lipopolysaccharide-responsive vesicle trafficking, beach and anchor containing gene) and only hepatocellular DILI (OR, 2.1; 95% CI, 1.6-2.7; P = 4.8 × 10-9). We did not associate any specific drug classes with genetic polymorphisms, except for statin-associated DILI, which was associated with rs116561224 on chromosome 18 (OR, 5.4; 95% CI, 3.0-9.5; P = 7.1 × 10-9). We validated the association between A*33:01 terbinafine- and sertraline-induced DILI. We could not validate the association between DILI and rs72631567, rs28521457, or rs116561224. CONCLUSIONS: In a GWAS of persons of European descent with DILI, we associated HLA-A*33:01 with DILI due to terbinafine and possibly fenofibrate and ticlopidine. We identified polymorphisms that appear to be associated with DILI from statins, as well as 2 non-drug-specific risk factors.

Death and liver transplantation within 2 years of onset of drug-induced liver injury.

Drug-induced liver injury (DILI) is an important cause of death and indication for liver transplantation (fatality). The role of DILI in these fatalities is poorly characterized, particularly when fatalities occur >26 weeks after DILI onset. We analyzed patients in the US Drug-Induced Liver Injury Network prospective study having a fatal outcome within 2 years of onset. Each case was reviewed by eight network investigators and categorized as DILI having a primary, a contributory, or no role in the fatality. We subcategorized primary role cases as acute, chronic, acute-on-chronic, or acute cholestatic liver failure. For contributory and no role cases, we assigned a primary cause of death. Among 1,089 patients, 107 (9.8%) fatalities occurred within 2 years. DILI had a primary role in 68 (64%), a contributory role in 15 (14%), and no role in 22 (21%); 2 had insufficient data. Among primary role cases, 74% had acute, 13% chronic, 7% acute on chronic, and 6% acute cholestatic failure. For the 15 contributory role cases, common causes of death included sepsis, malignancy, and severe cutaneous reactions with multiorgan failure. For the 22 no role cases, malignancies accounted for most fatalities. Higher bilirubin, coagulopathy, leukocytosis, and thrombocytopenia were independently associated with DILI fatalities. New R ratio Hy's law had a higher positive predictive value for overall fatality (14% versus 10%) and a stronger independent association with DILI fatalities within 26 weeks compared to the original version of Hy's law (hazard ratio, 6.2, 95% confidence interval 3.4-11.1, versus 2.2, 95% confidence interval 1.3-3.7). CONCLUSIONS: DILI leads directly or indirectly to fatality in 7.6% of cases; 40% of these had nonacute liver failure courses. New R ratio Hy's law better identifies risk for death compared to the original Hy's law. (Hepatology 2017;66:1275-1285).

Assessing agreement with relative area under the coverage probability curve.

There has been substantial statistical literature in the last several decades on assessing agreement, and coverage probability approach was selected as a preferred index for assessing and improving measurement agreement in a core laboratory setting. With this approach, a satisfactory agreement is based on pre-specified high satisfactory coverage probability (e.g., 95%), given one pre-specified acceptable difference. In practice, we may want to have quality control on more than one pre-specified differences, or we may simply want to summarize the agreement based on differences up to a maximum acceptable difference. We propose to assess agreement via the coverage probability curve that provides a full spectrum of measurement error at various differences/disagreement. Relative area under the coverage probability curve is proposed for the summary of overall agreement, and this new summary index can be used for comparison of different intra-methods or inter-methods/labs/observers' agreement. Simulation studies and a blood pressure example are used for illustration of the methodology. Copyright © 2016 John Wiley & Sons, Ltd.

Methods for assessing the reliability of quality of life based on SF-36.

The 36-Item Short Form Health Survey (SF-36) has been widely used to measure quality of life. Reliability has been traditionally assessed by intraclass correlation coefficient (ICC), which is equivalent to Cronbach's alpha theoretically. However, it is a scaled assessment of reliability and does not indicate the extent of differences because of measurement error. In this paper, total deviation index (TDI) is used to interpret the magnitude of measurement error for SF-36, and a new formula for computing TDI for average item score is proposed. The interpretation based on TDI is simple and intuitive by providing, with a high probability, the expected difference that is because of measurement error. We also show that a high value of ICC does not always correspond to a smaller magnitude of measurement error, which indicates that ICC can sometimes provide a false sense of high reliability. The methodology is illustrated with reported SF-36 data from the literature and from real data in the Arthritis Self-Management Program. Copyright © 2016 John Wiley & Sons, Ltd.

Amoxicillin-Clavulanate-Induced Liver Injury.

BACKGROUND AND AIMS: Amoxicillin-clavulanate (AC) is the most frequent cause of idiosyncratic drug-induced injury (DILI) in the US DILI Network (DILIN) registry. Here, we examined a large cohort of AC-DILI cases and compared features of AC-DILI to those of other drugs. METHODS: Subjects with suspected DILI were enrolled prospectively, and cases were adjudicated as previously described. Clinical variables and outcomes of patients with AC-DILI were compared to the overall DILIN cohort and to DILI caused by other antimicrobials. RESULTS: One hundred and seventeen subjects with AC-DILI were identified from the cohort (n = 1038) representing 11 % of all cases and 24 % of those due to antimicrobial agents (n = 479). Those with AC-DILI were older (60 vs. 48 years, P < 0.001). AC-DILI was more frequent in men than women (62 vs. 39 %) compared to the overall cohort (40 vs. 60 %, P < 0.001). The mean time to symptom onset was 31 days. The Tb, ALT, and ALP were 7 mg/dL, 478, and 325 U/L at onset. Nearly all liver biopsies showed prominent cholestatic features. Resolution of AC-DILI, defined by return of Tb to <2.5 mg/dL, occurred on average 55 days after the peak value. Three female subjects required liver transplantation, and none died due to DILI. CONCLUSION: AC-DILI causes a moderately severe, mixed hepatocellular-cholestatic injury, particularly in older men, unlike DILI in general, which predominates in women. Although often protracted, eventual apparent recovery is typical, particularly for men and usually in women, but three women required liver transplantation.

A simulation study of nonparametric total deviation index as a measure of agreement based on quantile regression.

Total deviation index (TDI) captures a prespecified quantile of the absolute deviation of paired observations from raters, observers, methods, assays, instruments, etc. We compare the performance of TDI using nonparametric quantile regression to the TDI assuming normality (Lin, 2000). This simulation study considers three distributions: normal, Poisson, and uniform at quantile levels of 0.8 and 0.9 for cases with and without contamination. Study endpoints include the bias of TDI estimates (compared with their respective theoretical values), standard error of TDI estimates (compared with their true simulated standard errors), and test size (compared with 0.05), and power. Nonparametric TDI using quantile regression, although it slightly underestimates and delivers slightly less power for data without contamination, works satisfactorily under all simulated cases even for moderate (say, ≥40) sample sizes. The performance of the TDI based on a quantile of 0.8 is in general superior to that of 0.9. The performances of nonparametric and parametric TDI methods are compared with a real data example. Nonparametric TDI can be very useful when the underlying distribution on the difference is not normal, especially when it has a heavy tail.

The use of routine non density calibrated clinical computed tomography data as a potentially useful screening tool for identifying patients with osteoporosis.

OBJECTIVES: To evaluate whether lumbar vertebral body density CT attenuation values measured in Hounsfield Units (HUs) on routine Computed Tomography (CT) examinations can be reliably measured with limited variability, and to evaluate for a correlation between HUs and bone mineral density as measured by dual energy X-ray absorptiometry (DXA) scan. METHODS: Retrospective review of a total of 249 routine MDCT examinations, performed to measure HUs at the first non-rib bearing lumbar vertebral body on axial images, cross-referenced to the lateral scout image. RESULTS: The overall ICC and RC for intra-reader variability on CT HU were 0.987 (95% CI 0.973 - 0.999) and 15.664 (95% CI 11.66-16.97). The overall ICC and RDC for inter-reader variability on CT HU were 0.952 (95% CI 0.892 - 0.999) and 30.20 (95% CI 23.73 - 34.48). The ICC and RC for interscanner variability were 0.98 (95% CI 0.95 - 0.99) and 16.67 (95% CI 13.13 - 22.85). The correlation between the L1 HUs and L1 BMD, L1 t-score, and overall t-score was 0.437, 0.392, and 0.400, respectively. CONCLUSIONS: CT attenuation values of the first lumbar vertebra can be measured on routine abdomen CTs with limited variability despite multiple readers and scanners. Correlation between HU and BMD as measured by DXA scan was only weakly positive, and by this method measuring the density of a lumbar vertebral body from a routine MDCT scan does not provide the sensitivity or specificity necessary for a screening test. However above a certain measured value (180 HU), patients have a low chance of osteoporosis and therefore may not need additional screening, potentially limiting radiation exposure and cost.

Reliability of causality assessment for drug, herbal and dietary supplement hepatotoxicity in the Drug-Induced Liver Injury Network (DILIN).

BACKGROUND & AIMS: Because of the lack of objective tests to diagnose drug-induced liver injury (DILI), causality assessment is a matter of debate. Expert opinion is often used in research and industry, but its test-retest reliability is unknown. To determine the test-retest reliability of the expert opinion process used by the Drug-Induced Liver Injury Network (DILIN). METHODS: Three DILIN hepatologists adjudicate suspected hepatotoxicity cases to one of five categories representing levels of likelihood of DILI. Adjudication is based on retrospective assessment of gathered case data that include prospective follow-up information. One hundred randomly selected DILIN cases were re-assessed using the same processes for initial assessment but by three different reviewers in 92% of cases. RESULTS: The median time between assessments was 938 days (range 140-2352). Thirty-one cases involved >1 agent. Weighted kappa statistics for overall case and individual agent category agreement were 0.60 (95% CI: 0.50-0.71) and 0.60 (0.52-0.68) respectively. Overall case adjudications were within one category of each other 93% of the time, while 5% differed by two categories and 2% differed by three categories. Fourteen per cent crossed the 50% threshold of likelihood owing to competing diagnoses or atypical timing between drug exposure and injury. CONCLUSIONS: The DILIN expert opinion causality assessment method has moderate interobserver reliability but very good agreement within one category. A small but important proportion of cases could not be reliably diagnosed as ≥50% likely to be DILI.

Numeric score-based conditional and overall change-in-status indices for ordered categorical data.

Planned interventions and/or natural conditions often effect change on an ordinal categorical outcome (e.g., symptom severity). In such scenarios, it is sometimes desirable to assign a priori scores to observed changes in status, typically giving higher weight to changes of greater magnitude. We define change indices for such data based upon a multinomial model for each row of a c × c table, where the rows represent the baseline status categories. We distinguish an index designed to assess conditional changes within each baseline category from two others designed to capture overall change. One of these overall indices measures expected change across a target population. The other is scaled to capture the proportion of total possible change in the direction indicated by the data, so that it ranges from -1 (when all subjects finish in the least favorable category) to +1 (when all finish in the most favorable category). The conditional assessment of change can be informative regardless of how subjects are sampled into the baseline categories. In contrast, the overall indices become relevant when subjects are randomly sampled at baseline from the target population of interest, or when the investigator is able to make certain assumptions about the baseline status distribution in that population. We use a Dirichlet-multinomial model to obtain Bayesian credible intervals for the conditional change index that exhibit favorable small-sample frequentist properties. Simulation studies illustrate the methods, and we apply them to examples involving changes in ordinal responses for studies of sleep deprivation and activities of daily living.

Secondary analysis of APPLE study suggests atorvastatin may reduce atherosclerosis progression in pubertal lupus patients with higher C reactive protein.

OBJECTIVE: Participants in the Atherosclerosis Prevention in Paediatric Lupus Erythematosus (APPLE) trial were randomised to placebo or atorvastatin for 36 months. The primary endpoint, reduced carotid intima medial thickness (CIMT) progression, was not met but atorvastatin-treated participants showed a trend of slower CIMT progression. Post-hoc analyses were performed to assess subgroup benefit from atorvastatin therapy. METHODS: Subgroups were prespecified and defined by age (> or ≤15.5 years), systemic lupus erythematosus (SLE) duration (> or ≤24 months), pubertal status (Tanner score≥4 as post-pubertal or <4 as pre-pubertal), low density lipoprotein cholesterol (LDL) (≥ or <110 mg/dl) and high-sensitivity C reactive protein (hsCRP) (≥ or <1.5 mg/l). A combined subgroup (post-pubertal and hsCRP≥1.5 mg/l) was compared to all others. Longitudinal linear mixed-effects models were developed using 12 CIMT and other secondary APPLE outcomes (lipids, hsCRP, disease activity and damage, and quality of life). Three way interaction effects were assessed for models. RESULTS: Significant interaction effects with trends of less CIMT progression in atorvastatin-treated participants were observed in pubertal (3 CIMT segments), high hsCRP (2 CIMT segments), and the combined high hsCRP and pubertal group (5 CIMT segments). No significant treatment effect trends were observed across subgroups defined by age, SLE duration, LDL for CIMT or other outcome measures. CONCLUSIONS: Pubertal status and higher hsCRP were linked to lower CIMT progression in atorvastatin-treated subjects, with most consistent decreases in CIMT progression in the combined pubertal and high hsCRP group. While secondary analyses must be interpreted cautiously, results suggest further research is needed to determine whether pubertal lupus patients with high CRP benefit from statin therapy. TRIAL REGISTRATION: Clinical Trials.gov Identifier: NCT00065806.

Cardiovascular toxicity of epoetin-alfa in patients with chronic kidney disease.

BACKGROUND: Recombinant erythropoietin has become a routine component of care of patients with chronic kidney disease reducing the need for blood transfusions but raising the risks for cardiovascular events. We undertook this secondary analysis of subjects enrolled in the Correction of Hemoglobin and Outcomes in Renal Insufficiency (CHOIR) trial to examine the interrelationships between epoetin-alfa maintenance doses utilized and achieved hemoglobin (Hb) irrespective of treatment target and randomized allocation. METHODS: We performed a post hoc analysis from the CHOIR trial. Inclusion criteria were Hb <11.0 g/dl and estimated glomerular filtration rates of 15-50 ml/min/1.73 m(2). To be included in the present analysis, subjects needed to be free of the composite event at 4 months, receive epoetin-alfa, and have ≥1 postbaseline Hb measurement. The mean weekly dose of epoetin-alfa received up to the time of first event or censure was the main exposure variable, while the achieved Hb at month 4 was the confounder representing the subject's underlying response to treatment. The primary outcome was the composite of death, heart failure hospitalization, stroke, or myocardial infarction. A Cox proportional hazard regression model was used in time-to-event analysis. RESULTS: Among 1,244 subjects with complete data, the average weekly dose of epoetin-alfa ranged 143.3-fold from 133 to 19,106 units/week at the time of first event or censure. Cox proportional hazard analysis found that those in the middle tertile of Hb achieved (>11.5 to <12.7 g/dl) and the lowest tertile of epoetin-alfa dose exposure level (<5,164 units/week) had the lowest risk. Irrespective of Hb achieved, the relative risk in the highest tertile (>10,095 units/week) of epoetin-alfa dose exposure level was significantly escalated (hazard ratios ranged from 2.536 to 3.572, p < 0.05, when compared to the group of middle Hb tertile and lowered dose tertile). In a multivariable model that adjusted for achieved Hb, albumin, cholesterol, age, prior heart failure, prior stroke, prior deep venous thrombosis, atrial fibrillation or malignancy, the average weekly dose had a significant (p = 0.005) relative risk of 1.067 per 1,000 units of epoetin-alfa for the primary end point. CONCLUSIONS: In the CHOIR trial, average epoetin-alfa doses >10,095 units/week were associated with increased risks for cardiovascular events irrespective of the Hb achieved within the first 4 months of treatment. These data suggest the weekly epoetin-alfa dose and not the Hb achieved was a principal determinant in the primary outcome observed implicating a cardiovascular toxicity of this erythrocyte-stimulating agent.

Multiparametric Quantitative Imaging Biomarker as a Multivariate Descriptor of Health: A Roadmap.

Multiparametric quantitative imaging biomarkers (QIBs) offer distinct advantages over single, univariate descriptors because they provide a more complete measure of complex, multidimensional biological systems. In disease, where structural and functional disturbances occur across a multitude of subsystems, multivariate QIBs are needed to measure the extent of system malfunction. This paper, the first Use Case in a series of articles on multiparameter imaging biomarkers, considers multiple QIBs as a multidimensional vector to represent all relevant disease constructs more completely. The approach proposed offers several advantages over QIBs as multiple endpoints and avoids combining them into a single composite that obscures the medical meaning of the individual measurements. We focus on establishing statistically rigorous methods to create a single, simultaneous measure from multiple QIBs that preserves the sensitivity of each univariate QIB while incorporating the correlation among QIBs. Details are provided for metrological methods to quantify the technical performance. Methods to reduce the set of QIBs, test the superiority of the mp-QIB model to any univariate QIB model, and design study strategies for generating precision and validity claims are also provided. QIBs of Alzheimer's Disease from the ADNI merge data set are used as a case study to illustrate the methods described.

Multiparametric Data-driven Imaging Markers: Guidelines for Development, Application and Reporting of Model Outputs in Radiomics.

This paper is the fifth in a five-part series on statistical methodology for performance assessment of multi-parametric quantitative imaging biomarkers (mpQIBs) for radiomic analysis. Radiomics is the process of extracting visually imperceptible features from radiographic medical images using data-driven algorithms. We refer to the radiomic features as data-driven imaging markers (DIMs), which are quantitative measures discovered under a data-driven framework from images beyond visual recognition but evident as patterns of disease processes irrespective of whether or not ground truth exists for the true value of the DIM. This paper aims to set guidelines on how to build machine learning models using DIMs in radiomics and to apply and report them appropriately. We provide a list of recommendations, named RANDAM (an abbreviation of "Radiomic ANalysis and DAta Modeling"), for analysis, modeling, and reporting in a radiomic study to make machine learning analyses in radiomics more reproducible. RANDAM contains five main components to use in reporting radiomics studies: design, data preparation, data analysis and modeling, reporting, and material availability. Real case studies in lung cancer research are presented along with simulation studies to compare different feature selection methods and several validation strategies.

Multiparametric Quantitative Imaging in Risk Prediction: Recommendations for Data Acquisition, Technical Performance Assessment, and Model Development and Validation.

Combinations of multiple quantitative imaging biomarkers (QIBs) are often able to predict the likelihood of an event of interest such as death or disease recurrence more effectively than single imaging measurements can alone. The development of such multiparametric quantitative imaging and evaluation of its fitness of use differs from the analogous processes for individual QIBs in several key aspects. A computational procedure to combine the QIB values into a model output must be specified. The output must also be reproducible and be shown to have reasonably strong ability to predict the risk of an event of interest. Attention must be paid to statistical issues not often encountered in the single QIB scenario, including overfitting and bias in the estimates of model performance. This is the fourth in a five-part series on statistical methodology for assessing the technical performance of multiparametric quantitative imaging. Considerations for data acquisition are discussed and recommendations from the literature on methodology to construct and evaluate QIB-based models for risk prediction are summarized. The findings in the literature upon which these recommendations are based are demonstrated through simulation studies. The concepts in this manuscript are applied to a real-life example involving prediction of major adverse cardiac events using automated plaque analysis.

Multiparametric Quantitative Imaging Biomarkers for Phenotype Classification: A Framework for Development and Validation.

This manuscript is the third in a five-part series related to statistical assessment methodology for technical performance of multi-parametric quantitative imaging biomarkers (mp-QIBs). We outline approaches and statistical methodologies for developing and evaluating a phenotype classification model from a set of multiparametric QIBs. We then describe validation studies of the classifier for precision, diagnostic accuracy, and interchangeability with a comparator classifier. We follow with an end-to-end real-world example of development and validation of a classifier for atherosclerotic plaque phenotypes. We consider diagnostic accuracy and interchangeability to be clinically meaningful claims for a phenotype classification model informed by mp-QIB inputs, aiming to provide tools to demonstrate agreement between imaging-derived characteristics and clinically established phenotypes. Understanding that we are working in an evolving field, we close our manuscript with an acknowledgement of existing challenges and a discussion of where additional work is needed. In particular, we discuss the challenges involved with technical performance and analytical validation of mp-QIBs. We intend for this manuscript to further advance the robust and promising science of multiparametric biomarker development.

A comparison of the effect of patient-specific versus weight-based protocols to treat vaso-occlusive episodes in the emergency department.

BACKGROUND: Vaso-occlusive crises (VOCs) cause debilitating pain and are a common cause of emergency department (ED) visits, for people with sickle cell disease (SCD). Strategies for achieving optimal pain control vary widely despite evidence-based guidelines. We tested existing guidelines and hypothesized that a patient-specific pain protocol (PSP) written by their SCD provider may be more effective than weight-based (WB) dosing of parenteral opiate medication, in relieving pain. METHODS: This study was a prospective, randomized controlled trial comparing a PSP versus WB protocol for patients presenting with VOCs to six EDs. Patients were randomized to a PSP or WB protocol prior to an ED visit. The SCD provider wrote their protocol and placed it in the electronic health record for future ED visits with VOC exclusion criteria that included preexisting PSP excluding parenteral opioid analgesia or outpatient use of buprenorphine or methadone or highly suspected for COVID-19. Pain intensity scores, side effects, and safety were obtained every 30 min for up to 6 h post-ED bed placement. The primary outcome was change in pain intensity score from placement in an ED space to disposition or 6 h. RESULTS: A total of 328 subjects were randomized; 104 participants enrolled (ED visit, target n = 230) with complete data for 96 visits. The study was unable to reach the target sample size and stopped early due to the impact of COVID-19. We found no significant differences between groups in the primary outcome; patients randomized to a PSP had a shorter ED length of stay (p = 0.008), and the prevalence of side effects was low in both groups. Subjects in both groups experienced both a clinically meaningful and a statistically significant decrease in pain (27 mm on a 0- to 100-mm scale). CONCLUSIONS: We found a shorter ED length of stay for patients assigned to a PSP. Patients in both groups experienced good pain relief without significant side effects.

Effect of hemoglobin target on progression of kidney disease: a secondary analysis of the CHOIR (Correction of Hemoglobin and Outcomes in Renal Insufficiency) trial.

BACKGROUND: Conflicting relationships have been described between anemia correction using erythropoiesis-stimulating agents and progression of chronic kidney disease (CKD). This study was undertaken to examine the impact of target hemoglobin level on progression of kidney disease in the CHOIR (Correction of Hemoglobin and Outcomes in Renal Insufficiency) trial. STUDY DESIGN: Secondary analysis of a randomized controlled trial. SETTING & PARTICIPANTS: 1,432 participants with CKD and anemia. INTERVENTION: Participants were randomly assigned to target hemoglobin levels of 13.5 versus 11.3 g/dL with the use of epoetin alfa. OUTCOMES & MEASUREMENTS: Cox regression was used to estimate HRs for progression of CKD (a composite of doubling of creatinine level, initiation of renal replacement therapy, or death). Interactions between hemoglobin target and select baseline variables (estimated glomerular filtration rate, proteinuria, diabetes, heart failure, and smoking history) also were examined. RESULTS: Participants randomly assigned to higher hemoglobin targets experienced shorter time to progression of kidney disease in both univariate (HR, 1.25; 95% CI, 1.03-1.52; P = 0.02) and multivariable models (HR, 1.22; 95% CI, 1.00-1.48; P = 0.05). These differences were attributable to higher rates of renal replacement therapy and death for participants in the high hemoglobin arm. Hemoglobin target did not interact with estimated glomerular filtration rate, proteinuria, diabetes, or heart failure (P > 0.05 for all). In the multivariable model, hemoglobin target interacted with tobacco use (P = 0.04) such that the higher target had a greater risk of CKD progression for participants who currently smoked (HR, 2.50; 95% CI, 1.23-5.09; P = 0.01), which was not present for those who did not currently smoke (HR, 1.15; 95% CI, 0.93-1.41; P = 0.2). LIMITATIONS: A post hoc analysis; thus, cause and effect cannot be determined. CONCLUSIONS: These results suggest that a high hemoglobin target is associated with a greater risk of progression of CKD. This risk may be augmented by concurrent smoking. Further defining the mechanism of injury may provide insight into methods to optimize outcomes in anemia management.